1999 Vol 6 Compilation

VALVE
the magazine of astounding sound
Buddha extreme FM-3 mod
Boulders a turntable odyssey, part two
Doc B.C4S’ing the S.E.X. amp
Epstein a visit with Fi’s Don Garber
Brainiac heavy metal
volume 6
number 1
1999
Doc B.
1998
Paraglow
a sonic masterpiece by Electronic Tonalities
VALVE
the magazine of
astounding sound
Editor and Publisher
Dan “Dr. Bottlehead” Schmalle
Chief Administrator
“Queen Eileen” Schmalle
Resident Smart Guy and Technical Editor
Paul “Braniac” Joppa
Graphic Design
Bruce “Papa” Borley
Resident Hot Iron and Dr. B’s Bodyguard
John “Smoothplate” Tucker
Big gun OEM advisor
Michael “Airgap” LaFevre
Guru’s guru
John “Buddha” Camille
Contributing Editors
David “Full Track” Dintenfass
“Crazy Eric” Lenius
Our mailing address:
VALVE
P.O. Box 2786
Poulsbo, WA 98370
by phone: 360-697-1936
business hours: 9-5 PST, Mon -Fri
fax: 360-697-3348
e-mail - [email protected]
website - http://www.bottlehead.com
VALVE in no way assumes responsibility for
anyone harming themselves through exposure to
the contents of this magazine. W e believe electrons
flow from minus to plus, and that they can kill you
along the way if you’re not careful. Vacuum tube
audio equipment operates at potentially lethal
voltages. Always treat it with respect.
Many ideas published in this magazine are untried,
and involve the use of potentially dangerous parts
and tools. In attempting any idea or project
published herein, you assume total responsibility for
your actions and any harm caused to yourself or
others. Please, be careful!
This publication is produced as a service to the
audio community and is wholly owned and
published by Electronic Tonalities.
The intent of this publication is to offer ideas to
inspire and educate audiophiles in an effort to
increase their understanding of the audio
equipment they use and cherish. Blatant copying of
the circuits published in this magazine for use in
commercial products shows a complete lack of
original thought.
editor’s
thing
Once again we greet you in a new
format, as we have upon occasion over the years (has
it been five already?!). Time and technology march on,
and we flounder to keep up.
Welcome to the first edition of the new electronically
published VALVE. The magazine has been produced
with the intention that it will print well, and we encourage you to try printing a hard copy from the .PDF
file.
We had hoped to continue to be able to put out VALVE
in both electronic and hardcopy formats, but costs have
been rising and we just can’t handle the expense involved in putting out printed issues of VALVE anymore.
But- on the bright side, subscriptions are a thing of
the past. VALVE will appear each month, downloadable
for free at bottlehead.com.
Yup, for free.
Along with creating a new downloadable issues
throughout the year, we will be changing the way we
package our back issues. We have been taking a great
deal of time to scan and convert our oldest issues, which
were pasted up in the traditional fashion, and also to
convert our later electronically published issues all into
Acrobat .PDF format. We hope to have them all available on CD-ROMs in the next month or so. The 1999
issues will also be available on CD ROM in December
of 1999, in case you miss any issues before they are
taken down from bottlehead.com and replaced by a
new issue during the year.
What else is new for 99?
A new website look - hopefully on your way to finding
this issue you found our new website. After doing a
nice job of getting bottlehead.com up and running,
Dave Stagner has handed the reins over to my fatherin-law, Bruce Borley. Bruce has many years of experi-
ence as a graphic designer, and consequently we will
be getting a pretty hip makeover as the months
progress.
New digs - we’re outgrowing our disorganized “home
orifice”setup that we’ve been working in for the past
seven years, and this year we will being remodeling to
make us more efficient. The ol’ basement workshop
will be revamped to include the office that up to now
has taken over our family room and also a new kit packing facility.
And thanks to a wonderful Xmas gift of cash from the
local VALVE members, instigated by Richard Riley and
Paul Joppa, we will begin construction on our long
awaited new listening room, to be built in the free
standing 12’x20’ garage that has been the recycling
receptacle at Bottlehead ranch for the last 12 years.
Thanks again for your generosity guys, Paul was tickled to see that for the first time since he’s known me I
was speechless. Bottlehead is taking trips to the recycle
center daily and he’s almost found the garage floor.
We’ll cover the adventure of putting this thing together
in future issues.
Don’t let the blue smoke out,
Doc B.
on the cover
Ray Kuehlthau sent us this photo of his latest woody
maker, a two chassis amp based on the now famous
Single Ended Direct Coupled Active Loaded Parallel
Feed 45 amp from October ‘97 VALVE.
This two chassis stunner features carbon fiber chassis
plates, zebrawood bases, brass bottom plates,
MagneQuest parafeed iron, the brass channel frames
of which Ray polished, and the rest of which he shot
with faux granite finish.
Below is another of Ray’s projects, this one his Paul
Joppa 300B S.E.X. amp mod. This guy can crank out
gorgeous projects like nobody.
New exotica from
First Impression Music
FIM CD019 Favorite Chinese Instrumentals
The Jin Ying Soloists
1 Like Wave Again The Sand (Pipa solo with ensemble) 4:36 2 Singing
the Night among Fishing Boats (Jung solo) 4:15 3 Night (Percussion
ensemble) 5:54 4 Love Song Of The Grassland (Tung-hsiao with
ensemble) 2:34 5 Ducks Quacking (Percussion ensemble) 4:00 6 Love
At The Fair (Er-hu solo with ensemble) 5:04 7 The Fishing Song (Bawoo solo with ensemble) 4:53 8 Happy Reunion (Xylophone solo with
ensemble) 2:55 9 Chinese Martial Art (Emsemble) 1:43 10 The Flowing Stream (Er-hu solo with Yang chin) 8:25 11 Spenpadei Folksong
(Yang-chin solo with ensemble) 4:32 12 Autumn Moon (Er-hu solo with
ensemble) 7:06 13 Moonlight Over The Spring River (ensemble) 9:22
14 Variation On Yang City Tune (Ku-jung solo) 5:36
FIM CD020 A Vocal Tribute to Ben Webster
Jacintha, Vocal
Teddy Edwards, tenor sax; Kei Akagi, piano; Darek Oles, bass; Larance
Marable, drums
1 Georgia On My Mind (Gorrell-Carmichael/BMI) 5:19 2 The Look Of
Love (Baccharach-David/ASCAP) 4:08 3 Danny Boy (Traditional) 7:26
4 Somewhere Over The Rainbow (Harburg-Allen/ASCAP) 9:43 5
Startdust (Parish-Carmichael/ASCAP) 6:40 6 In The Wee Small Hours
Of Morning (Hillard-Mann/ASCAP) 4:15 7 Tenderly (Lawrence-Gross/
ASCAP) 5:25 8 Our Love Is Here To Stay (George & Ira Gershwin/
ASCAP) 3:19 9 How Long Has This Been Going On? (George & Ira
Gershwin/ASCAP) 5:28 10 Pennies From Heaven (Burke-Johnston/
ASCAP) 3:12
#189 16149 Redmond Way
Redmond, WA 98052
http://www.fimpression.com
Tel: 425-868-5326
Fax: 425-836-9061
Buddhafying the
Dyna FM-3
By John “Buddha” Camille
You guys who know Buddha also know I let a djinni
out of the bottle when I asked him for a few simple
ideas on modding a little ’ol FM-3. Hang on to your
hats with this one bottleheads, by the time Buddha gets
through with you, you’ll not only have the meanest
FM-3 in town, you’ll have the cleanest tubes too. - Doc
B.
Dear Dr. R sub g (Rg),
Was very amused at the third issue of “V”. Your new
subtitle bears out the caliber of your readership. (Buddha is of course referring to the unending string of
corrections I got regarding that dumb grid resistor I
left on the output tube in the Blues Master article - B.)
The FM-3 has turned into a real challenge. My thoughts
of hot-rodding the tuner and IF strips were dashed by
the wide open construction of the unit. Using higher
gm tubes for the mixer and IF strips was not possible
due to the lack of shielding.
Above a certain total gain
point the RF-IF strip takes off
and will oscillate at around
1
0
7
MHz, the tenth harmonic of
the IF. The feedback path is
primarily through radiation
from the last IF stage back
through the RF stage grid circuit, in concert with direct
conduction through the B+
and heater circuits. There is
also a lot of direct radiation
above chassis from tubes and
the mickey mouse tube and IF
can shields. This problem prevented me from getting 60 dB
plus of limiting on noise alone
by using hotter tubes in the IF.
I did achieve a happy balance of gains between the RF
and IF that permits better limiting, slightly increased
total gain, and much better noise figure with the 6DJ8
front end. Several college stations thirty to forty miles
way now come in full quieting with a 3 foot clip lead
antenna.
Eventually I hope to fully shield the RF section and
possibly the bottom of the IF with proper interstage
shielding and tons of bulkhead feed-throughs. I did
cheat and use button mica caps on the RF board while
trying to calm oscillations. These caps get the RF off of
the lower cathode and upper grid and of course increased RF gain considerably. This required lowering
the gain on the IF strip. This was done by lowering
plate and screen voltages, which had the added benefit of providing better limiting. Virtually any signal
will now produce a limited 4Vpk-pk signal on the plate
of V-6. A slight bit of AM’ing is still visible on positive
peaks, however. Before modding, AM was significant
on both positive and negative peaks. The beast is really
starting to sound solid.
I have not had time to redo the first audio stage, V-7,
and the audio output stage V-72. Plan to use a 5687
for V-7 and 12B4’s for V-72. Of course, C4S’s will be
used for cathode and plate follower feedback on the
output. Will probably change the volume control to a
6dB stepped attenuator for better high frequency response.
The present 19,38 and 67kHz attenuation really sucks.
Plan to use plenty of LC traps and filtering here, along
with a proper de-emphasis filter. May need a cathode
follower to drive these filters properly. The present system has a lot of 38 kHz on the output and tons of 67
kHz SCA on some stations.
Hum has always been a problem on my unit. Got rid of
about 95% of it with additional filtering and direct
wiring the primary power. Putting 120 VAC (to the
on-off switch) within 1/2 inch of the first audio stage
is not very cool. Am planning to eventually move the
power supply off chassis and DC everything in a large
audio visual console I had built in the Phillipines back
in the 60’s. The wood, Nara, is too hard to work so I’ll
have to stick with the heavily modded stuff that will
fit the holes on the cabinet work.
Tw eaks that m ight be applicable to the ex perim enter
RF-IF Tube Shielding
The original tube shields are tube killers. I had replaced
the tin can shields with IERC style shields when the
unit was built back in the early 60’s. All tubes still have
plenty of emmision with probably 100K hours of use.
These shields are still hot with RF however, as the original ground strip corrodes pretty quick. A better answer is my combination tube shield/microphonic attenuator. This process should be used on all tubes in
the set, especially the RF-IF tubes. Some folks might
not like the sonics when this technique is used on the
audio tubes because much of the reverb is lost...
Procedure
1
2
3
Select a new tube that checks way over good.
Wire brush pins until shiny with a gun bore
brush chucked up in a drill motor or Dremel tool.
Practice on an old tube until you figure it out.
Wash tube thoroughly with tooth brush and 409.
Rinse and blow dry. Do not touch pins with fin-
gers again. We’re talking clean room plastic
gloves here (food handler gloves are cheaper)
4
Wrap one layer of 1” 3M foil shielding tape
around tube and solder the seam in several places.
Connect shield to the appropriate pin with 24
ga. wire wrap wire, see figure 1. The RFI tape
(Digi-Key 3M1181C) is pretty pricey but will be
a lifetime supply for most experimenters. I use
tons of it for ground planes, power busses, shielding, etc. The stuff is severly addictive.
5
Clean rosin from all solder points with multiple
Q-tips and anhydrous alcohol (Caig Caikleen) or
toluene. This process must be repeated several
times in order to remove all film traces.
6
Wash tube again with 409 and toothbrush (especially around tube pins). Rinse and blow dry.
Blow dry is important in order to prevent leakage from water rings from solutes. Do not try to
wipe dry.
7
Mask tube pins (use blue 3M masking tape) and
shoot tube with flat black Krylon. The paint and
copper drop envelope temperatures, significantly
increasing tube life. The glass/adhesive/copper
composite reduces microphonic ringing measurably. Note: for those using regular masking tape
- clean adhesive off of tube pins with Q-tips saturated with Goof-Off. Then degrease with 409,
etc.
8
Chuck one half of a
Q-tip in a drill motor and add
a small drop of Caig Pro-Gold
to the cotton. Use this rotary
buffer to polish tube pins.
Repest process with clean Qtips until pins produce no
more black deposits on cotton.
Surprise, surprise, you
thought they were clean.
Note: A short lead may be soldered to the copper shield and
run to the nearest analog
ground when all tube pins are
floating. Where necessary, a
1/16” hole may be drilled in
the chassis for this ground
connection. RF-IF and other
non-audio signal tube shields
may be returned to a chassis mounted solder lug, a
ground plane or a ground buss. Make the connection
short and sweet.
IF Transformer shielding
The IF can mounting lugs do not make a serious electrical connection to the aluminum can. All of my cans
were initially hot with RF. Physical movement of the
can would also produce intermittent contact and
detuning of the transformer. The IF transformers were
removed for overhaul and clean-up. While I was at it,
the winding assemblies were also removed from the
cans and the tank capacitors were replaced , and all
internal solder joints were sucked dry and resoldered.
Several joints looked suspicious where the magnet wire
insulation had not been volitized completely. In addition, several windings had broken loose from the wax
tie down process. The wax was removed with a slightly
warm solder iron tip and the wire was rewound in
place and recoated with the wax removed previously.
The warm solder iron can be used to remelt the wax
blob onto the winding and the coil former. A step by
step follows:
12
1
2
3
4
5
6
7
8
9
10
11
Witness mark IF cans and circuit board with a
scratch awl.prior to removal. The green alignment mark will go away later. Early cans had a
slot punched into the top of the can and this will
suffice.
Remove solder from board with solder sucker (6
places) and gently rock the can ‘around the clock’
as joints are heated in succession until the can is
free.
Witness mark the can and the coil assembly. A
green dot may or may not still exist an the winding frame as an alignment marker, just make sure
you get it right. For those who are reading the
instructions after the fact, the primary winding
is on the bottom next to the tank capacitors, see
schematic symbology.
Number each can in accordance with its circuit
location with the scratch awl, i.e., T-1 would be
numbered 43201-1, T-2 would be dash 2, etc.
Flatten the four coil frame mounting indentations
in the can with the needle nose pliers and gently
remove the winding assembly from the can. Done
right, the winding can be pulled out with finger
pressure. Don’t grip the solder lugs with pliers
and pull or you may have a real fabrication problem - the plastic frame is brittle.
Polish the can with 409 and a piece of 3M
Scotchbrite. The tin plated steel mounting lugs
should be brite for later soldering. rinse and blow
dry.
Wrap one turn of 1/2” wide copper tape around
the can so that approximately 1/2 of the mounting lug is covered. Solder the copper to both lugs
and tack the top of the seam, see figure 2.
Remove rosin from solder joints and degrease
with 409. Rinse and blow dry.
Shoot can with Krylon to forstall galvanic action.
Flat black is best for tube life.
Clip out the 22 pf disc capacitor from the primary side of the transformer (lower winding)
and replace with a 22 pF/N750 disc ceramic,
DIGI-KEY 1305PH. When soldering the new part
into the circuit, first wrap the leads around the
solder joints for the coils leads. This technique
will possibly solve a bad joint existant since the
original manufacture. Remove iron as soon as the
solder flows as the coil frame plastic is pretty
poor.
Clip out the 10pF disc capacitor attached to the
secondary terminals. Replace with a Phillips 10
pF /N750 disc DIGI-KEY 1301PH observing the
above cautions.
13
14
15
16
17
Measure coil resistance after modification.
Proper primary resistance is ~0.75Ω and the
secondary resistance is ~ 0.78Ω.
Align witness marks on the IF can and the coil
assembly and insert coil assembly into the can.
Insure coil form is fully seated in the circular boss
on the top of the can.
Use awl to indent the can back into the grooves
in the coil assembly base frame (four places).
Place 0.1 in. length of small teflon tubing or wire
insulation over pins 2 and 4 of the IF can.
Drill clearance holes in PC boards at the pins 2
and 4 locations for the above teflon insulation.
(typical 1/16” - 2 places per transformer). New
wiring will lift RF hot wiring off the lossy PC
board material
Strip PC board solder pads arpund these holes
and lift trace off board back to the plate or coupling cap solder pad. Apply soldering iron to the
plate lead or coupling cap and they may be lifted
off the board in one fell swoop. The idea is to get
as much of these RF leads as possible off the
crummy board material and build a decent Rf
transmission line. See fig. 3. Another aim is to
reduce circuit “Q” and widen bandwidth. The
direct wiring spec’d also reduces radiation ef-
fects.
Tube Sockets
The tubes sockets used in the FM-3 are a disaster. Each
socket required a half dozen round toothpicks dipped
in Pro-Gold to remove corrosion, etc. Twirl the picks
in the pin contacts until they come away clean
Grounding Tuning Cap Shaft
The tuning shaft is fairly hot with RF and needs to be
positively grounded. I used the old ham radio trick of
fabbing a clockwork spring buss between the shaft and
the shield. The shield around the variable cap also needs
to be tack soldered to the chassis every inch. This process is fairly easy when the filter cap (C-32) is replaced.
See next page.
RF Frequency Shift
Joe Roberts mentioned that the FM-3 is well known
for tuning instability, something I have never encountered. Having done a lot of VHF/UHF work, I’ve always been amazed how stable the thing is. Well, the
palmprint on the forehead occured when thumbing
these since some are 2pF feedthrus for signal leads most are 470 to 1500 pF for power leads and bypasses.
Other mods
1
2
3
the through the assembly manual and seeing the “real”
C-8. I had replaced the cheapass ceramic trimmer with
a Johanson air variable when I built the thing, so I had
experienced no problems. Digi-Key carries the
Sprague-Goodman line of trimmer caps that should
work well. I used a style 9, 1-10pF cap. pn SG4004.
Looks like a good fit. Surplus is better.
This type of cap, along with the button mica feedthrus,
can be found in those old RF assemblies sold by the
pound at the flea market. I paid 50 cents for a filter
box that contained 8 sapphire insulated Johansons
about a year ago. Samo, samo on bulkhead, ceramic,
feedthrus such as the SCI’s sold by Newark, p 151. Old
TV tuners that sell for a buck or two are usually filled
with feedthrus. Will need a capacitance checker with
4
The HP 5082-2810 diodes also work well for the
demultiplex circuit and they are a helluvalot
easier to match. I am going to put the 47K resistors and the diodes in the can for T-73. Leads are
too long the way they are presently scattered over
the PC board.
The whole de-mux board needs to be replaced
with an FR-4 proto board with ground plane.
Will probably steal the design for a good de-mux
from a Fisher or something highly acclaimed.
Might move V-7 over to the de-mux board and
put the whole thing in a die-cast box.
You are right getting rid of the six for a nickel
parts on the rear deck.(I suggested getting rid of
the crummy antenna connector screws and installing a nice F connector - B.) I am going to
BNC on audio stuff.
Well, enough B.S. for a while -
Brainiac’s
Heavy Metal
ably use a 6J5 or 6SN7 or 12AU7 tube in single ended
drive at reasonable current.
sidual ringing at around
250kHz - not bad at all!
Both sides are -1dB at
10kHz; they are down 3dB
at 25 and 28kHz. Bass extends to -1dB at 45 Hz, 3dB at 20 Hz with small
signals (7 volts rms)
This one is in a U-channel frame, with a lamination
stack of 1.875 by 1.5 by 0.625 inch and mounting
holes 2 3/8 inches apart. It weighs 8 oz. on my kitchen
scale.
...an occasional review of some nifty chunk of iron
(usually cheap!)
Low frequency data:
Today’s Topic:
the Allied “Ally” interstage transformer
single-ended
Turns ratio:
OK, so I couldn’t think of a clever name for this one! If
you have a better name, let me know!
Primary:
Secondary:
DC current:
This unit is available for $15 from Allied Radio, and is
one of the small class of cheap interstage units available. There are several in the $10 to $20 range, made
for old radios usually. Nothing fancy, but their small
size makes for low leakage inductance even without
interleaving and they can sometimes be used with decent results. They are usually designed for SE drive of
push-pull output tubes, so they are not optimum for
single-ended output. In my experience these transformers are often very inaccurate in their specifications;
they should be measured before a circuit is designed
around them. This one is attractive because it operates
at a relatively high impedance level. It can comfort-
Peak output:
1:1CT; specified for
10k:10kCT impedance
1240 ohm DC
798+731 ohm DC
Inductance with 12v/
60Hz excitation is 44H at
0v, falling to 22H at 24mA.
I would rate it at 12mA,
where inductance is 35H
52.5v + 52.5v at 40 Hz.
High frequency data (resistances for a decent-looking square ware at 5kHz)
push-pull:
Source impedance 8k plus
a load impedance of 12k +
12k leaves only a tiny re-
Seems to work best with 8k
source and a 15k load endto-end. It’s -1dB from
45Hz to 11kHz; -3dB at
20Hz and 20kHz. Not as
good as push-pull, but not
terrible either.
This is pretty decent performance, considering the
price. It could, just barely, drive a pair of 2A3’s in pushpull. The fact that it could do so with a 56 as driver is
attractive, though another gain stage would be needed
for any reasonable sensitivity. The relatively high impedance and decent current handling means the driver
can operate at reasonable voltages, as much as 300400v, something other interstage transformers have a
hard time with.
Paul Joppa
ActiveS.E.X.
6Kohm
3W
Camille Cascode
Constant Current
Source (C4S)
R1 = 0.95V/Iplate
R2 = Vtop of load/.002mA
300VDC
R1 60 ohm
1/4W
47 mfd
450V
Remove 1K WW resistor from power
supply. Replace second 10 mfd filter
410VDC
capacitor with a 47 mfd @ 450V
electrolytic capacitor
original SEX OPT
as plate choke
MPS4250 or
2N2907
3mfd 630VDC
MJE 350
R2 3 series 50K
ohm 1W
paralleled 6DN7
triodes no.2
2mA
250VDC
paralleled 6DN7
triodes no.1
.47 mfd
630VDC
270K
1/4W
100K
1/4W
A C4S adaptation for the
S.E.X. amd ParaS.E.X. kits
By Doc B.
8VDC
1Kohm
1/4W
16mA
new SEX parafeed
OPT
220mfd
35V
orange
220mfd
35V
390
ohms
2W
green
Fi‘n’ Art
by Jeremy Epstein
Don Garber, the man behind the Fi amplifier company, has an enviable life. Ensconced in a commodious
hundred-and-fifty-year-old Brooklyn carriage house,
Don paints abstract paintings and builds a unique line
of bottle-driven audio equipment. He has the good fortune to be supervised by Max, his shaggy big dog, and
inspired by Ikuyo, his muse, who was responsible for
my visit.
My wife was talking to Ikuyo, Don’s wife, at a party a
few months ago. She came and told me, “Ikie’s husband builds tube audio gear like that stuff you’re into.”
I explained to Don that I was a neophyte bottlehead,
and he was very kind. Later, I checked up on Fi and
quickly learned that my neighbor was responsible for
some of the most interesting wattage-challenged gear
around, and that he had been a pioneer in the emergence of triodemania here in the US.
So I imposed on his hospitality and
invited myself over to see what a real
professional amp builder looks like.
Piled neatly in the Fi production
space were dozens of chassis plates,
which Don had just finished drilling for the next batch of amps. Trannies and tubes were stacked in neat
little rows and aluminum sawdust
was everywhere. While Don chooses his parts with deliberate care,
according to quality, price, and their
synergy with the circuit application,
he doesn’t seem to me to be using
anything
wildly
esoteric
or mysterious. Since he builds in
small quantities, he gets his parts in
reasonably-sized batches : there
were no tell-tale crates revealing
bargain bulk sourcing. I buy my
parts from some of the same places
Don does. Most of the tubes he uses
are current production. About the only evidence of a
tweaky mentality I saw was the silver wire used inside
the amps, hardly voodoo. And Don admits, “I’m no
great audio designer, and I don’t want to contend that
I am. I guess I’ve looked at a lot of things, tried a lot of
things. My approach is more empirical than anything
else.” What he shoots for is simplicity, value, precise
workmanship and a clean look. He doesn’t much like
talking about audio, and hype is notDon’s style. It’s an
awful lot easier to get him talking about music.
”I’ve been struck by audiophiles in general, (I tend to
think my customers are a little better than most, or is
that just my own ego in there?) the music they listen
to - you’ll try to get past my snobbishness - is such
trash! But most of them, it’s just to demonstrate
their systems. And I find that this tends to be less true
among people that build their own stuff. You would
think it might be the other way around but it’s not.”
Don and his wife regularly calibrate their ears by attending chamber concerts at Bargemusic, a floating
concert space moored under the Brooklyn Bridge. A
window onto the lower Manhattan skyline forms a
spectacular backdrop for the world-class musicians
who perform there. Perhaps it’s his experience dealing with the audio salon crowd (Fi was
started partly to feed product into a retail store of the
same name Don ran a while back) but Don Garber is
no fussbudget. When I arrived at his shop/soundroom
he methodically set up his wonderful 2A3 monoblocks
to drive Exemplar horns with zip cord (Yo, Don,
Smoothplate’s gonna crap his pants when he reads this
-B.), no-name interconnects, and a little beer. Carefully, he tuned the room by moving a pile of cardboard
boxes out from in front of the right speaker, and
powerline conditioning was taken care of by a 50' orange extension cord and some cheater plugs. Together
we listened to some mutual favorites, like the new
Dylan live set and Lucinda Williams’ long-awaited new
album, and I turned him on to the terrific guitar
Goldberg Variations by Kurt Rodarmer. We talked
about a lot of things : Brooklyn real estate, family, slide
guitar, and even the business of audio. Neither one of
us was unduly bothered by the cheap interconnects,
let’s just say, and my cassette of our interview has the
same relaxed, open, effortless sound as the original
playback. At one point, Don commented, “I can get
into a discussion about a specific thing or a specific
problem or approach to something and do that, but
the ‘audio bullshit session’ just kind of drives me up a
wall.” So I shut up and listened a bit closer. I heard an
awful lot of music coming out of Don’s system, both
with the 300B X3 and the monoblock 2A3 amps : we
listened to different styles of music on each
but I preferred the 2A3’s and Don concurred. He said,
“The 2A3 amps are the better amp, but you have to
compare them on a level playing field. These speakers
are efficient enough that it truly is a fair comparison,
but if you get, say, a pair of Spendors, and you play the
300B’s with them, they do this” - and he proceeded to
caricature the stance of a pumped-up muscle guy.
Don has enjoyed some attention with his recent new
product, a stereo 2A3 amplifier named the X. “I wanted
to build the cheapest good 2A3 amp that
I could. There were a lot of people that wanted to buy
one and couldn’t afford it.” He has done that and more.
The X features a chassis in the shape of an extruded
letter “x” standing on its end. It looks a little
like a tank trap with tubes and transformers peeking
out of various planes. I doubt I am the first to predict
that the Fi X will someday be in the permanent design
collection of the Museum of Modern Art. Its
construction literally turns amp design on its ear. The
X arrangement is more than just startlingly cool,
though : it provides physical isolation between power
supply and audio circuitry, permits excellent ventilation, and cuts the audio signal path shorter than a New
York minute. He has built on the novel X with two new
products : the Y line stage and the X3, a 300B stereo
power amp.
I asked Don how he thought up the X chassis concept,
and he said, “I’d been thinking of doing a cheap version of the 2A3 amp for a long time and I knew partswise what to do, how to keep the price way down. But
I didn’t want to put it on a little Bud-box chassis, because then it would look just like every other cheap
little amp. I tried this idea and that idea, with the idea
of keeping the power supply and the amp seperate. I
didn’t come up with anything that was any good at all.
I was riding the subway one day, when I was just . . .
there it was!” He smiles and his fingers make a little
“x” as the words trail off. “I quickly sketched it on
newspaper, came home, figured it
out, and everything just fell into
place. It worked perfectly. I’m not
sure how it happened, it was just
one of those inspirations that happens.” It looked like his amplifier
brainstorm went on to inspire a
painting : hanging on the wall was
a bichromatic canvas with cornerto-corner diagonal elements.
He was clearly proud of another
creation : an elegant WE 421Astereo
amp sporting only a dual triode
driver, the dual triode power tube,
and a rectifier nestled amidst big
iron. He took it down off a shelf to
show me. “This was the first one to
use the multiple levels of chassis
plates,” the signature style of the
earlier Fi amps. He showed me how
the design inherently shields and
physically isolates the audio path
from the power supply. Naturally, it
also looks very stylish. Here is a man whose products
embody modern industrial design, who paints in a
clean, abstract style, and yet spent about a half hour
lovingly displaying some of his antique tubes, including the bizzarre 316A, which looks like a tiny
weathervane inside the brain of Rosie, the Jetsons’ robot maid. Like his 3-tube stereo amp, Don Garber may
look simple but he works on many levels.
While Don has been refining his taste in circuitry for
decades now, he allows room for improvement : “Somebody said I got to the point where I kept taking things
out until it stopped working. And then put that part
back in, and that was it. Not quite, but almost. From
what I hear, simpler is usually better. Now you could
make that into a ‘rule,’ and go and carve it into a beam
up over your shop. But then tomorrow you might find
out that you’re wrong, so I don’t like to get into making rules. As soon as you make something into a rule,
you close doors.” Doing that may make your life simpler, but . . .” Don’s affection for simplicity is reflected
in his preference for direct-coupled circuits. When we
were discussing some of the tradeoffs in designing the
inexpensive X, Don pointed out that a direct-coupled
design is more forgiving of LF limitations in the output transformer : there is no rolloff from the coupling
cap to exacerbate any bottom-end phase shift in the
transformer.
While the original (2A3) X is direct-coupled, he was
clearly disappointed that he could not direct-couple
the new 300B version and still make it work in the X
chassis. The more complicated power supply
also would have driven the price of the X3 too high to
fit in with the conception of the X as an affordable
product. I wouldn’t be surprised if an all-directcoupled 300B amp comes out of Fi in the near future :
Don has a pretty clear idea in mind of how he wants
one built.
Also in his plans : a phono stage for the Y preamp, and
experimenting with the sound of the beefy KR 2A3
when run at a higher dissipation. Don’s curious,
though he’s not sure a move up to six watts will be
worthwhile. The remarkably musical 2A3 sound may
or may not show up using that operating point, he suspects, and also he’s not sure a six-watt amplifier compliments many more speakers than a three-watt amp
does. Garber is the easternmost member of the “Audio
Dream Team,” in charge of chassis and layout for the
upcoming hush-hush project soon to come out of Area
51 in the Mojave desert. Parallel-fed, active-loaded,
transformer-coupled, this X-wing fighter is going to
pack a fair amount of iron on whatever chassis Don
comes up with. I joked with him about a Kustom Kandy
Kolor for such a hot rod, and he scowled. (Don scowls
a lot.) “I like things to look like what they are. If it’s
metal, it should look like metal.” Later on, he pointed
out a small cosmetic flaw in a transformer prototype
he had rejected for other reasons. Clearly, this is a picky
guy with strong opinions about how his equipment
looks as well as how it sounds. I asked Don if he
thought the work he puts into his gear satisfies the
side of his soul that is also a fine artist. “It’s nothing
like painting, that’s a totally different thing. I’ve read
people talking about, building amps and so forth is an
art of its own. No. No, it’s craft.” He pointed towards
the speaker, from which played a legendary
live recording by another cranky master. “What’s coming out of that, that is the art.”
I hope Don Garber gets the satisfaction someday of
seeing his paintings hanging in a museum. But painting is a crowded, competitive field, and many who excel are never recognized. He may have to settle for seeing his X amp on a pedestal - it’s a shoo-in.
Re : the schematic. Don said, “I wish I had priced the X
at $892.00 instead of $895.00.” I asked why. “So it
would be exactly 1/ 100th of the price of the Ongaku.
W henever you hear, ‘cost-no-object,’ cost IS the
object”.
X
da’ basics
job is a done deal.
What do those stripes on the resistors mean?
Improving our soldering skills
by Doc B.
Here’s a little reparte I had with a novice Foreplay customer a while back after helping him get his Foreplay
up and running by resoldering all the joints:
I practiced soldering with my sca-35 and replaced
some its dried up wires and got that working. That
was my first solder experience. I guess I will solder
next time by heating up the end device first, say a lead
or terminal connection then try to get the solder on
top of the wire and then apply tip to the solder.
The trick is to get the tip of the iron in such a position
that it heats both the wire and the terminal it is being
attached to at the same time. This is critical, otherwise
the solder may only adhere to the one surface
that’s hottest. You should get both the wire and terminal hot enough that you will melt and flow the solder
when it is applied to the joint itself, not the tip of the
iron.
If this creates a non shiny solder then should I reheat
the connection until it hot again and see if that settles
it?
In general yes, although some of the newer types of
solder like the 97% tin /3% copper “lead free” solder I
use flashes to a dull finish when cool, even on
a perfect joint. The thing to look for is an even coating
of solder over both the wire and the terminal, and that
the solder has flowed enough to form a more or
less concave “fillet” instead of a convex “blob” or ball
at the joint. When reheating a joint it is a good practice to add just a bit more solder to the joint, which
helps stem the inevitable pull of gravity drawing the
remelted original solder down away from the joint. If
a joint is already gobbed up with a lot of solder, use
solder wick or a desoldering bulb to remove the excess
before resoldering the joint.
I think I got the basic idea behind it (not the dexterity
quite yet) but would appreciate any more soldering
tips. Since I want to jump to a 2a3 amp next I dont
want to keep on making those errors.
Part of the trick is to attach all the wires securely by
wrapping them around the terminal before you solder
the joint. This frees you from trying to hold
the wire in the terminal while soldering, leaving one
hand free to hold the iron, and one hand free to apply
the solder. Holding a wire in place while soldering
is not steady enough, the joint will inevitably crystallize while cooling due to tiny movements in the wire.
Make sure the solder gets hot enough to
freely flow over and through the joint. The whole joint,
all around the terminal, should be penetrated and
evenly coated with solder.
Using a hot enough iron is important too. Most of the
hotshit solderers I know have an assortment of irons.
Always included in this assortment is a big ugly 45W55W iron with a big chisel tip. If their fancy little temperature controlled model doesn’t heat a thick joint
enough, like say in the case of soldering speaker wire
to terminal posts, they bring out the big mutha and the
We get a certain number of calls complaining that
we don’t hand label each 1/4 watt resistor in our
kits. Newsflash, folks. They are already labeled. It’s
just a matter of knowing what those little stripes
mean.
The basic layout used involves four or five stripes.
On older style “precision” resistors there are usually four stripes, on newer metal films there are
often five to accomodate the hair splitting values
available in the 1% types.On either style of resistor
all but the last stripe denote number values according to the following plan:
black = 0
brown = 1
red = 2
orange = 3
yellow = 4
green = 5
blue = 6
purple = 7
grey = 8
white = 9
On the four stripe variety, the first two numbers
denote ordered values. On five stripers, the first
three stripes denote ordered values. For example if
the first two stripes on a four striper are red, purple,
we are looking at 27.
The next (second to last) stripe on either value denotes a “power of ten” multiplier. For example,
abrown (value 1) stripe denotes that you should
multiply the first two (or three) stripes by the factor 101, or ten. So a four stripe resistor with the
first three colors of colors red, purple, brown
would translate to 27 X 101 = 270 ohms. Red,
purple, yellow (value 4) would be 27 x 104 or
270,000 (a.k.a. 270K) ohms, etc.
To denote the 270 ohm value on a five striper, the
first four stripes would read, red, purple, black,
black, which translates to the first three stripes
denoting the number 270, and the fourth stripe
denoting a multiplication factor of 100, which is
of course, one.
What about the last stripe? It denotes the tolerance, or accepted plus and minus range of variance of the resistor’s value from that stated by the
stripes. Mostly these days we see gold, which denotes a 5% tolerance; red, which denotes a 2% tolerance; and brown, which denotes 1% tolerance.
There are other codes not so often seen anymore,
unless you are fooling with old carbon composition resistors - silver is 10% tolerance, no stripe is
20% tolerance, and even more obscure, yellow denotes 4% tolerance and orange denotes 3% tolerance.
So, you are now thoroughly confused and your
middle aged eyes can barely make out the stripe
colors anyway?
That’s why God invented the volt-ohm meter,
homes. Just measure the damned things with the
meter set on the “ohms” scale and forget this.
experiments with
a turntable
- part two
By Rene Bolders
a new platter
After the successes mentioned last issue I began to look
at the glass platter a little more critically, and I wondered if I could possibly improve here. At that time
acrylic platters became very popular in Germany and
elsewhere, and since acrylic or Perspex is very close to
vinyl materialwise, I felt that it could be a good match.
I chose Perspex as a suitable material because it has
high density but it is not too hard and is easily machined. I bought a block 14”x14”x3” and started by
cutting a rough round disc with a bandsaw. Since I’m
a machinist/fabricator by trade, it wasn’t very difficult for me to turn out a nice platter on the lathe.
The height is the same as the glass platter since I wanted
to do a direct comparison between the glass and
Perspex without changing the tonearm height. Well,
it was worth the trouble, the acrylic platter produces
an overall deeper bass. I feel that the bass and midbass
both increased, and the soundstage also seemed to be
deeper and more convincing.
I tried the two platters back and forth the following
days but the acrylic platter was always the clear winner. Weight-wise I didn’t gain or lose anything, important to preserve the bottom bearing load.
speed control
Next I began to look at the platter speed, because I had
noticed for some time that it wasn’t as constant as I’d
wished. The little motor is basically a 110 volt motor,
24 pole synchronous type, but it was wired for 230
volts, which means if you look at a stroboscopic disc
with a neon light you’ll notice that the little dots or
stripes do not stand still, but rather they kind of go
forward and backward all the time as if they were vibrating
The root problem is that the 24 pole synchro motor
gives 24 little jerks in one revolution. This is visible
with a good stroboscope. Om top of that the overall
speed varies quite a lot depending on the time of day
because there is not always 50Hz (or 60Hz in the US)
coming out of the wall outlet.
I measured and monitored the line voltage and frequency for a couple of days, only to see it fluctuate
from 218 volts in daytime to 238 volts at night. the
frequency varied from 48 Hz to 53 Hz. This is not a
very desirable situation, and in my opinion is audible.
My good friend Bernd (still lives in Berlin) came up
with a solution. He designed an external power supply
circuit board for 110/220V 50/60Hz operation. All I
had to do was build a nice enclosure, install a few plugs
and rewire the motor to 110V. See the photo showing
the power supply sitting next to the turntable. The
round thing on top is a record clamp.
The power supply generates a precise square wave
through a high frequency quartz generator and digital divider. Frequncy can be set for an exact 50Hz or
60Hz. The signal then goes through an active filter and
gets converted to a pure sine wave.
A nine watt class A amplifier amplifies the sine wave
and feeds it straight into the motor. Because the 110V
motor was wired to 220V through a voltage divider, I
gained some torque as I hooked it back to 110V. This
was to my advantage because of the relatively high mass
platter. The constant feed of pure 110V 50Hz into the
motor independent of mains fluctuations made the
motor turn very constant. When I measured the speed
with the stroboscope, to my satisfaction the stripes on
the disc were nearly standing still. Of course the motor is still a 24 pole synchro motor, but with a clean
signal feed it works great.
I would describe the advantage of an external power
supply to be better control in the bass region, a broader
soundstage, and a cleaner sound overall.
two steps forward, one step back
The next experiment went wrong.
I noticed that on Pierre Lerne turntables the surface of
the platter is not straight but concave. The reason for
this seems to be that the record has good contact with
the surface of the platter. The record clamp or weight
forces the record to the surface, so even if a record is
slightly warped it will still sit solidly on the platter and
not “wave” in the air.
Some records are concave themselves which results in
them siting on the platter either as a “little hill” or when
you turn them around as a frisbee up in the air. I know
one should try to avoid buying frisbees, we all like 180
gram audiophile virgon vinyl limited series handpicked
and so on. But if you like older stuff like me, jazz from
the 50’s and 60’s, and you finally find the one record
you were looking for for two years, you can’t be picky.
You have to find a way to make it work. Thus, following this spirit, I machined the platter slightly concave.
I adjusted the cartridge parallel to the platter’s surface
to have a good vertical azimuth. First I used a precision ruler to align the cartridge body to the platter and
later I checked with a test record (1kHz test tone) and
a digital voltmeter for fine tuning.
Now all this works great when you play frisbee’s from
the 70’s and 80’s, pop music and so on (Dynaflex
records work great). But it all starts going wrong when
you play LSC’s, Blue Notes, and especially 180 gram
pressings, because they don’t give as easily as you would
think. Plus, because of the differences in thickness the
vertical azimuth changes all the time, which results in
decreasing channel separation as well as lost balance
and soundstage width. On top of that, when it gets to
an extreme one channel plays louder than the other,
there is excessive groove wear on one channel and the
antiskating is affected.This is all very theoretical, but I
still don’t like it. On high quality turntables everything
matters and everything is audible.
I fought with these problems for a month or two, but
then I lost patience and resigned, which resulted in
my machining the platter straight and setting the cartridge back to where it was. A nice vacuum hold down
might do the job, but I haven’t gotten to this yet.
tonearm cable
As I mentioned earlier, my turntable came with a Grado
Signature Tonearm and Grado interconnects. The interconnects proved to be very good, but they are as old
as the turntable itself (15-20 years, who knows?).
Through time they may have become corroded or
something, anyway I told myself this as an excuse to
try something new.
I bought van den Hul phono interconnects with a
straight DIN-plug and liked them, but only until my
friend Bernd suggested trying a Straightwire Microlink
cable. I like this cable a lot, in fact I still use it today, it
is specially designed for low level analog signals and
doesn’t cost a fortune. I’m lured by some of today’s top
cables and I shall try to replace the whole cable (including the tonearm, and removing the DIN connection), probably with Kimber Silver, Cardas, or Discovery.
cartridge
The Grado Signature cartridge installed on the tonearm
when I bought the turntable had a bent cantilever. My
friend Bernd just happened to have a brand new Benz
MC-2 and since the price was right, it went straight
under the Grado headshell.
The Benz family of cartridges is somewhat of an interesting story. The Benz Company ( from Switzerland) is
a manufacturer of high quality cartridges. They design and manufacture cartridges under thier own name
as well as producing cartridges for other companies,
which sometimes leads to confusion. My cartridge is
called the MC-2 (Europe), however kit is called the
MC-3 in the US, and is identical or very close to the
Madrigal Carnegie 2, Taurus DDT, Empire MC1000,
van den Hul MC-10, MC-one, MC-two,
Empire MC-2000, and has lots of similarities with some Ortofon units. This is not to
say a bad word about Benz & Co., but rather
to take some of the hype out of cartridges.
The same goes for Scan-Tech of Japan, who
also make a whole range cartridges under
different names.
The Koetsu is very sensitive to all parameters. After
much trying and experimenting I ended up with a VTA
of 25 degrees, tracking force of 2.0 grams, 47Kohm
loading and a 100 pF cap. The difference between 1.8
grm and 2.0 gram tracking is very audible. It’s fun to
experiment with this Koetsu because when everything
is right it really sings.
tonearm
All this experimenting led to modification of the Grado
tonearm.
It’s Grado’s approach to not use only gravity or only
spring force to adjust tracking force, but rather both.
For 1-1/2 grams of tracking force, the setup uses 1
gram of gravitational load and 1/2 gram of spring
While my cartridge was at the factory I still wanted to
play records, so I checked out my friend Bernd’s collection to see what he had that I might like. As it happened a Koetsu Black Magic Gold had just arrived from
van den Hul where it was serviced and retipped. I purchased it on the spot and it really made me happy, in
fact I still use it today. There is something special about
Koetsus. I heard a Koetsu Rosewood on Bernd’s
Goldmund turntable, so I know what I’m talking
about). The Koetsu organizes the sound and puts everything in the middle where it belongs, but with a
wide and deep soundstage. (“Honey” is the term my
Aussie buds use to describe the K. sound - I’m in love
with my own Koetsu Black as well - B.)
I constructed a new counterweight which is solid
aluminumand steel by measuring the original weights
and trying to get as close as possible. Of course the
first attempt ended up way too heavy, so I had to machine the sides a little more until it worked. Adjusting
the tracking force is a little difficult now - if you move
the weight a little too far you’re immediately off 1015 grams! I didn’t change anything on the tonearm, so
I could always go back to the factory setup
But it was worth the trouble. The sound is
very airy now. Remember the tracking force
is still 2.0 grams, but the damping, other
than the friction in the bearings, is gone. I
like it better the way it is now, but some
reader will disagree and use a spoon wading through a basin of silicone oil!
Anyway, my Benz MC-2 worked great for a
year or so but then the sound slowly began
to degrade. At first I thought I had some bad
records, but I began to notice how the strong
bass and clear highs were dissappearing,
and the soundstage width and depth began
to dissappear. After some troubleshooting
we tracked the problem down to the cartridge. It became so bad that I had to send
the cartridge back to the factory. It turned
out to be a suspension related problem.
What happened was the little ”rubber” suspension ring had dried out and become hard
so that it limited the cantilever free travel. I
later heard that this happened to other cartridges too. The factory replaced the
“rubber”ring with one of some other material and all work fine now.
For me this cartridge is a typical MC cartridge - deepbass, laid back midrange and
good highs with a lot of detail. I used 2.1
grams tracking force and 24 degrees VTA
(rear of cartridge/tonearm slightly down).
the factory recommends a tracking force of
1.8 grams, but I didn’t like it so much, it
was somewhat thin sounding. The channel separation
is given as 28dB and it puts out 0.3 mV which means
that an excellent phono stage is a must.
the little weights off the tonearm? The idea was that
by constructing a new weight I could lower the center
of the counterweight compared to the turnaxis (center of gravity) and move the weight closer to the vertical axis. This would give a more stable tonearm with
respect to up and down movement (faster standstill)
so it would play slightly warped records more easily.
See drawing three for inspiration.
Shortly after I completed the experiment
with the counterweight I spotted an article
about a Roksan tonearm upgrade called the
Tabriz Tonearm with 2i counterweight. On
the Roksan Artemiz counterweight this idea
has been taken to the limit, the weight rests
only on a spike so it can swing freely without moving forward or backward and altering the tracking force. This works great
and I’m a little surprised not to see more
manufacturers picking up the idea. It’s
probably a matter of taste.
Right now I use all the above mods with great
satisfaction. I use the Welborne Labs Phono1 kit in connection with a German made
pre-preamp. I have about 63 dB gain
(Phono-1, 36dB, pre-pre, 27 dB) which is a
lot, but then again quite a lot of cartridges
put out something in the neighborhood of
0.3 mV, so this is about what is needed.
load. With a combination of a built-in spring and several weights this is a relatively easy adjustment to
achieve on a Grado tonearm. Once the tonarm is set
up it is a great tracker, but some recent developments
made me curious.
Some friends in the German audio community believe
that a tonearm should have no damping at all, but
rather should move as freely as possible in order to get
that last little bit of air. One friend uses a Grado tonearm in which he took the damping completely out no spring load, just the counterweight load, and he
reports good results.
The Grado tonearm has one big counterweight and two
small ones, so called micro-weights, for setting the
gravity to spring force ratio. This brought up an idea what if I take out the spring load, construct one weight
which is lower, actually below the cartridge, and take
The loading is 47kohms. I tried 100K, 30K,
10K, and even no loading in the past, but I
go back to 47Kohms every time. The capacitive loading is 100 pF. My system consists of a switching unit with stepped attenuator, two Welborne Laurel monoblocks with Cetron 300Bs and a full range
Edgarhorn system (see October 97 VALVE). I found it
unecesssary to use a preamp because all components
put out more than 2 volts, enough to drive the laurels/Edgarhorns to disco level. However, I installed a
buffer to compensate for the 20 ft. Kimber interconnects.
Maybe Ron Welborne was right when he mentioned
in his catalog “This is the last phono amp you will ever
need”. Maybe that and a pair of triode amps and good
horns is all we really need.
MagneQuest
Electronic Tonalities &
making the world a better place for DIY
Parallel Feed Output Transformers
EXO-36 2.5K:8 ohms, 3 watts max, permalloy core, solid brass channel frame. 2A3,
6A3, 6B4 - $270 pr.
Airgapped Single Ended Output Transformers
TFA-204 3K:16,8,4 ohms, 8 watts, 60 mA - $198 pr.
DS-025 2.5K:16,8,4 ohms, 20 watts, 60 mA - $298 pr.
EXO-35 2.5K:16 ohms version of above - $270 pr.
DS- 050 5K:16,8,4, ohms, 20 watts, 40 mA - $350 pr.
EXO-45 5K:8 ohms, 2 watts max, permalloy core, solid brass channel frame. 45,
71A, 417A/5842, 6CK4, 6DN7 - $270 pr.
EXO-46 5K:16 ohms version of above - $270 pr.
TFA-2004 3K:16,8,4 ohms, 15 watts max, black end bells, M4 core, 300B, VV32,
VV52 - $350 pr.
TFA-2004 Pinstripe ( M6 with Permalloy “pinstripes”), polished solid brass brass
bell ends, 300B , VV32, VV52 - $500 pr.
M
Q
RS-330 3K:16,8,4, ohms, 30 watts, 80 mA, - $500 pr.
RS-500 5K:16,8,4 ohms, 40 watts, 80 mA, - $600 pr.
RS-520 2.5K:16,8,4 ohms, 50 watts, 120 mA - $700 pr.
FS-100 10K:16, 8, 4 ohms, 40 watts, 80mA - regular $850 pr. - SPECIAL SALE,
$450 pair!
Push Pull Output Transformers
TFA-2004 all Permalloy core, 12 watts max, polished solid brass bell ends, 300B,
VV32 - $600 pr.
NEW - EXO-50 5K:16,8,4 ohms, 20 watts max, M6 with Permalloy “pinstripes”,
solid brass bell ends, 300B, 845, SV811A - $550 pr.
Parallel Feed Linestage Transformers
B7 15K:500 ohmsCT, parallel feed line stage transformer, permalloy core, brass
channel frame, 56,76,6J5, 6SN7,5965,5670, 12AV7 - $198 pr.
B7 5K: 500 ohmsCT, as above, 417A./5842, 437A/3A-167M, 5687 - $198 pr.
MQ-420-CL 6600CT:16,8 ohms 30 watts - regular $298 pr. - SPECIAL, $238 pair!
MQ-431 4300CT: 16,8 ohms 60 watts - regular $378 pr. - SPECIAL, $298 pair!
MQ-451-CL 2200CT:16,8,4, ohms, 120 watts - regular $458 pr. - SPECIAL, $398
pair!
MQ-470-CL 4300CT:16,8,4 ohms, 35 watts - regular $298 pr. - SPECIAL $238
pair!
Peerless 20-20 series model S-271-A, configure as 5000CT:16,8,4,1 ohms, 200 mA,
80 watts or 1250CT:16,8,4,1, ohms, 400 mA, 80 watts - $1200 pr.
Parallel Feed Interstage Transformer
RIT-5 5K:5K (1:1), M6, channel frame - $250 pr.
RIT-5 as above, but with bell ends - $300 pr.
add $100 per pair for Permalloy core
Parallel Feed Plate Loading Chokes
Peerless 20-20 series S-275-A, configure as 4000CT: 16,8,4,2 or 8000: 32,.16,8,4
or 12,000:48,24,12,6 ohms, 120 mA, 80 watts - $1200 pr.
Power Transformers
PGP 8.1 120V primary, secondaries 350-0-350 75 mA, 6.3V 2A, 2X 2.5V 1.25A,
horizontal (drop through) mount, solder terminals - $165 pr.
BCP-16 1mA, 425H, 4400oms DCR , channel frame - $80 pr.
BCP-14 10mA, 100H, 2075 ohms DCR, channel frame, - $90 pr.
EXO-01 20 mA, 100H, 985 ohms DCR, brass channel frame $110 pr.
5504 universal primary, secondaries 350-0-350 100 mA, 5V 3A, 6.3V .5A, 2X 2.5V
1.25A, black bell ends - $350 pr.
Filter Chokes
BCM-19 10H, 200 mA, 109 ohms DCR - $150 pr.
NEW EXO-99 dual bobbin, configure for10 mA 600H or 20mA 150H, 1129 ohms
DCR, brass channel frame - $198 pr.
NEW B.F.F.C. 10H, 200 mA, 18 ohms DCR!!!!!!!! - $298 pr.
BCP-15 50mA, 40H, 550 ohms DCR, channel frame - $90 pr.
EXO-03 60 mA, 30H, 344 ohms DCR, brass channel frame - $130 pr.
EXO-04 60 mA, 50H, 317 ohms DCR, black bell ends - $198 pr.
B.A.C. 80 mA, 50H, black bell ends, - $298 pr.
Electronic Tonalities
360-697-1936
www.bottlehead.com
VALVE
the magazine of astounding sound
Groove Thang Doc’s phono
preamp
O-Glow,
converting the
Baby Ongaku to
Active Loaded
Direct Coupled
operation
Brainiac on the
British ultrafi
scene and more
Heavy Metal
direct coupled
differential PP
EL84 amp
why can’t Johnny
read
schematics?
more
VSAC 98 photos
in color!
volume 6
number 2
1999
B-Glow
the latest buzz - a single ended 300B amp kit
from Electronic Tonalities
active loaded 5965 driver
MagneQuest TFA-204 air gapped output transformer
8 watts of sonic honey
$900 the pair, you just supply the 300 bees, er, 300Bs
VALVE
the magazine of
astounding sound
Editor and Publisher
Dan “Dr. Bottlehead” Schmalle
Chief Administrator
“Queen Eileen” Schmalle
Resident Smart Guy and Technical Editor
Paul “Braniac” Joppa
Graphic Design
Bruce “Badd Dawg” Borley
Resident Hot Iron and Dr. B’s Bodyguard
John “Smoothplate” Tucker
Big gun OEM advisor
Michael “Airgap” LaFevre
The Guy With Answers
John “Buddha” Camille
Contributing Editors
David “Full Track” Dintenfass
“Crazy Eric” Lenius
Our mailing address:
VALVE
P.O. Box 2786
Poulsbo, WA 98370
by phone: 360-697-1936
business hours: 9-5 PST, Mon -Fri
fax: 360-697-3348
e-mail - [email protected]
website - http://www.bottlehead.com
VALVE in no way assumes responsibility for
anyone harming themselves through exposure to
the contents of this magazine. We believe electrons
flow from minus to plus, and that they can kill you
along the way if you’re not careful. Vacuum tube
audio equipment operates at potentially lethal
voltages. Always treat it with respect.
Many ideas published in this magazine are untried,
and involve the use of potentially dangerous parts
and tools. In attempting any idea or project
published herein, you assume total responsibility for
your actions and any harm caused to yourself or
others. Please, be careful!
This publication is produced as a service to the
audio community and is wholly owned and
published by Electronic Tonalities.
The intent of this publication is to offer ideas to
inspire and educate audiophiles in an effort to
increase their understanding of the audio
equipment they use and cherish. Blatant copying of
the circuits published in this magazine for use in
commercial products shows a complete lack of
original thought.
editor’s
thing
Did we do the right thing?
We expected some bruhaha about our
changing over from hardcopy to what
seems to have affectionately been dubbed
CYBERVALVE by our readers, but in fact,
the positive response has been almost overwhelming.
Naturally we had a few folks come to us
with technical problems, but we seemed
to have sorted most of them out as being
due to the use of outdated versions of Acrobat Reader.
The folks who didn’t have computers were
urged to check out CYBERVALVE at the local library or cybercafe, and those who did
seemed pleasantly surprised.
A few folks complained about the download time. Ever hear the saying “anything
worth having is worth waiting for?” I don’t
plan on cutting the quality of this publication just to improve the download time,
so that’s just the way it is. A 20 minute
download is a lot shorter than three days
in the mail, homes.
The bottom line is this VALVE in hardcopy peaked at a total of
650 subscribers.
The first issue of CYBERVALVE, a bigger,
better issue than ever before, was accessed
over 7500 times in four weeks, from an
average of 1063 unique addresses per
week.
Did we do the right thing?
Can you say “no brainer?”
and another thing
We welcome new advertisers to this new
version of VALVE. Don’t let the fact that
this mag is coming from bottlehead.com
keep you from talking to us about advertising a competitive product. United we
stand and all that, we want to see everyone do well in this biz. You may note that
one of our biggest competitors (and a very
good friend at that) advertises in VALVE.
If you’d like to advertise in a magazine that
gets 7500 accesses a month, by folks who
are looking for kits, parts, and any tube
audio related stuff, get in touch with us.
We offer rates that are much lower than
the hardcopy mags - $100 per issue for a
full page (7.5”x10”) color ad, and can deliver a new ad to the public in a much
shorter time frame than the printed media.
You can send us the ad in Pagemaker, Publisher or Acrobat format, or send us your
photos and copy and we can put the ad
together for you and get it into the next
issue for an additional $40.
Our goal is to get a new issue out every six
weeks, so advertise early and advertise often!
Don’t let the blue smoke out,
Doc B.
on the cover
This is the prototype of our new B-Glow
300B SE amp kit.The conventional
airgapped, RC coupled, fixed bias circuit
has really turned out to be a pleasant surprise - great bottom end, absolutely killer
midrange, sweet clear highs. The prototype
was finished with gloss black powder coat
on the chassis
plate, a clear polymer coating on the
poplar base, and
metallic
faux
granite on the
i nductor s. Brass
hardware
replaced the stock
plated steel.
The wire is a combination of Jena
Labs
18
ga.
hookup and our
new house brand
20.5 ga. solid wire.
Groove
Thang
a phono stage
by Doc B.
“If your CD sounds better than your vinyl, there’s something seriously wrong
with your vinyl setup, B.”
Geez, was I getting tired of hearing that
one.
a little history
I’ve been futzing around with a few different pieces of phono gear over the years,
but never really reaching anything like satisfaction with the setups I’d use.
Been through the TD-124/SME3009/
Denon DL-103 route, actually that was my
second 124 (hell, man, I had a TD-124
when I was 18, with a big clunky Rek O
Kut arm and an original ADC-1 cartridge,
what took the rest of you so long to figure
out they’re a good Ttable?)
For now I’ve settled on a big ol’ Denon DP6000 I got as a commission on a fairly hefty
antique radio sale I did a few years ago. It
has a Denon DA-305 tonarm on it, nothing to write home about, save for the fact
that it is fairly massy. The table also has a
second arm mounting board, which has
been through a Grace 747 and now sports
a Linn Basik that’s been sleeved with
heatshrink and the counterweight has been
replaced with one of Andy Bartha’s littlest
Whatchamacallits with a hole poked in it
to add a little mass and damping. The
search for a decent tonearm continues, and
there’s a certain long skinny piece of ebony
that sits in my woodpile that subliminally
beckons me to try to DIY one whenever I
walk by.
The concern for mass stems from the fact
that I love low output moving coil cartridges, and them moving coils want to
hang off a beefy stick. They also want some
step up, and for this I have used a Denon
step up transformer, for lack of anything
else. I did have a Levison JC-1 for a while,
but somehow talked myself into the whole
“gotta have transformers” bit at one point
and sold it. To it’s credit the trans is very
versatile, it has switched inputs for two
cartridges and 40 ohm and 3 ohm taps.
But unfortunately I’ve heard Koetsu and
Sun MC step trannies, so I’m well aware
that this ain’t the best there is, although a
replacement of the stock output cable with
our new trick wire helped open things up
a wee bit.
Mikey keeps talking about these incredible step up trannies he wants to wind
someday, but he knows I’m too much of a
rag picker to be able to afford all that fancy
ass triple mu metal shielding, so we won’t
hear a MagneQuest MC stepup until someone with deeper pockets than Doc B. steps
up to the plate and custom orders a few
sets.
OK, I skipped over the cartridge, I know
that. Been through my fair share, over the
years, from that ADC-1, through Sonus
Blues, cheapo Empires and ATs, Grados of
various persuasions, cheapo to middling
good Shures, an Audioquest AQ-404 that
I bought in a fit of self pity when my first
wife dumped me, a DL-103 I scored for
$50 right before everyone started saying
how great they were, etc., etc.
No doubt in my mind that the low output
MCs always worked best for me. I note that
in the 50’s MC cartridges were called dynamic cartridges. I’ll drink to that. So now
I have a Koestu Black I picked up used from
Victor Meurisse, and a Fidelity Research
FR-1 that my bud Stan Webb found at
Goodwill, in the box, with the papers, for
something like 50 cents. The man has an
eye for the good deal, and I felt like I got a
great deal even when I offered him a
10000% profit for it
So now I have two carts with piddling output, about 0.3mV, and distinctly different
characters. The K is like honey on top, so
sweet you think it might be rolled off, until you hear a cymbal. It’s clean and clear,
but laid back. Bass? Holy shit! Put on some
70’s fonk with that really heavy analog
synth, and tear the roof off the mutha
sucka. The K is quite happy at 40 ohms, I
found it quicker but a bit dry into 47K.
The FR-1 is the other end of the spectrum,
more like what we’ve come to accept as
the “MC” sound, I guess. Very clean and
beautifully open, very present, to the point
of accentuating surface noise a bit too
much. Maybe not so hot on the bottom,
demanding 3 ohms to sound with any
weight at all. Fooling with the VTA helps
this somewhat, but my hunch is that this
is in part due to the Basik it’s mounted on,
and I hope at some point to try the FR-64
tonearm the cart was designed to mount
in, which I understand is the Hulk Hogan
of tonearms, with a bigass armtube and a
2 lb. weight hanging from it’s base, under
the turntable plinth.
a starting point
So enough with the preamble, you thought
this article was gonna give you a cookbook phono preamp and Doc is just reminiscing, right? Well, this ain’t that.
I was fortunate enough to have been given
one of George Wright’s first phono stage
prototypes (OK, I admit it, George has been
so kind as to send me three preamps over
the years, two of which have wandered
off to other guy’s systems, never to be
heard from again). How good are they?
Suffice it to say that it’s the one I recommend when anyone calls asking, and everyone who buys one loves it.
But hey, I’m Doc B., no matter how good
this thing is I just gotta build my own and
convince myself it’s better, ya know?
George based his design on the now classic RCA phono stage, a two stage preamp
with passive RIAA eq, which can be found
in the RC-30 tube manual. He modified
the design extensively, incorporating very
high gain 6ER5 tubes. He designed his
preamp to work with MM cartridges, and
to drive any kind of preamp you might
want to plug it into, including a cathode
follower output for the purpose.
Check out May 95 VALVE for the scoop.
I wanted to do a couple of thing differently, so I kinda cobbled up one of the early
protos George had given me. The earliest
experiment was to upgrade coupling caps
and wire, the usual “know nothing” type
of upgrades. Nice subtle changes, but this
was not what I was about, so the first thing
to happen was the cathode resistors on the
6ER5s went in the dumper and I shoe-
horned in four Ni-Cd AA batteries instead,
into a box even smaller than the production preamp come in.
George's preamp is setup for MM cartridges, like I said, and George has worked
hard to get a very extended top end (some
say a bit bright). This is a superb complement to Grados and Shures, but some of
the MCs I was using got pretty lean sounding through the preamp.
The AAs did a nice job of kicking up the
bass, not necessarily with a great deal of
control, but with ample quantity. I should
point out that in the process of adding the
batteries, the operating points remained the
same, with the 6ER5 grids biased to -1.2V
giving a low, like around 1 mA current
draw. This low current operation keeps the
6ER5s, with their rather high gain, very
quiet. One thing I wanted to mess with was
trying a higher current draw to see if I
could richen the sound up a bit without
adding too much noise.
I planned on running this preamp into a
Soul Sister line stage with 100K input impedance, through a short cable, so I didn't
feel the need for the cathode follower output stage George designs into his preamp.
Output now comes directly off the coupling cap that follows the second 6ER5
stage.
And of course I had to try to build the
preamp with C4S loads. These could turn
out to be the perfect load for a phono
preamp gain stage, they give two advantages we seek - high gain (the full mu of
the tube) and low noise, doing a superb
job of isolating power supply noise, and
replacing the typical noisy plate load resistor. Since I wanted to hear the 6ER5s
with higher current running through
them, I set the C4S atop each 6ER5 to run
at 10 mA and went back to resistors on
the cathodes. A word to the wise here, this
preamp is still in the cut and try stages.
Fiddle with the cathode bias resistor to get
the plate voltages you want.
So the basic two stage with passive RIAA
circuit came out as shown below.
I had been very pleasantly surprised at the
effect of using gaseous shunt regulators
with the Soul Sister line stage (really opens
up the top end), so I employed the same
ideas with the phono preamp, using the
original Wright prototype series regulated
power supply chassis as the starting point
The series reg and tube rectifier came out
and the VR tubes went into the existing
holes. It would be much better to put the
shunt regulator as close to the C4S boards
as possible, and I will probably move these
tubes to the holes where the cathode followers lived on the phono preamp chassis
in the future.
Ultrafast recovery soft start diodes replaced
the existing rectifiers in the high voltage
supply and heater supplies, and I reworked
the existing caps and resistors into a CLCR
filter ahead of the shunt regulator.
Groove Thang Phono Stage - one channel
300VDC B+
R1
100 ohms
100 ohm
C4S
active loads
2N2907A
2N2907A
MJE 350
R2
150K 1/2W
MJE 350
190VDC
150K ohms
190VDC
6ER5
6ER5
.47 mfd 400V
402K
1/4W
in
0.1 mfd 400V
47K
1/4W
the adventurous
will want to play
with this cartridge
loading resistor
536K
1/4W
1.2VDC
120 ohms
20K 1/4W
0.012 mfd
.003
mfd
passive RIAA
EQ
1.2
VDC
120 ohms
out to
line
stage
The whole PS is shown below:
And now a confession. I was hoping to get
low enough noise to be able to plug the
cartridges directly into the input of this
preamp, and get rid of the step up device.
I installed a 47 ohm input resistor and
cranked the volume pots on the Soul Sister waaay up. The sound was marvelous,
faster with more depth and detail, but just
a little too noisy and a bit shy gainwise.
Maybe some future improvements in layout will help, but right now the limiting
factor to direct input from low output cartridges seems to be finding 6ER5s with a
low enough noise level. I suspect it woud
be interesting to try this preamp with
nuvistors, too. I don’t have a clue whether
they would be any better,(they have a rep
for lots of microphony, so maybe not, but
it would be fun to try.
So back went a 47K resistor across the input, and in went the step up trans. Hmmm,
not bad! The new circuit seems to have
alleviated some of the vagaries I felt the
step up trans was producing.
medium Whatchamacallits. A few small
Whatchamacallits went around the
armboard. All of this extra mass seems to
tighten things up and offer a bit better isolation.
So you need a description of the sonics
now, huh. Well, I finally like my vinyl set
up better than my CD. It's a little fatter than
the CD setup, but the dynamics are like,
there, yowsa!!
But the biggest improvement was when I
pulled off the old rubber mat and put six
evenly spaced small Whatchamacallits on
the aluminum platter. The records now
"float" on these pucks. I had to add a piece
of 1/4" aluminum tubing to extend the
spindle to keep the center hole of the
record from slipping around and creating
wow, and of course I had to raise the tonearm height. Once this was done the improvement was, as we say, not subtle! The
reduction of background noise was quite
significant.
Bass is really something with the Koetsu, it
gets way down there, and with good control.
Living Stereos that seemed too soft and lush
now have impact and better, if not perfect, presence. Vocals are tighter with less
euphony. Mo' natchul.
I should mention my other recent tweaks.
Along with using our new long crystal copper cables I covered the Denon plinth with
big Whatchamacallits (well, one in each
corner really) and set the base on three
Give Andy Bartha a call at 954-583-7866
ET and get some of these ugly little pucks,
they are really fun to play with.!
Groove Thang Q&D power supply
(derived from Wright phono preamp supply)
adjust this resistor to drop
raw B+ to 300V at the VR
tubes. Use a curent draw of 70
mA to calculate R=V/I and
VxIx2 to calculate wattage
UF4007
fast recovery diodes
1A
600VCT or
higher
70 mA
120VAC
300V to C4S loads
8H
47
mfd
Schottky diodes
12.6V
1A
470
mfd
0D3
0D3
0.33
ohms
1W
2x22000 12VDC to
mfd series wired
heaters
For best performance put the power supply in a chassis separate from the phono preamp.
The VR tubes should be on the preamp chassis very near the C4S boards for best
regulation
New exotica from
First Impression Music
FIM CD019 Favorite Chinese Instrumentals
The Jin Ying Soloists
1 Like Wave Again The Sand (Pipa solo with ensemble) 4:36 2 Singing
the Night among Fishing Boats (Jung solo) 4:15 3 Night (Percussion
ensemble) 5:54 4 Love Song Of The Grassland (Tung-hsiao with
ensemble) 2:34 5 Ducks Quacking (Percussion ensemble) 4:00 6 Love
At The Fair (Er-hu solo with ensemble) 5:04 7 The Fishing Song (Bawoo solo with ensemble) 4:53 8 Happy Reunion (Xylophone solo with
ensemble) 2:55 9 Chinese Martial Art (Emsemble) 1:43 10 The Flowing Stream (Er-hu solo with Yang chin) 8:25 11 Spenpadei Folksong
(Yang-chin solo with ensemble) 4:32 12 Autumn Moon (Er-hu solo with
ensemble) 7:06 13 Moonlight Over The Spring River (ensemble) 9:22
14 Variation On Yang City Tune (Ku-jung solo) 5:36
FIM CD020 A Vocal Tribute to Ben Webster
Jacintha, Vocal
Teddy Edwards, tenor sax; Kei Akagi, piano; Darek Oles, bass; Larance
Marable, drums
1 Georgia On My Mind (Gorrell-Carmichael/BMI) 5:19 2 The Look Of
Love (Baccharach-David/ASCAP) 4:08 3 Danny Boy (Traditional) 7:26
4 Somewhere Over The Rainbow (Harburg-Allen/ASCAP) 9:43 5
Startdust (Parish-Carmichael/ASCAP) 6:40 6 In The Wee Small Hours
Of Morning (Hillard-Mann/ASCAP) 4:15 7 Tenderly (Lawrence-Gross/
ASCAP) 5:25 8 Our Love Is Here To Stay (George & Ira Gershwin/
ASCAP) 3:19 9 How Long Has This Been Going On? (George & Ira
Gershwin/ASCAP) 5:28 10 Pennies From Heaven (Burke-Johnston/
ASCAP) 3:12
#189 16149 Redmond Way
Redmond, WA 98052
http://www.fimpression.com
Tel: 425-868-5326
Fax: 425-836-9061
you
asked
for it
O-Glow
a conversion of the “Baby Ongaku”
into a big brother version of Afterglow
By Doc B.
We’ve received several calls lately from
folks who’ve built Gordon Rankin’s great
“Baby Ongaku” design, so named in an
article by Frank Reps in Issue 9 of Sound
Practices. The design was originally drawn
up by Gordon for Mike Lafevre to use as a
giveaway application for MagneQuest customers who purchased DS-025 output
transformers and 5504 power transformers. A lot of folks have called us asking if it
would be feasible to convert the circuitry
of the Baby-O, as it is affectionately called,
to roughly the same direct coupled, active
loaded topology as the Afterglow.
Sure. The basic parts are there, and the
quality of the iron should yield up a sort
of “superglow”. NOTE that I don’t have a
Baby-O on hand as I write this, so it may
take a bit of tweaking to get the voltages
of this direct coupled conversion just the
way you want them. Such are the trials of
the hardcore direct coupler.
To take the design as close as possible to
the Afterglow in terms of sonics the 5V4
should be replaced with a pair of ultrafast
recovery soft start rectifier diodes. This will
lower the power supply impedance which
can help dynamics. It will also help to get
the voltages up where we want them to
be, as the voltage drop across the solid state
rectifiers is less than the tube rectifier.
The .68 mfd cap at the input of the power
supply filter must be increased, maybe to
40-100 mfd to get the B+ up to what we
will need for direct coupling of the circuit, a critical contributor of the Afterglow
sound. With the power transformer secondary voltage of 720VCT we should get
into the ballpark of voltage we get from
the PGP 8.1 trans used in the Afterglow.
In a pinch that 10 mfd cap connected to
the 12AT7 SRPP could be used to replace
the .68 cap if it has a high enough voltage
rating, otherwise use a 600V or higher
rated capacitor and go ahead and up the
value to 100 mfd.
While you’re at the output of the PS filter,
disconnect the 390K and 100K voltage divider resistors and the 22uf bypass cap that
connect to the CT of the 6.3V filament
winding, you won’t need ‘em for the new
driver stage. The center tap can be left
floating, or you can tie it to ground. Choose
the way that is quietest.
The plate of our 2A3 will be sitting at
roughly 400-440 VDC, and we will want
the 2A3 filament to float at about 145VDC.
The 2A3 will run at 60 mA, so we will
need bias of -45 volts between the filament and the grid. This puts our direct
coupled 2A3 grid and 5965 plate both at
100VDC. To set the filament at 145VDC
at 60 mA, we will need to change our
1Kohm 2A3 cathode resistor to 3000 ohms
(yes this is the right value, you’ll see why
when we explain the connection of our
5965 active load) and it will need to be a
10 watt wirewound resistor.
may be paralleled to it to achieve the 50
mfd value, although you may be sacrificing a bit of bass performance compared
to the stock Afterglow.
We will use 1/2 of a 5965 for our active
loaded driver, because a 12AT7 will become so sensitive with the C4S as to make
the amp’s volume a bit hair trigger to adjust. Using 1/2 of the 5965 is kind of cool,
if one triode wears out, we can always wire
up the other half!
The SRPP is replaced by a C4S active load
on top of the 5965. In the process of removing the SRPP the 7.5K ohm dropping
resistor, the 10 uF bypass cap, and the 333
ohm bias resistor associated with the top
half of the SRPP are removed. We want to
run our 5965 at 100V and 5 mA, so R1 on
the C4S board should be about 190 ohms,
to set our 5965's current draw to 5 mA.
R2 needs to be about 73Kohms. See the C4S
manual for details on setting this up. The
cathode bias resistor of the 5965 needs to
be about 280 ohms. No cathode bypass cap
is needed.
The input grid resistor of the 5965 can
stay at 221Kohms. Obviously the 221K
grid resistor on the 2A3 grid, and the .47
mfd coupling cap are removed for this direct coupled amp and the plate of the 5965
is coupled directly to the 2A3’s grid.
NOTE once again, beware that these voltages may be off a bit from what you actually get. The point here is to get 45V difference between the 2A3 filament and the
grid. If the plate voltage is a little lower or
higher, don’t sweat it. If it’s a lot different
(hey we don’t have a Baby-O here to test),
let us know and we can clarify the values
needed.
Another unusual feature borrowed from
the Afterglow design is the connection of
the C4S board’s input, the R1 current set
resistor to the top of the 2A3’s cathode resistor. This seems an odd way to supply
the driver tube, but in practice works extremely well, the most notable improvement being a reduction of the noise floor
when compared with the more conventional approach of feeding the load thru a
dropping resistor directly from the output
of the power supply. This connection also
explains why the 2A3’s cathode resistor is
3000 ohms. Only 50 mA passes thru the
cathode resistor, while another 7 mA
passes thru the C4S, for a total of 57 mA
total current draw through the 2A3.
A 200 mfd capacitor rated for at least 200V
should be used as the cathode bypass cap,
replacing the 30 mfd cap spec’d in the
original design. If the original 30 mfd cap
is rated for 200V or better, the 20 mfd @
250V cap removed with the voltage divider
So there you have it, now you can try out
another nice 2A3 circuit on that Baby-O
you built. And if you are looking to build
a super 2A3 amp on your own chassis, the
design would certainly be a worthwhile
contender for a scratchbuilt project.
o-glow
a little
journey
to the
home of
Kevin
Scott
By Paul Joppa
When I was in England in 1997, I tried to
visit Simon Shilton, the transformer maker.
That didn’t work out, but he recommended
that I try to see Kevin Scott and hear his
speakers. I’d never heard of him, his company, Living Voice (#2 Carlyle Road, West
Bridgeford, Nottingham NG2 7NQ United
Kingdom), or the Air Partner and Air Scout
speakers. But it seemed like a worthwhile
thing to try, as long as we were going to
Nottingham anyhow.
We stayed at a nice pub that must have
been built about the time trains were invented. It was close to the train station
anyhow, and a short walk to the castle.
There are several pubs near the castle, built
into the old tunnels under the castle, that
all claim to be the oldest pubs in England.
Everything seemed easily accessible, and
West Bridgeford was also pretty close - but
Kevin suggested strongly that a cab would
be the best way to find his place. He was
right. I had a map, and was diligently following the cab’s route - for about the first
half mile! After that, it was hopelessly confusing for an outsider like me. Eventually
however, we pulled up in front of his house
on a pleasant residential street and I
knocked on his door. I walked into his fairly
small living room, and was stunned by
these huge monoliths, in piano black and
beautifully finished birds-eye maple. They
stood nearly 4 ft high with a nearly 2 ft
square footprint. Then I noticed the bass
horn between them, maybe 30 in high but
at least 5 ft wide and 2-3 feet deep. The
speakers completely filled the narrow end
of the room. OK, I’m thinking, the satellites are 16 cubic feet each this guy must
be pretty serious!
He immediately put some music on, before I can say more than “Hello, I’m Paul...”
- a real believer in the VALVE philosophy,
Shut Up and Listen! The sound is simply
stunning - see the comments later. The system that we listened to included a Wadia
860 CD, Simon Shilton’s lovely preamp
(basic “anode follower”, choke-input
power supply - see his web page for details), and an Art Audio amplifier with
VV30B outputs driving Sowter output
transformers for 13 watts.
The bass he feels is far more important
than you would think - any shortcomings
here will muck up the midrange. It’s surprising, he says, how often a change in the
bass results in a percieved change at much
higher frequencies. It’s also the biggest
challenge; it’s very difficult to obtain clean
high-level bass. I must say, his system certainly supported this, showing the best
combination of extension, resolution, and
impact that I’ve ever heard from a speaker
system.
In between the cuts, we talked about various issues. Kevin has lots of opinions, and
wasn’t shy about sharing them. Most seem
to be based on his own experiences, and
he often deplored the incestuous repeating of other’s opinions which characterizes much of the high end dialog today.
Perhaps I’d better describe the speakers we
were listening to. These are his smaller
systems; there is also a larger model without the subwoofer. All his systems are made
as much as possible by other craftsmen;
he noted it took a long time to find a good
enough cabinet maker but it was worth
the effort. He’s right, the cabinets are real
works of art. The drivers are all from
Vitavox, except the tweeter ( JBL), but most
are modified in various ways to his specs.
The systems are way out of my league at
some $30,000 a pair plus another $12,500
for the woofer - I felt lucky to have been
able to spend so much time listening to
them.
He runs the business as much from his
home as from his (by-appointment-only)
shop. He seems to spend his time listening
to music and talking about music systems
with other enthusiasts, with occasional
breaks for beer and mountain biking. What
a great life!
On amplifiers, he thinks most single-ended
amps have inadequate power supplies.
There’s a guy named Gary Dews who
makes amps and power supplies under the
Border Patrol name (Simon Shilton also
mentioned him) who has experimented
extensively, and in fact sells upgrade power
supplies for many SE amps. Kevin claims
these power supplies make major improvements in sound, even on cheap amps. He
also thinks there’s no point in bothering
with 211s, 845s, etc - they just never have
the full-range resolution of the smaller
audio tubes. In general, he has found that
high-frequency extension is surprisingly
important - hence for example interstage
transformers are to be avoided unless you
can afford a really good one.
Kevin is very enthusiastic about component quality. I gathered he had spend considerable effort investigating components
for every detail of these speakers. He likes
Hovland capacitors, finding most others
dull (and even the best electrolytics dark
sounding). He uses big, air-core inductors
- and at the 90Hz subwoofer crossover,
they must be pretty darn big!
Starting at the bottom, the sub has two 12"
Vitavox woofers (a custom design for this
speaker) in a classic folded horn like the
original Shearer theater design. It covers
40-90Hz. I did not find out anything about
the crossover, except that it was passive,
everything running off the Art Audio amp.
The woofer of the “satellites” is another
12" Vitavox, the same one as the subs Kevin feels this helps the integration. The
suspension is quite stiff, and I assume the
cone is quite light to obtain the claimed
105dB sensitivity. The surround is the
tried-and-true doped folded paper; the
cone itself is undoped paper. The magnet
is quite large, and ceramic - he actually
prefers the sound of these drivers with
ceramic magnets to that with Alnico, after
trying both. It is front-radiating and backloaded with a large horn, exiting with a
mouth about 18" square at floor level. Visible internal curves are smooth. The horn
is said to go down to 50 or 60 Hz in-room
without the sub; with the sub it covers 90
to probably around 200, then direct radiation takes over up to the crossover at
500 Hz - much like a classic Lowther horn
in the bass, but with dimensions twice as
big, and a shorter horn due to the higher
cutoff.
The midrange crossover is basically a classic second-order type, with one trick that
he was a little cagey about except to say it
was quite cheap but makes a significant
difference. For the most part, effort was
spent on component selection rather than
crossover topology elaboration.
The midrange is a Vitavox horn and driver
with a 2.75 in custom diaphragm - very
similar to the classic large Altec units. The
diaphragm surround is a polymer with
radial corrugations. The horn is interesting - it looks a lot like a large Altec (it’s a
full 2 ft wide, and I suppose 10" high), but
the dividers go almost all the way back to
the throat - and the side walls are curved
in the half closest to the throat, so that it’s
more exponential than radial in the first
half of the horn. It becomes radial for better directivity control in the mouth half.
Specified directivity is 120 by 90 degrees.
Brainiac’s
Heavy
Metal
...an occasional review of some nifty
chunk of iron (usually cheap!)
by Paul Joppa
Today’s Topic:
the “Champ” output transformer
The dividers ring a bit when tapped, but
the horn walls are very well damped.
The mids run to 10 kHz with a natural
rolloff; the last octave is handled by a JBL
slot tweeter which he prefers by a wide
margin over the much-touted Audax.
Unusally, it’s rotated so the slot is 45 degrees from the vertical.
This is not a review, just some impressions
based on a few hours listening in Kevin’s
room, to his music in his system - but this
puppy sounds darn good. The system was
being played at levels considerably higher
than I would use myself, and I was impressed with the effortless ease with which
these levels were handled. I have never
heard bass before with such convincing
solidity. Not even the big Wilsons can do
what a really good bass horn can do. The
sound was wonderfully well integrated. I
have admired John Tucker’s Exemplars in
this respect ever since I first heard them,
as being far and away better than any other
horn system I’ve heard - but this was if
anything better. And the treble nasties,
which I hear in every horn system I’ve ever
run across to some extent, seem to be entirely missing. (The treble itself was certainly not missing - cymbals for example
were especially crisp and lifelike). I spent
some 3 hours with Kevin, probably 2/3
of it listening to this system at levels I would
not normally tolerate, and I was still not
tired of it at the end. I don’t know what it
is, but there’s some real magic here. Only
twice did I hear any hint of coloration (and
we went through a wide variety of music)
- a bit of “shout” on male voice, and some
low violin once. I suspect these were horn
resonances, but in any case they are rare
exceptions to the generally uncolored
sound. There was no chestiness, no boom,
no tizz, no high-level treble crunch (even
with massed strings at 10 dB above concert-hall levels). Percussive transients were
especially well-handled. Unusually in my
experience, these systems were very listenable fairly close in - even 5 ft from the
speakers the sound was still well integrated, and backing up as far as I could,
some 15-20 feet away made it worse rather
than better.
The “Champ amp” is a small (5-watt) guitar amplifier, usually with a 6V6 output
tube and an 8-inch speaker of 4 or 8 ohms.
This transformer is sold as a replacement
for burned-out units in these amps. I got
mne from Angela Instruments for $25; they
claim it is a little heavier duty than the
usual replacement unit. It weighs 15 oz
on my kitchen scale; the lamination stack
is 2.25" tall x 1.875: wide x 0.75" thick. It
appears to be air-gapped, secondary outside, primary inside. On my test bench I
measured the following values:
slope. With tweeters that have an inductive impedance rise, the treble should be
even better.
Turns ratio:
26.5
Primary:235 ohm DC
Secondary:
0.36 ohm
Leakage ca. 0.07 H
Capacitance:
ca 34 pF
DC current:
Inductance with 12v/
60Hz excitation is 10.5H at 10 mA, falling
to 5.25H at 90 mA. I would rate it at 45
mA, where inductance is 8.7H.
With a 4 ohm load, impedance is 3300
ohms and losses are 0.7dB. It should give
3.3 watts at 60 Hz. . Small signal frequency
response with a 1.5 to 2k plate resistance
is within 3dB from 21 Hz to 10 kHz.
This unit is unusual in that there is one
lamination on each side that is interleaved,
giving an unrealistic 15H at zero current.
The capacitance is surprisingly small, so
that high-frequencies are limited only by
the leakage inductance giving a gentle
The primary impedance with an 8 ohm
load is 6100 ohms, and losses are small at
0.4 dB. Figuring power bandwidth as the
frequency where the inductive impedance
equals the load resistance, I rate this unit
for 6 watts at 110 Hz. Small signal frequency response with a 1.5 to 2k plate resistance is within 3dB from 25 Hz to 17
kHz.
Overall, while this is not the greatest transformer available, it is good enough to hear
the music. I’ve used it in my original “spud”
amplifier with 417A/5842s; it’s also suitable for many TV vertical output triodes
like the 6DN7, 6CK4, or 6EM7. A 45
would be perfect, though at today’s prices
they are beyond the budget area this transformer represents. And of course it was
designed for 6V6 or 6BQ5 service.
Out wit h t he Old
We are closing out the KR Enterprise VV Valves to make room for the new generation of KR tubes.
All sale tubes below have the KR 1 year warranty from date of purchase. Visit our website for KR
pictures, specs and curves. http://www.welbornelabs.com/krhome.htm
Please Note: These VV Valves all have 2A filament current requirements. Please check your equipment specification before ordering and make sure they will operate in your amps.
No returns on these sale items.
VV300B...the following are all electrically the same tube, just different glass shapes and color. The
cylinder tubes are just that...they have a cylindrical shape with a flat top.
VV300B regular transparent glass $200 per pair
VV300BL
regular blue glass
$200 per pair
VV302BL
regular blue glass
$200 per pair
VV302BC
blue cylinder
$200 per pair
VV32B...these VV32Bs are all electrically the same tube except for glass color and shape.
VV32B
VV32BL
VV32BC
regular transparent glass all sold
regular blue glass
$250 per pair
blue cylinder
$250 per pair
VV52B...these VV52Bs are all electrically the same tube except for glass color.
VV52B
VV52BL
regular transparent glass $250 per pair Hurry, almost gone
blue glass
$250 per pair Hurry, almost gone
Hurry, quantities are limited.
All sales are final.
Buy 2 pair... take a 5% discount.
Buy 3 pair... take a 10% discount.
In with the New
New From KR Enterprise. All tubes feature a 2 year warranty.
KR10 Dual Triode Pre-amp and Driver Tube.
A 6SN7 on Steroids? $395.00/pair
KR2A3 Single-Plate Vacuum Tube
A great new single plate 2A3. 2.5V/2.5A filaments. 2-6 watts of glorious sound. $395.00/
pair
KR PX25 Vacuum Tube
A replica from the past. 4.0V/2.0A filaments. 8 watts. $435.00/pair
KR300B Vacuum Tube
An awesome 300B replacement. 7 to 14 watts, 5.0V/1.2A filaments. $435.00/pair
KR300BXLS Vacuum Tube
8 to 20 watts of power and dynamics. 5.0V/1.2A filaments. $575.00 per pair
Welborne Labs
ph: 303.470.6585 fax: 303.791.5783
http://www.welbornelabs.com
e-mail: [email protected]
direct coupled differential driven El 84 amp
by
Feeling burnt out on single ended amps?
Try this on for size - how about direct
coupled differential driven EL 84 amp.
That does not cost a million dollars to
make. The amp featured in this article is
one that I built for Mike Lafevre, using a
design that Mr. Komoro developed. I was
given permission to build this amp with
his schematic, also to write a future article.
I would not recommend this amp to beginners as it is not an easy amp to construct. Most single ended amps are a lot
easier to construct than this amplifier. How
would I rate the difficulty? Intermediate,
at least a couple of amps under your belt
or a couple of belts under your amp? For
resonace dampening of course!
To DHT or to not DHT that is the question? This amp was designed not to compete with the harmonically rich DHT PP
amps that are more expensive to build.
These amps are built to a price point. The
greatest cost incurred is the cost of the
Dynaco SCA 35.
Down and out need an audio fix, cannot
find one of these amps to strip? Mikey to
the rescue, Magnequest is reproducing the
Dynaco Z 565 transformer that is used in
both the SCA 35 and ST35 amplifiers. The
transformers are of better quality or equal,
of course people will whine that they are
not original. But they will not whine when
they sound better.
this EL 84 amp is like Sauvignon Blanc,
slightly sweet with a clean finish. “I said
sound, you clown”. Sorry, writing an article will drive you to drink. The amp has
a very powerful, sweet and transparent
sound - can you say holographic? I knew
you could.
As for the power tranny good luck,
Hammonds suck. Be sure to dig deep if you
go to Leeds Electronics, they might have a
surplus transformer that will work.
Electroprint will build custom power
trannies to order.
Where‘s the beef? Solid state rectification,
it’s true. I swear. Ask Herb, he will tell you.
How about the ultimate version? Tube
rectified, carbon resisters, monblocks,
separate cathode bias resistor and
Blackgate caps - these things are as smooth
silk. Do not forget, too smooth, too boring! It’s a tonal balancing act, every part is
a natural tone control. Every part has a
different resonance. The trick is to get every thing to resonate in natural harmony
how will you know its right? You will feel
it in your heart.
How does it sound? One could say that
Lets talk techno, geek! The cathodes of
the 6DJ8 are feed a negative voltage, the
voltage and current are set by the cathode
resistor. Why ? This enables the tubes to
operate with equal bias on each section.
Here lies the kicker, this amp is direct
coupled any changes in the driver tube will
effect the output tube, so the filaments of
the 6DJ8’s are regulated. This ensures that
the thermal emission of the heaters, are kept constant. I used a
Velleman kit regulator board that
is cheap enough to offset the cost
of my labor that would be required to hard wire one. Powering the regulator I used a Radio
Shack 12 volt transformer.
Regulating the B+ would have
taken up more real-estate then I
had available. Also the sound that
B+ regulation inflicts upon the
signal is not my sound.
A balance pot is used for each
input tube to adjust the balance
between each section. In order to
monitor the balance of sections
of the input tubes I connected a
battery powered LCD meter. I
would then adjust the pot until 0.00 was
displayed on the meter. A drift up to 1.00
is acceptable, this much imbalance will not
harm the tubes. The output tubes are biased with a cathode resistor on each channel. These resistors are mounted on the
chassis for better thermal transfer.
This amp is one of my favorite amps that
I have built. We shall see how it fairs
against the highly touted Acro 20/20 that
I am restoring.
The EL 84 tube is one that should be taken
seriously.
da’ basics
Schematics look like Greek to you?
by Doc B.
A lot of new folks have been joining our
ranks lately, which is awesome. While a
few come from an electronics background,
the majority of new bottleheads are die
hard audio nuts who know what they like,
but haven’t had the opportunity to get up
to speed on the design end of the DIY game.
While we go to great pains to make our
kits buildable without any working knowledge of circuitry, anyone who is serious
about this stuff and wants to continue to
improve their system will eventually need
to know how to read a schematic drawing
of a circuit.
To this end I will start to present here some
basic symbols seen in typical tube audio
circuits. Please don’t bust my chops if you
think I forgot something, there will be
more of this in the future, and count on
the fact that we’ll cover what each component symbolized here really does in a
circuit over the next few months.
Here we go:
Resistor - The squiggle
implies that the electrons
flowing through the resistor are forced to burn
off a little energy as they fight their way
through.
Potentiometer - Now our
squiggly line has an arrow
pointing to it. The arrow represents the wiper that runs
over the resistive track in the
pot when you crank it up to
scare the neighbors. Can also indicate any
other form of adjustable resistor
Capacitor - The two perpendicular lines with the
gap between them represent the two charged
surfaces (usually foil or a metallized coating) in the capacitor. The insulator,
polypropylene, oil or whatever, would be
in the gap. You will see here that one of
the two perpendicular lines is curved.
Sometimes both lines are straight. This
curved line sits to the side with the lowest
potential of a DC rated capacitor, the “minus” terminal.
Inductor- The curly
line represents the
inductor coil and
the straight lines represent the core material. One curly
line makes a
choke,
two
curly lines, one
on each side of
the straight
lines, makes a
t r a n s f o r m e r.
Taps are signified by straight lines attached
to the curly lines.
Ground - a ground connection
is represented in a lot of different ways, but the most common
is either the three lines shown
here or a triangle pointing down
(to the ground, get it?)
heated rectifier A directly heated tube is a
tube in which the electrons boil directly
off the filament. The V can also be called
the heater in an indirectly heated tube. An
indirectly heated tube has an additional element called the cathode which is heated
by the heater to the point where it boils
off electrons. The indirectly heated cathode is the sort of upside down L shaped
thingy right above the heater (the filament
in a DHT is often called the cathode too).
Triode - Like the
vacuum diode, the
circle defines the entire tube, so a dual triode is often diagrammed with only
half the circle enclosing the internals. The
new element added to our basic vacuum
diode is dashed line in the middle which
represents the grid.
Tetrode - This is like
the triode, but now
there is an additional
dotted line above the
grid. This is the
screen.
Diode (Solid State) The triangular arrowhead dealy bob represents the direction of current flow through
the rectifier, and the line represents the
cathode end, through which current is
kept from flowing back in the reverse direction.
Light Emitting Diode ( LED)
- These are similar to the
solid state diode, with the
addition of little arrows to
represent the light emission.
Diode (Vacuum) Here’s a little more
c o m p l e x
heiroglyphic. The
circle defines the entire tube. The upside
down T at the top is
the plate of the diode. If there is one of
these in a tube designated as a rectifier
(which turns AC power into DC power) it
is a half wave rectifier. If there are two of
these in the circle, the tube is a full wave
rectifier. The upside down V at the bottom is called the filament in a directly
Pentode - This is like
the tetrode but yet
another dotted line is
added, which represents the suppresor.
cold cathode.
Voltage Regulator Tube - the
gaseous shunt regulator tube
is designated with the
familar upside down T for
the plate, and a circle for the
Switch - gotta turn the
thing off (unless you just
unplug stuff like Doc ).
The angled line represents the moving switch contact.
Fuse - the curvy line represents the part inside
the glass that goes poof.
Jack - the outer ring represents the
part of the jack that usually goes
to common or ground, the inner
circle the signal carrier. Now they
can’t say you don’t know jack.
more VSAC 98 pics
Now that we can give you color photos, I dug back
through all the nice photos we received from VSAC 98
attendees.
Tom Vetromile sent us the nice pic above, of the
awesome Tube Research Labs amps and phono preamp
built for Winston Ma. These babies greeted everyone in
the show office as they signed in.
Dick Olsher supplied the photo of a happy Ed Billeci
and his Sato horn. Ed came away with a nice sampler
of Audience caps as an award for his horn and the
250TL amp shown in 5/98 VALVE
Dorothy Harwood of
Acoustical Magic and Betty
Kalmus of Antique Sound
USA proved that audio is no
longer just a man’s game.
Photo by Tom V.
Ron Welborne introduced the new
VV52BX powered Apollo amp kit,
replete with beautiful chassis work,
teflon sockets and a cool Japanese
meter
Photo By LynnO
You have no idea what Doc goes
through to get you a good price
on MagneQuest iron
Photo by LynnO
Dan Meinwald and Alan
Kafton set up a radical
looking system in the EAR
USA/Samahdi/Sahuaro
room.Yes, those are speaker
cables behind the Ichiban
speakers and V-20 amp
Photo by Tom V.
So what are Eduardo de Lima,
Lynn Olson, Gary Dahl, and the
rest of the gang looking at?
Gary’s Ariels
in the Craftsman Room,
of course.
Photos by Hank Murrow
These pics just barely scratch the surface of what went on at VSAC 98.Check out Lynn Olson’s Ariel website, Dick and Leslie
Olsher’s Black Dahlia website, and David Robinson’s coverage of VSAC 98 in Positive Feedback, Volume 7 Number 6 for some
other perspectives.
So when’s the next one? The superb quality of the exhibits and the major enthusiasm of the attendees has inspired us to make
VSAC even bigger and better in the future. To give us a enough time to do it right, we’ve decided the next VSAC will be VSAC2000.
We’ll put up info about where and when at bottlehead.com as soon as we have a venue and a date.
By the way, in case you didn’t guess, putting on a show of this type is a gigantic undertaking. If you would like to volunteer to help
put on VSAC2000, give us a call. We’d love the assistance, and you’d get to spend even more time involved in things tube audio.
Entwined
interconnect cable kits from Electronic Tonalities
pre-cut twisted pair 20.5 guage High purity continuous cast long crystal solid copper
conductors with a high temperature insulative coating, pre-cut polyethylene outer jacket
and Vampire Wire RCA plugs
$49 per 1 meter pair $10 each additional meter
VALVE
the magazine of astounding sound
hardcore action - Buddhafied Afterglow
Deep Throat - a Lowther front horn
C4Sing the Decware Zen
da’ basics - V=IR, it’s not just a good idea,
it's the law!
mo'’ Heavy Metal
volume 6
number 3
1999
B-Glow
single ended 300B amp kit
another masterpiece from Electronic Tonalities
active loaded 5965 driver
MagneQuest TFA-204 air gapped output transformer
8 watts
$900 the pair, you just supply the 300 Bs
VALVE
the magazine of
astounding sound
Editor and Publisher
Dan "Dr. Bottlehead" Schmalle
Chief Administrator
“"Queen Eileen" Schmalle
Resident Smart Guy and Technical Editor
Paul "Braniac" Joppa
Graphic Design
Bruce "Badd Dawg" Borley
Resident Hot Iron and Dr. B’s Bodyguard
John "Smoothplate" Tucker
Big gun OEM advisor
Michael "Airgap" LaFevre
The Guy W ith Answers
John "Buddha" Camille
Contributing Editors
David "Full Track" Dintenfass
"Crazy Eric" Lenius
Our mailing address:
VALVE
P.O. Box 2786
Poulsbo, WA 98370
by phone: 360-697-1936
business hours: 9-5 PST, Mon -Fri
fax: 360-697-3348
e-mail - [email protected]
website - http://www.bottlehead.com
VALVE in no way assumes responsibility for
anyone harming themselves through exposure to
the contents of this magazine. W e believe electrons
flow from minus to plus, and that they can kill you
along the way if you’re not careful. Vacuum tube
audio equipment operates at potentially lethal
voltages. Always treat it with respect.
Many ideas published in this magazine are untried,
and involve the use of potentially dangerous parts
and tools. In attempting any idea or project
published herein, you assume total responsibility for
your actions and any harm caused to yourself or
others. Please, be careful!
This publication is produced as a service to the
audio community and is wholly owned and
published by Electronic Tonalities.
The intent of this publication is to offer ideas to
inspire and educate audiophiles in an effort to
increase their understanding of the audio
equipment they use and cherish. Blatant copying of
the circuits published in this magazine for use in
commercial products shows a complete lack of
original thought.
editor's
thing
Yo, bottleheadsDiscussion came up at the last meeting that
while the new surge in local VALVE membership has been bringing in lots of folks
who are eager to learn the basics of design, there seems to be a general swing towards requests for "cookbook" circuits in
VALVE by the worldwide membership.
(definition: if you read VALVE, you're a
VALVE member, we are quite democratic)
How can I put this diplomaticallyAin't gonna happen, no way, no how.
For those of you who have recently discovered the pleasures of tube audio, and
are finding a desire to construct your own
gear, maybe partly for economy's sake
(Amen, bro!), maybe partly for the satisfaction of telling others "I made it myself!",
we aren't going to let you off that easy.
See, VALVE was formed by a bunch of guys
who didn't just roll tubes, or hook up a
new piece of used gear every week. The
guys who started VALVE learned early on
to dig up the forgotten nuggets of information in arcane texts, apply that knowledge in their own experiments, and share
that knowledge - not just the end result,
but the method behind the madness as well.
In fact the most hardcore DIY guys I know
are usually more proud of their tube electronics book collection than they are of
their projects!
What we intend to present in VALVE is not
just a canned circuit full of exact Digi-Key
numbers and a preprinted circuit board,
but rather a general design and just enough
information to make you work a bit to get
the project finished.
This may seem a bit overwhelming the first
time you work through a project published
here, and that's part of the reason we’ve
started "da' basics" column, to help you
through the math and other basic knowledge you will use again and again in designing tube gear. You'll usually find the
math to help you through the tough spots,
and once you finish a project or two, our
hope is really that you'll go off on your
own, devise some devilishly clever new circuit, and share it with the readership
(complete with color photos, of course).
Which brings me to another important
pointWe need article submissions like yesterday.
We need not only your new circuit designs,
but also your photos of finished ET kits
and other projects, stories of your experiences learning the craft of tube audio, and
your pieces on the theory of various aspects of design. You needn't be intimidated
by the fanatical attention to detail of Buddha or the way math rolls off the pen of
Brainiac. Some of our most popular articles are those like the original old S.E.X.
speaker piece, where we slapped a couple
of drivers on a cardboard box! And in fact
I can't think of two guys to whom the sharing of knowledge is more important. Believe me, they will appreciate your contributions more than anyone, no matter what
your skill level at the time of writing.
So fire up the iron and the pen,
but- don't let the blue smoke out,
Doc B.
on the cover
Mike Connly sent us these photos of his
S.E.X. amps. Mike not only built the amp
kits and made the custom walnut and
maple bases for the amps, he also made
the chessboard upon which they rest. Mike
was the first bottlehead to build a Blues
Master, and rumor has it he will be putting the completed amp in a new chassis
soon. Another cover girl, maybe?
New exotica from
First Impression Music
FIM CD019 Favorite Chinese Instrumentals
The Jin Ying Soloists
1 Like Wave Again The Sand (Pipa solo with ensemble) 4:36 2 Singing
the Night among Fishing Boats (Jung solo) 4:15 3 Night (Percussion
ensemble) 5:54 4 Love Song Of The Grassland (Tung-hsiao with
ensemble) 2:34 5 Ducks Quacking (Percussion ensemble) 4:00 6 Love
At The Fair (Er-hu solo with ensemble) 5:04 7 The Fishing Song (Bawoo solo with ensemble) 4:53 8 Happy Reunion (Xylophone solo with
ensemble) 2:55 9 Chinese Martial Art (Emsemble) 1:43 10 The Flowing Stream (Er-hu solo with Yang chin) 8:25 11 Spenpadei Folksong
(Yang-chin solo with ensemble) 4:32 12 Autumn Moon (Er-hu solo with
ensemble) 7:06 13 Moonlight Over The Spring River (ensemble) 9:22
14 Variation On Yang City Tune (Ku-jung solo) 5:36
FIM CD020 A Vocal Tribute to Ben Webster
Jacintha, Vocal
Teddy Edwards, tenor sax; Kei Akagi, piano; Darek Oles, bass; Larance
Marable, drums
1 Georgia On My Mind (Gorrell-Carmichael/BMI) 5:19 2 The Look Of
Love (Baccharach-David/ASCAP) 4:08 3 Danny Boy (Traditional) 7:26
4 Somewhere Over The Rainbow (Harburg-Allen/ASCAP) 9:43 5
Startdust (Parish-Carmichael/ASCAP) 6:40 6 In The Wee Small Hours
Of Morning (Hillard-Mann/ASCAP) 4:15 7 Tenderly (Lawrence-Gross/
ASCAP) 5:25 8 Our Love Is Here To Stay (George & Ira Gershwin/
ASCAP) 3:19 9 How Long Has This Been Going On? (George & Ira
Gershwin/ASCAP) 5:28 10 Pennies From Heaven (Burke-Johnston/
ASCAP) 3:12
#189 16149 Redmond Way
Redmond, WA 98052
http://www.fimpression.com
Tel: 425-868-5326
Fax: 425-836-9061
Buddhafied
Afterglow
Hi Dan,
Talked to Ken Dangerfield and he has been busy on the Internet. He told some
group about the beefed up design I gave him for a two chassis layout for the
Afterglow. (we had this info posted on the old website -B.)
He has been getting numerous questions about the power supply, switching, noise
suppression and CMCs. He came up with the idea of putting the mod on your
internet site, so I cleaned up the drawings so you might scan them into the
gadget. Evidently you are assembling a compendium of A-glow mods - pretty
neat.
Included a short theory-of-op’s to save you from a bunch of questions. Hope you
do not find all of this too presumptuous.
Regards,
John C.
Can you believe this guy? Too presumptuous? Zowie, this article has some incredible gems in it! Study these pages hard, learn from Buddha, and be a better
builder for it!
Doc B.
the Deep Throat Horn
a front horn for Lowther drivers
by Bent Audio
John Chapman of Bent Audio sent us these
wild horns a few months back, and Doc
just finally got around to setting them up.
The tractrix expansion contour horns are
molded from ABS and have a theoretical
cut off frequency of 140Hz.
The horn mouth is about 34" in diameter,
and the horn is 18" deep, hence the name
Deep Throat. The throat opening is 7-1/
2", just right for fitting snugly against the
frame of a Lowther driver.
The driver mounts to a piece of PVC pipe
which comes cut to just the right length to
make a nice mounting flange, as seen in
the picture to the right. John suggests attaching the flange with 5 minute epoxy
putty, which worked nicely. The putty is
rolled into a snake of a length approximately equal to the throat circumference
and then pressed in the tapered gap between the outside of the horn throat and
the PVC flange.
Once the epoxy dried holes were drilled
for 3/16" hanger bolts and the drivers
were slipped on and secured with brass
nuts and washers.
The backs of the drivers were left open for
the first audition, and fiberfill was piled
over the backs to damp the rear wave. A
quick and dirty frequency sweep by ear
indicated that the horns went down pretty
smoothly to 200hz, where they begin to
roll off.
Anyone who has listened to Lowthers
knows they have a 'presence peak' which
is generally centered around 2-2.5kHz.
I have measured
my PM2As as having more of a plateau from 2 kHz
clear up to 10 kHz,
so I was curious to
see if these horns
would fill in the
midrange and upper bass.
Yup! The peak is
still there a bit, but
the front loading
does wonders for
the balance without losing any of
the speed and life
that the Lowthers
are known for. The
dynamics are stunning, I have a
drum solo that will
blow your mind on
these guys. I estimate overall sensitivity at something
a r o u n d
101dB@1W@1M.
I have been using a
Whamo subwoofer temporarily, but we'll
develop a woof that's a better match both
in terms of bandwidth and sensitivity in
the future. These babies may not be
‘plug'n'play', but they are really worth trying if you're a hardcore Lowther tweak.
The Deep Throat Horns are $520 the pair,
plus shipping. Contact John Chapman at
Bent Audio, [email protected] or
604-533-6684 for more info.
A similar horn called the Oris 150 has been
developed by Bert Doppenberg. More info
about it can be had at http://home.wxs.nl/
~lowther/final.htm
The Deep Throat Horns, at work in Doc's Chamber of Horrors
and Kit Packing Dungeon
Out wit h t he Old
We are closing out the KR Enterprise VV Valves to make room for the new generation of KR tubes.
All sale tubes below have the KR 1 year warranty from date of purchase. Visit our website for KR
pictures, specs and curves. http://www.welbornelabs.com/krhome.htm
Please Note: These VV Valves all have 2A filament current requirements. Please check your equipment specification before ordering and make sure they will operate in your amps.
No returns on these sale items.
VV300B...the following are all electrically the same tube, just different glass shapes and color. The
cylinder tubes are just that...they have a cylindrical shape with a flat top.
VV300B regular transparent glass $200 per pair
VV300BL
regular blue glass
$200 per pair
VV302BL
regular blue glass
$200 per pair
VV302BC
blue cylinder
$200 per pair
VV32B...these VV32Bs are all electrically the same tube except for glass color and shape.
VV32B
VV32BL
VV32BC
regular transparent glass
regular blue glass
blue cylinder
all sold
all sold
$250 per pair
VV52B...these VV52Bs are all electrically the same tube except for glass color.
VV52B
VV52BL
regular transparent glass
blue glass
all sold
Hurry, quantities are limited.
All sales are final.
all sold
Buy 2 pair... take a 5% discount.
Buy 3 pair... take a 10% discount.
In with the New
New From KR Enterprise. All tubes feature a 2 year warranty.
KR10 Dual Triode Pre-amp and Driver Tube.
A 6SN7 on Steroids? $395.00/pair
KR2A3 Single-Plate Vacuum Tube
A great new single plate 2A3. 2.5V/2.0A filaments. 2-6 watts of glorious sound. $395.00/
pair
KR PX25 Vacuum Tube
A replica from the past. 4.0V/2.0A filaments. 8 watts. $435.00/pair
KR300B Vacuum Tube
An awesome 300B replacement. 7 to 14 watts, 5.0V/1.2A filaments. $435.00/pair
KR300BXLS Vacuum Tube
8 to 20 watts of power and dynamics. 5.0V/1.2A filaments. $575.00 per pair
Welborne Labs
ph: 303.470.6585 fax: 303.791.5783
http://www.welbornelabs.com
e-mail: [email protected]
New KR2A3
you
asked
for it
C4Sing
the
Decware
Zen
By Doc B.
We've hade several calls from folks who’ve
been using the Decware SE84B Zen amp,
a single ended SV83 amp manufactured
by Steve Deckert (www.decware.com).
A few of the callers have listened to the
Zen amp and have asked us for a way to
make the Zen sound a bit more dynamic.
I point out here that the folks who have
called liked the Zen amp, this article is not
intended in any way to be a flame job of
the stock amp. This mod is offered strictly
in the spirit of experimentalism that makes
VALVE what it is.
The choice of tubes is cool - The driver is
Svetlana's 6N1P, sister to the 6922, but
with a bigger cathode and vastly reduced
microphony. In fact we like the tube so well
we are using it in a new prototype parafeed
line stage here at Rancho Tonalities. The
SV83 output tube is a sister to the 6BQ5,
which any triode guy will tell you is the
best sounding pentode made.
the voltage swing requirements for the
6N1P are minimal.
The circuit is as simple as possible, the only
adjustment being a choice of cathode resistance on the straightforward voltage
amp driver stage. The output stage is operated in triode mode with a 1K resistor
tying the screen to the plate.
To hit our numbers means replacing the
2.7K and 1.5K resistors with a single 833
ohm (820 is a common value that would
be close enough) 1/4W resistor on each
cathode.
The simplicity of the circuit makes the C4S
installation easy. The 6N1P is set up to run
at a plate voltage of 156VDC with a cathode bias of 4.3VDC and a cathode resistance of 2.7K, or a plate voltage of about
104VDC, cathode bias of 2.3VDC and a
cathode resistance of 964 ohms.
The first operating point gives us a plate
current of 1.6 mA, and the second operating point gives us a plate current of
2.4mA.
Version 1
Looking at the curves published by Svetlana
for this tube, these are really kind of low
operating points. I'll do a conversion sticking with a low 2.4mA current first, and
using the curves I'll shoot for an op point
of about 110VDC on the plate and 2VDC
of cathode bias.
The SV83 grid is biased only 10V or so, so
Our C4S's R1, the current set resistor, will
need to be .95V/0.0024A, or 396 ohms
1/4W (390 is close enough).
To set the bias resistors we will need to go
back upstream in the circuit and figure out
what our voltage should be at the top of
the C4S loads and how to make it so.
At point b in the power supply schemo we
know we have a pretty stable 260 VDC or
just a few volts less available.
We will have a total draw of 6.8mA from
our new circuit, 4.8 mA total thru the two
plates and 2 mA total thru the bias diodes
of the two C4Ss.
The 15K and 10K resistor in parallel between points b and c equal a 6K resistor.
The drop across this puppy at 6.8 mA
would be 22.8V, so let's say we would have
around 256 - 22.8= 233VDC available at
point c. The actual value may vary a bit
but this is close enough for jazz.
two C4S loads with
R1 = 390ohms
R2 = 220K ohms
each cathode
resistor
is 820 ohms
to points “d”
with C4S active loads
version 1
two C4S loads with
R1 = 95 ohms
R2 = 120K ohms
each cathode
resistor
is 200 ohms
to points “d”
with C4S active loads
version 2
replace these two
resistors with a single
1K ohm 2W
wirewound resistor for
version2
C4S
loads
replace 15K
resistors
with wire
jumpers
Now we calculate our bias resistors R2 on
our C4Ss as 233V/.001mA = 233K ohms.
A 220K 1/2watt resistor will do. Substitute wire jumpers for the 15K resistors between points c and d. The C4S loads replace the 47K plate loading resistors between points d and e, obviously...
Version 2
For a higher current version that will have
the 6N1P running at a point we like in
other circuits we use, let's shoot for 10 mA
at 200V as an operating point for our
6N1P. The SV83s draw a maximum of 66
mA total, so the total draw of two SV83s
and the two halves of the 6N1P running
with C4Ss will be666 + 20 + 4 = 90 mA,
well within the maximum 150 mA current rating of the Hammond 372-FX
power trans.
R1, the current set resistor on the C4S
board, will need to be .95V/.01A = 95
ohms.
10V of drive for the grid of the SV83.
The cathode bias on the 6N1P halves will
need to be 2V, so we need a cathode resistor on each triode of 2V/.01 mA = 200
ohms.
We can now compute the value for R2.
With 236V available at the top of the load
and a 2mA current through the LEDs we
need a 236V/.002A = 118K resistor rated
for 1 watt. 120K is plenty close.
Let's bump the LED's current draw up to
2 mA. We already figured our current
draw by the 6N1P halves and C4Ss to be
24 mA. The voltage drop across that 10K
and 15K resistor from point b to point c
would now be 144V, way too much!!
No sweat, let's just change the 10K and 15K
resistors in our previously modified power
supply (modified by the jumpers replacing the 15K resistors at points c and d) to
a single 1K 2W WW resistor. This would
give us 24V of drop, and our voltage at
the top of the C4S loads should be about
260-24=236V. This is just enough plate
to source compliance for our C4S load to
work, and remember, we only need about
What else can you do??
Parafeed it, Baybee
Disconnect the secondaries of the existing
output transformers, and cap couple a suitable PA matching type transformer, ala
ParaS.E.X., to the plates of the SV83. Try a
2-3mfd coupling cap.
The plate resistance of the SV83 may be a
bit high with respect to the inductance of
the Zen output-trannie-now-plate choke,
so you may not get much sub 30 Hz bass,
but there's only one way to find out, and it
would be a cheap experiment..... B.
da' basics
V/I=R
Ohm’s law the most important formula you’ll need
Substituting 43 Volts for V and .060A
(that’s the same as 60 mA, get it?) for I we
get
If you can remember this one, you can figure about 80% of the math out in a given
circuit design. Ready?
V=IR
where V= Voltage
I= Current
R= Resistance
How about some applications, you say?
OK, here’s one:
You need to calculate the value of a cathode resistor for a 2A3 self biased at 43 Volts
and drawing 60 mA of current.
We use our junior high algebra and rearrange Ohm's law to read
Brainiac's
Heavy
Metal
...an occasional review of some nifty
chunk of iron (usually cheap!)
by Paul Joppa
Today’s Topic:
the De Young Mfg. S.E.X. kit
“universal” power transformer
P= 43V x .06A
which equals 2.58 Watts.
43V/.06A=R
which equals about 717 ohms.
Now a nice feature of Ohm's Law is that
we can relate power to it with the following formula
P=VI
where P = power and the other variables
keep the same definitons.
With this information we can calculate
what kind of power rating our 2A3 cathode resistor needs. We know V = 43 Volts
and I = .06A, so we can find P by substituting
This is the transformer supplied with the
S.E.X. kit for international orders. It differs from the U.S. version in having more
primary windings so it can be wired for
several power line voltages, it is rated for
50 or 60 Hz operation, and it has less B+
current capability than the Magnequest
unit offered for the domestic market. It
weighs about 3.5 pounds on my kitchen
scale, and the lamination stack is 3” x 2.5”
x 1.5”. The laminations are moderate,
about 0.01”, and they are interleaved one
by one. It is mounted horizontally with an
upper metal end bell, and takes mounting
holes 2” x 2.5”. All the wires come out on
the bottom.
The primary can be wired for 100, 120,
or 240 volts and measures 7.1 ohms DC
wired for 120 volts. There are four secondaries. The ratings and my measurements are as follows:
Rated voltage
2 x 2.5v
6.3vCT
350-0-350v
Rated current
2 x 1.25A
2A
60mADC
Open circuit voltage
2.90 + 2.90v
6.67v
366-0-366v
DC resistance
0.12 + 0.12 ohms
0.17 ohms
404 ohms CT
By the way a good rule of thumb is that
you want the power rating of a resistor to
be at least twice ( the conservative say four
times) the power radiated by the resistor.
In this case you would want to use a 5 watt
resistor at the very least, and a 10 watt
resistor would be safer.
There may be cases where you are working on a design and don't have one or the
other of V or I through a given component, but you know the value of the resistor. You can still calculate the power rating.
Since we know that V=IR, and hence
I=V/R and we can substitute for I in our
power formula:
P=VI=VxV/R=V2/R
or P=IRxI=I2R
Cool, huh.
Estimated voltage at rated current
2.66 + 2.66v
6.13 v
343-0-343 v*
* the high voltage winding was estimated
on the basis of 60mA RMS through the
whole secondary, which is a reasonable
approximation for capacitor-input circuits
but is not exact. Leakage inductance was
not evaluated but may reduce the B+ output voltage slightly.
Equivalent source resistance in the B+ is
about 267 ohms, or about 4% of the likely
load resistance. With a capacitor-input filter and using silicon diodes, a maximum
of 445 volts could be generated at 60mA.
Remember however to allow for the drop
in the DC resistance of any filter choke
used, plus the loss in the rectifier tube if
used. With a choke-input filter an output
of 305 volts at 90mA could probably be
obtained without excess heating, again
minus the losses in chokes and rectifiers.
Unusual in smaller power transformers,
the regulation is pretty good. You can
probably draw more B+ current if the filament windings are not fully utilized, but
no more than half the extra volt-amp capability can be used this way without excess heating.
jc morrison smashes
record for the worlds
largest S.E.X. speaker...
This is a photo of one of a pair of speakers constructed by jc morrison and demoed
at nyNoise, a very cool show put on in
Hoboken, NJ by jc and pal Blackie Pagano,
which we attended on March 13. The
speaker is a 9' tall open baffle with 18 of
the drivers that Doc worked with MCM
Electronics on a couple years back (PN 551870). The tweeter is a Raven (sounded
very nice!). There was also a TAD based
subwoofer with this system. The overall
sensitivity was quoted as 100dB, and it
sounded like it.
You long time bottleheads will note a basic resemblance to the open baffle S.E.X.
speaker born from the unstable minds of
Doc and Brainiac at the first VALVE Xmas
party after several bottles of Champagne.
If you want to take a project like this on,
we might recommend that you try what is
often called the "original S.E.X." drivers instead (PN 55-1290), With their high Q
they are quite a bit more open baffle
friendly, going lower on an open baffle
(about 110 Hz vs. 200Hz for the drivers
shown here) and a bit more efficient to
boot. See our new Whamo webpages for
lots more info.
Thanks to jc and Blackie for great hospitality and a fun show. We hope to twist
their arms and get them to come to the
next VSAC. Hey jc, will those things ship?
now that looks dangerous
All you gotta do to get bottleheads to look
at the camera is say "beer".
Here's Smoothplate, John Hoffman, Brad
Brooks (he's the one guy who’s working),
Tom Vetromile, Richard Riley and Brainiac
in his disguise, hard at work on the new
VALVE listening room. Those joists that
Tucker's almost hitting his head on will
come out, and if you look close you will
see the new ones he put in at 9-1/2 feet.
The room will finish out to about 19' x
11'.
We're planning on installing two layers of
1/2" sheetrock, and there's four hospital
grade sockets on the front wall and one
side wall, in case we audition speakers that
like long walls, like Ariels.
Oh yeah, there'll be a small wine bar in
the back corner too. Brainiac smiled very
wide when I announced this...
Doc,
This is Joe from Australia. Thanks for the
Foreplay Kit with the upgrades. I have received the package in December last year.
Unfortunately, I was away in Malaysia for
a couple months. So, I was able to finish
the kit just a few weeks ago. Anyway, I have
managed to build another deluxe version
of the kit using better parts like the
Hovland musicap, Holco, Kiwame resistors.
I have used VSE Allen Wright's voltage
regulator together with the power supply
of the Foreplay to provide the B+. Heater
is also DC regulated as well. I have put the
whole package in a Hammond Aluminium
box with the valves sticking out proudly
(see pic).
I have managed to have a power transformer custom wound for operating in
240V country. I went to Octave Electronics (www.members.xoom.com/octave/) in
Malaysia and they did it for me at a reasonable price (works out to be about US$27
/ VA). So if someone from 240V territory
needs the service, I would recommend
them.
Anyway, thanks again.
Regards,
Joe Ling
Entwined
interconnect cable kits from Electronic Tonalities
pre-cut twisted pair 20.5 guage High purity continuous cast long crystal solid copper
conductors with a high temperature insulative coating, pre-cut polyethylene outer jacket
and Vampire Wire RCA plugs
$49 per 1 meter pair $10 each additional meter
VALVE
the magazine of astounding sound
Hawaii 5.0 two ways to
play a RS 5"”
driver
mo'’heavy
metal
300B taste
test
new Foreplay
techniques
Buddhafied
Afterglow part two
some ideas
for SE 845
parafeed
da'’basics determining
triode plate
loads
volume 6
number 4
1999
B-Glow
single ended 300B amp kit
another masterpiece from Electronic Tonalities
active loaded 5965 driver
MagneQuest TFA-204 air gapped output transformer
8 watts
$900 the pair, you just supply the 300 Bs
editor's
thing
VALVE
the magazine of
astounding sound
Editor and Publisher
Dan “Dr. Bottlehead” Schmalle
Chief Administrator
“Queen Eileen” Schmalle
Resident Smart Guy and Technical Editor
Paul “Braniac” Joppa
Graphic Design
Bruce “Badd Dawg” Borley
Resident Hot Iron and Dr. B’s Bodyguard
John “Smoothplate” Tucker
Big gun OEM advisor
Michael “Airgap” LaFevre
The Guy With Answers
John “Buddha” Camille
Contributing Editors
David “Full Track” Dintenfass
“Crazy Eric” Lenius
Our mailing address:
VALVE
P.O. Box 2786
Poulsbo, WA 98370
by phone: 360-697-1936
business hours: 9-5 PST, Mon -Fri
fax: 360-697-3348
e-mail - [email protected]
website - http://www.bottlehead.com
VALVE in no way assumes responsibility for
anyone harming themselves through exposure to the
contents of this magazine. We believe electrons flow
from minus to plus, and that they can kill you along
the way if you’re not careful. Vacuum tube audio
equipment operates at potentially lethal voltages.
Always treat it with respect.
Many ideas published in this magazine are untried,
and involve the use of potentially dangerous parts
and tools. In attempting any idea or project
published herein, you assume total responsibility
for your actions and any harm caused to yourself or
others. Please, be careful!
This publication is produced as a service to the audio
community and is wholly owned and published by
Electronic Tonalities.
The intent of this publication is to offer ideas to
inspire and educate audiophiles in an effort to
increase their understanding of the audio equipment
they use and cherish. Blatant copying of the circuits
published in this magazine for use in commercial
products shows a complete lack of original thought.
Yo, bottleheads,
This issue I’d like to present Brad Brooks
as our guest editor. Brad has been a local
VALVE member this past year, and has really impressed us with his enthusiasm and
commitment to the hobby. Below is Brad’s
response to an inquiry by a newbie about
learning the fine art of design, as posted to
the Bottlehead ForumDon’t let the blue smoke out,
D o c B.
The following was posted by Scott Wobecke
to the Bottlehead Forum:
How long does it usually take a
bonehead in math/construction
skills to build tube amps? I have
a lot of ambition...Which books
do you recommend reading, I’ve
come across some recommendations at the Audio Asylum. I
figured if there is anyone to ask
its the master of DIY tube gear.
Thanks Doc.
To which Brad replies:
Here is my two cents. I agree with Quest
that you should jump in and build something. Here is what I recommend, and
please keep in mind that I do not consider
myself a designer. I would like to think that
someday I would be able to sit down with
an idea, and find the right solution based
on the combination of other people’s experience, my own experience, and research
and experimentation.
Here are some things that I have done to
help me along this path.
1. Read, Read, Read—all you can about all
aspects of stereo, not just tube, design. Decide for yourself what is funk, and what is
bunk. Understand that alot of what is written out there is designed to sell products.
Decide what your dream system consists of,
and prepare to have your ideas blown apart.
Develop enough character to deal with
blows to your ego.
2. Decide what sort of music you really like,
and what you really like about it. Calibrate
your ear with live performances. Learn to
love music again, unless you only like talk
radio, and even then, learn to love the human voice.
3. Listen to as many different systems, tube
and transistor, as you can.
4. Build something simple, like Electronic
Tonalities kits, especially since there is
enough information there to help you understand the operation of the circuit—in
fact, pretty much all of the ET kits are made
with the idea that you will continue to refine, redesign, and experiment. Learn by
doing.
5. Listen to what other competent experimenters and designers say, and listen to how
their systems sound.
6. Identify small milestones of accomplishment, experimentation, design. Small
things, like how to determine the best operating points of a tube, or to find what
design factors are affect bass performance.
Experiment.
7. Continue to evaluate everything learned,
experimented with, etc...throw out useless
information and disproved theories.
8. Associate yourself with experimenters—
people who actually like music and build
the equipment that they listen to it on. Find
a club of experimenters and music lovers.
9. Have fun, and ignore the bigots and jerks.
10. RCA Radiotron Designer’s Manual, 4th
ed., 1950’s era Radio Amateur’s Handbooks, RCA Receiving Tube Manuals,
Valve Online Magazine (Free!), Positive
Feedback, Vacuum Tube Valley, Glass Audio.
Sorry for rambling on...
Brad
on the cover
Pictured this issue is the Foreplay preamp
I decided to build for myself a while back,
to try out some mod ideas I had.
This preamp is now doing a tour of the staff
at Audio Asylum, in our opinion the best
audio forum on the web. In fact we think it
is so good that we have decided to be a sponsor, and we are proud to announce a new
Bottlehead Forum on the Asylum, in conjunction with this sponsorship. The forum
will be a more timely way to get info out
and offer answers to your DIY questions.
Check it out at
http://www.audioasylum.com/forums/
bottlehead/bbs.html
new
classics from
First Impression Music
FIM XRCD 022, The Artistry of Linda Rosenthal
Linda Rosenthal, violin, Lisa Bergman, piano
A collection of timeless violin solo favorites, by the most
talented student of Jascha Heifetz with dazzling skill, dynamics and colors. You have to listen to believe. A must
for demo.
FIM XRCD 2000, Messiah by George Frideric Handel
Yvonne Kenny, Paul Esswood, Martin Hill, Magnus
Linden
Members of the SWEDISH RADIO SYMPHONY ORCHESTRA
THE STOCKHOLM BACH CHOIR ANDERS OHRWALL.
Conductor
Recorded live at the Adolf Fredriks Church in Stockholm,
Sweden, on February 13 & 14, 1982
#189 16149 Redmond Way
Redmond, WA 98052
http://www.fimpression.com
Tel: 425-868-5326
Fax: 425-836-9061
Hawaii
5.0
by Gregory Hee and Jason Takamori
The Weems’ Project
In the mid-90’s, Hiroshi Ito began showing me literature describing triode amps,
horns, and speakers claiming to play “music”. I was curious, but unwilling to plunk
down $$$ for something I never heard.
At the library I discovered a speaker building book by David Weems. No cutting edge
stuff, but something caught my eye. Weems
described a single driver project using a
RadioShack No. 40-1354A 5 1/4" full range
driver that “gets the midrange right”. Actually that’s all he said.
Cheap enough. I picked up a pair to check
out Mr. Weems’ hearing. I sized the box to
my liking and built particle board enclosures from discarded shelving. He was
right! With no crossover, dynamics weren’t
restrained; no phase shifts, just music.
Without hesitation, I built some real 3/4"
MDF boxes. I surface mounted the drivers
and applied a 1/8" cork sheet to the front.
Kimber 8TC was used for internal wiring.
Hardware was all stainless. The insides
were lined with carpet pad on all surfaces.
I would guess the response to be 60 to
15,000 hz. Perfect for my small room.
Deep bass? Forget it, but then the highs
are limited too. The sound is balanced.
For 3 years, I listened in blissful secrecy.
Enter Jason Takamori
Then at a Jazz and Audio Society meeting,
Jason showed up. He asked me what I had,
but I talked about coherency and emotion
instead. Soon afterwards, I invited Jason
and Hiroshi over. It was the first time I let
anyone hear my secret system. At one point
Jason looked at his watch; it was past 1:00
in the morning. He kept saying, “It grabs
you”. And this is through my B&K ST140.
Jason asked me what else did I have going.
Well, after my experience I purchased a pair
of full-range Diatones that have been waiting for cabinets, but no further experiments
with the RadioShack. They weren’t perfect, but aside from an Edgar tractrix horn
TAD system with straight mid-bass horns,
I hadn’t heard anything else that made me
really want it.
Jason’s Super Boxes
RadioShack provided the motivation; the
drivers went on sale for a mere $10 each.
It was decided that the box needed to be
more rigid and the drivers braced. Bracing was easy; a T-nut and a bolt with a wood
plug through the rear of the cabinet sufficed. As for the box, Jason had other ideas,
like 3/4" Corian. Yes, cast cultured marble
would be more rigid, but he exercised restraint. His other deviation was to follow
Weems’ design exactly. So how did it
sound? “I can’t believe it’s . . .a
RadioShack.” Music was clearer and dynamics were more powerful. Some of the
upper midrange edge present in the MDF
versions were gone. Jason’s Spicas went
into his daughter’s bedroom.
A few days later, Jason brought over a
Corian version of my box (see picture and
plans). They sounded wonderful too. The
major difference is the larger soundstage
projected by Weems’ large front baffle. My
speaker presents a deeper front to back perspective. It is personal choice. By the way,
the enclosures are heavy!
Due to his 24’L X 16’W X 8’H room, Jason augmented his Corian speakers with
two bandpass subs that Dan Schmalle de-
signed for his Whamodyne project, powered by a Phase Linear 400. Upper frequency output from the subs is somewhat
controlled by adjusting cabinet stuffing.
The ShackAttack!
Can a cheap speaker get better? Hiroshi
calls it the ShackAttack.
It started with 300 hz tractrix horns intended for Dynaudio D54 drivers. For fun,
we put a pair of RadioShacks inside. The
sound was awful. Removing the cabinet
back was a transformation. The sound was
effortless, smooth and big, but midrange
detail was washed out. Adding a peizo
tweeter to provide some “air” didn’t help,
but the potential was there. So with the
aid of Speaker Builder magazine and the
internet, Jason and I charged forwarded.
Following Dr, Edgar’s construction suggestions in Speaker Builder, Jason built a new
300 hz horn (no back enclosure). We
couldn’t believe it. The midrange presence
was back even though highs were rolled
off as expected. Stereo images were deep
and solid even 30+ degrees off axis. The
horn mouth was still 14" X 11", but the
throat was enlarged to 3 5/8" square. The
drivers were mounted with a 1/4" gap from
the throat. According to literature, this acts
as a filter and will differ with another driver.
Despite success, there was a hole between
the lower midrange and sub, and images
were a bit small. Thus our final iteration,
a 240 hz horn to extend the coverage lower.
We also modified the throat mounting
board. The horns do what was intended
and present a realistic soundstage even
when sitting only 10 feet back. (See photos)
Where Are We Now
We still listen to the Corian projects. Be-
ing a Lowther PM5A front loaded 140 hz
round horn setup in Jason’s system, we
needed to test the line between practical and
extreme. It shows promise with lower distortion and more top end.
Here and Back
The RadioShacks turned out to be the key
that allowed us to learn what the “to lust
for” drivers do without breaking the bank.
For a list price of $16.99 each, it is a bargain basement over achiever, in whatever
project configuration you choose.
Beginners take heart. Enough information
is out there so that even a pair of novices
can get started. Don’t blindly believe
claimed efficiency. Crossovers can suck
power. The 89 db rated RadioShacks in a
box take less to drive than some 94+db
speakers with crossovers. Build the box
more rigid than you think necessary. You
will hear it. Brace the drivers. Even the
horn drivers were braced. Lastly, go for it.
Aloha!
sides working in smaller spaces, they “grab
you” in a different way. Like the horns,
they convey emotion. The horns by contrast are smoother and more effortless. The
horns also spurred our quest for a better
sub. Remember Schmalle’s beast is an under $100 project that needs no excuses.
And the search is on for cheap drivers with
lower distortion and a better high end.
A 40 hz bass horn, driver by a JBL D140
coupled to a 6L6 bass amp now shares low
end honors with the bandpass..
Are there any cheap drivers better than the
RadioShack? We tried the 5" MCM driver
used in the Whamodyne project. It is a good
driver, but doesn’t work in horn application. We also tried kevlar, coated paper and
poly drivers under $60. The performance
of the kevlar driver was frequency dependent and never cohesive. The horn just
magnified the plastic sound of the coated
and poly drivers. Voices and instruments
lacked the quality of being real. It appears
that paper drivers rule.
We also insertd the $300/pair Diatones
which breaks the cheap rule, but after hear-
Brainiac's
Today’s Topic:
the Allied 6K3VG power transformer
This low-cost leader (only $22.95) is one
of several in the Allied catalog. All are
cheaper than their Hammond counterparts,
but this one seems to be the best bargain of
the bunch. It looks about right for hefty
preamps or small (2-watt) power amps. It
weighs 2# 7oz on my kitchen scale, and
the lamination stack is 3” x 2.5” x 1.06”.
The laminations are thick, about 0.025”,
but they are interleaved one by one. It is
mounted vertically with metal end bells,
and takes mounting holes 2” x 1.75”. All
the wires come out on one side.
Heavy
Metal
...an occasional review of some nifty
chunk of iron (usually cheap!)
by Paul Joppa
Rated voltage
5v
6.3vCT
325-0-325v
Rated current
2A
2A
40mADC
The primary is rated at 117 volts and measures 11 ohms DC. There are three secondaries. The ratings and my measurements
are as follows:
* the high voltage winding was estimated
on the basis of 40mA RMS through the
whole secondary, which is a reasonable
approximation for capacitor-input circuits
Open circuit voltage
5,81v
7.35
367-0-367v
DC resistance
0.224 ohms
0.305 ohms
500 + 468 ohms
but is not exact. Leakage inductance was
not evaluated but may reduce the B+ output voltage slightly.
Equivalent source resistance in the B+ is
about 600 ohms, or about 6% of the likely
load resistance. With a capacitor-input filter and using silicon diodes, a maximum
of 435 volts can be generated at 40mA.
Remember however to allow for the voltage drop in the DC resistance of any filter
choke used, plus the loss in the rectifier tube
if used. With a choke-input filter an output
of 300 volts at 60mA could probably be
obtained without excess heating, again
minus the losses in chokes and rectifiers.
As is typical of smaller power transformers, the regulation is not the best, so be sure
to add series resistance to the filaments if
you don’t use the full rated current. You
can probably draw more B+ current if the
filament windings are not fully utilized, but
no more than half the extra volt-amp capability can be used this way without excess
heating.
Estimated voltage at rated current
5.13v
6.45v
333-0-333v*
300B
tube
tasting
By Paul Joppa
At VSAC 98, one of the seminar
events was a listening comparison
of six different 300B tubes. This event came
together almost by accident, starting with
the realization that there were two or three
types available in the local club. By simply
calling around, several others were obtained
- everyone was very helpful, and it is to the
credit of this hobby that there is so much
good will and enthusiasm on the part of
everyone involved.
The tubes tested were the following:
The setup was a good ordinary system, except for the Exemplar speakers which were
definitely a couple ranks better than the rest.
A decent CD player was followed by a
5670-powered preamp. The power amp was
the 300B-modified S.E.X. amp, running the
tubes at 325v and 55mA through a TFA204 output transformer; the driver was a
6SN7 mu-follower coupled with a Hovland
capacitor. Music was one jazz cut, followed
by some symphonic Mozart - about one
minute of each was played. Tubes were listened to blind the first time, then the whole
was repeated after the tubes were identified. About 40 people were in the highceilinged room, of whom some 23 left copies of their notes.
One major limitation of this test was the
lack of warmup time. Each tube was operated for one minute before listening began.
All but the AVVTs had at least several tens
of hours on them, but all were cold when
the testing started. This is thought to be a
particular problem for the Valve Arts 5300
with its massive graphite plate. So as always, take these results with a few grains
of salt!
I would only add a statement that both the
operating voltage and current were quite
low and not optimized for any particular
tube. (I know that AVVT definitely likes
higher voltage and current).
300B
“best” tube. But it did offer the opportunity
to discover what kinds of differences were
heard. As can be seen in the results, the
range of perceptions was quite wide - in
fact, one conclusion is that each person
hears different things, and weighs what they
hear differently, so that every tube sounded
best to someone. But there is also some family character to each tube that seems to come
out.
(I’ll add my 2 cents too. The VA5300 was
positively strangled at the op point used
for this test. A more appropriate Valve Art
tube would have been the C60, but these
were not available at the time of the test.
I’ve been using the VA5300 for several
months now, at about 480V and about 160
mA, and it’s a big winner in my book - B.)
1998
Clearly there are too many confounding
factors for this to be a test of which is the
n
Sovtek 300B, purchased from New
Sensor by the author for the original
version of the amplifier.
n
Western Electric 300B, a matched
pair loaned by Charles G. Whitener of
Westrex Corporation
n
KR Enterprise VV300B, loaned by
Ron Welborne who was exhibiting at
the show
n
AVVT AV302SL, loaned by Tony
Bombera, the North American dealer
for AVVT
n
Svetlana 300B, loaned by Eric
Barbour of Svetlana
n Valve Art VA5300, loaned by Ed Fallon
Part One listening impressions
I made several attempts to organize the listening comments, sorting them into groups
by various criteria of similarity. None of
them seemed to really help, and I continually worried that this approach was filtering other people’s perceptions through my
own unconscious prejudices. Finally I gave
up on interpreting the results. In the table
below each row contains one listener’s comments. The only editing is that I’ve removed
specific relative ranking comments. All
comments on the sound itself are reproduced pretty much as they were written.
Sovtek 300B
Westrex 300B
KR Enterprise
VV300B
AVVT
AV302SL
Svetlana 300B
Valve Arts
5300
Natural, easy,
pleasant
Air; more HF
Less air, nice
mids, bass OK
Soft, pleasant
Nice air, mids
natural, tight
tuneful bass
larger overall
sound
Lush, dynamic
Soft overall, less
air, mid OK,
bass is fat not
tight
Open, detailed,
but slightly
veiled
Nice air, natural
decay of piano,
full overall
sound, bass a
little recessed
Very nice,
dynamic
Veiled, lacks
sparkle, soft;
bandwidth
limited,
listenable
though.
OK, not much
top
Quiet, articulate, Midrange good,
nice
bland overall
sound
Nice overall
balance, highs
tinny on
occasion, lacks
articulation
Dulled
transients but
nice. Spitty.
OK, nice piano
& guitar, limited
bass, sibilant
treble, lack of
air. Nice violins.
Again nice
piano & guitar better bass.
Quite tight
(faster) - better
treble & good
air
Very good
Fuller on piano,
more detailed,
faster
Fairly good.
Nice, mellow,
smooth, laid
back.
Nasty sibilance,
fuzzy, boomy
Clear, smooth,
balanced, not
overly warm
Involving
midrange,
lacking in bass
punch, highs
seem a little thin
Nice sound
Thin, not full
sounding
Nice dynamics,
clear and lively.
Somewhat
blunt;
More bass,
same top
Open, natural,
some
brightness at
very top
Warm, rich,
clear
Soft or slightly
veiled mid,
background
could be
quieter. Voices
clear, brass
grainy
Good low end.
Nice flow to the
music
Some bass
punch
Warm & rich
sound, slightly
less dynamic
Good detail &
dynamics,
natural tonal
balance, best of
six
Extended,
clean, forceful,
dynamic
Pleasant but
light weight
Recessed but
more natural,
more depth,
more open
Drier piano,
good guitar,
treble not as
good as #2 and
slightly less air
Dull but mids
and drum good.
Cymbals flat but
pleasant.
Limited bass;
piano & guitar
OK. - sibilant
treble, not
much air
Bass less but
good. Treble
OK
I like, good,
natural. Can
feel a little bass.
Good piano &
guitar, bass not
quite there treble OK good air.
Piano OK limited bass
again - boomy
piano, sibilant,
OK air.
OK
Piano not as
good as 2;
good and fast
though on
instrumental
A little less airy
and natural but
excellent dark
dynamics
Cool, not as
engaging, but
sounded
accurate
Clear highs,
good detail but
want more
bass. Detailed
voice, resolved
violin
OK
Good piano,
more sustain.
Great pace on
drum. I liked
this one.
decent
Very good
Piano murkier,
more spurious
sounding on
instrumental
less
Slightly brighter
Piano murky,
OK
instrumental,
some lower mid
peakiness
boomy bass
Cooler, OK
Very nice tone;
balanced
Appealing,
balanced
Nice midrange,
reserved top
end. Voices
could be more
resolved.
Balanced, good
definition, quiet
background.
Nice vocal
separation &
violins.
Not as
involving. Could
be more well
defined.
Soft bottom,
good clarity
Good balance
top-bottom
More up front &
vivid that others
---
Somewhat
forward
midrange
Warm but
mushy.
Resentful
sounding
Overall average
Very quick &
detailed
Muddy
sounding
Sweet fullbodied sound.
Good dynamics
Beautiful clear
rich, good
dynamic,
human and
Hollow & dry,
mechanical.
Good dynamic,
no soul to it
Sovtek 300B
Westrex 300B
KR Enterprise
VV300B
AVVT
AV302SL
Svetlana 300B
Valve Arts
5300
Nice sound
Good low end.
Nice flow to the
music
Some bass
punch
Soft bottom,
good clarity
Good balance
top-bottom
More up front &
vivid that others
---
Somewhat
forward
midrange
Warm but
mushy.
Resentful
sounding
overall.
slightly smeary
Overall average
Very quick &
detailed
Muddy
sounding
Sweet fullbodied sound.
Good dynamics
Beautiful clear
rich, good
dynamic,
human and
soulful
Very open.
Hollow & dry,
mechanical.
Good dynamic,
no soul to it
Beautiful tone &
clarity cooler &
more neutral
A little recessed
and fuzzy. A
very good tube.
Thin, not full
sounding
Nice dynamics,
clear and lively.
Somewhat
blunt;
aggressive
Blunt attack
Warm & rich
sound, slightly
less dynamic
OK, congested
Slightly warmer;
great clarity.
Mids more
involving. Good
dynamics,
impact &
punch; open
Thinner, less
clean, less
detail
Very clean,
natural inner
detail
Nice and clear good
discernible
detail - low &
high better high
Full bodied,
forward a little,
good bass and
clarity
Better balance,
good dynamics
About the
same, a little
more midrange
Bright but not
harsh. Good
dynamics.
Soundstage
more forward
yet polite,
vanilla.
Compressed,
uninvolving,
veiled.
Forward, good
Thinner still.
Relaxed, quiet
timbre, a little
Smaller
mellow,
upper freq
soundstage,
rounded
glary
transients, good emphasis
pace
Very nice; a little Nice - maybe
Very full
more presence not as good
sounding,
clearer, more
detail
Good
dynamics, tilt
towards upper
mids
Good -?-
White,
bleached. Less
dynamic. Good
soundstage,
less detail.
Pretty good,
fuzzier
Good sense of
depth, mellow,
lighter bass but
clean
Thinner and
less depth,
lighter and airy,
boring
Good balance
and clarity.
Moderate bass
and sweet
highs
Nice space and
air. Clear sweet
highs and
snappy
dynamics
More lower
bass but a little
loose, more
extended highs.
Nice midrange.
Bright
Smooth.
Mellow, more
depth
Lush, warm,
Crisp, detailed,
lacks bass,
cold, shallow.
mushy, soft,
Dynamic,
voices not clear voices layered,
good on
symphony
A little thin in
Lots of detail, a
mids. Good
little thin, crisp,
frequency
good tight bass
range, fair bass. control, very
nice
Real good
luscious sound.
Slight upper
midrange
suckout.
Less defined
Good detail,
bass a tad light
Very good
dynamic, good
detail
Balanced
Depth, weight,
definition
Balanced
Lush, warm,
musical.
Dynamic,
voices clear,
bass OK
Sharp, harsh,
voices unclear,
bass mushy
Sharp, harch,
marginal bass,
OK
Good bass,
good midrange,
dynamic
OK, not as
dynamic
Nice, bass
good, musical.
Voices not
clear, not
detailed on
symphony.
Full bass,
dynamic,
smooth, good
frequency
range
No upper
harmonics, no
extended bass.
Nice midrange.
Easy to listen
to. Clean highs
but lighter
bodied piano
sound. Lighter
bass, fine
balance
Very good,
images well
Different, good
sound but
some weird
harshness, a
little lacking in
Sovtek 300B
Westrex 300B
KR Enterprise
VV300B
AVVT
AV302SL
Svetlana 300B
Valve Arts
5300
Unremarkable,
lacking detail
Detailed, rich,
nice high end
Slightly edgy
high end but
detailed mids
Nice air but
slightly edgy
high end
Clear, good
highs and
natural sound
Slightly edgy
but otherwise
good
300B tube tasting , Part Two - Tests on 300Bs
After VSAC 98, I had six different brands of 300Bs on hand. I had to return two of them at the show, but four remained, and I couldn’t
see returning them without making at least a few measurements to back up the listening tests. This article is the result.
I had recently developed a test method which would allow me to find the plate resistance, gain, and the nonlinearity factor which I
derived recently (see VALVE v.5 n.3-5). The method involves a simple experiment and a messy calculation. The experiment consists of
powering up the tube in fixed bias with a plate resistor of about the normal load value, 2000-2500 ohms for 300Bs. Grid bias is varied
from zero to cutoff, and the plate voltage and current are measured. The circuit is in Figure 1. Six or seven data points are more than
enough.
+300v - 400v
ca 2000
ohms
(variable)
-150v
Monitor
plate voltage
and current
Monitor
grid
voltage
300B
50k ohms
16
ohm
5v
5v
Fig. 1 - test setup.
The calculation is a nonlinear least-squares fit of the data to my model of tube nonlinearity, in which current I is the integral of the
3/2 power law over a Gaussian distribution of gain:
1+3σ
exp
I( Eb, Ec, σ )
( µ 1)
2
2 .σ
2.5.σ
2
.( Eb µ.Ec)1.5dµ
1-3σ
Eb is plate voltage, Ec is grid voltage normalized to a m of 1, and s is the standard deviation divided by the average gain. The curve
fit is done on equations of the form Ib K .I(Eb , Ec .µx , NLF) where the variables in the curve fit are perveance K, average gain
i
i
i
mx, and the nonlinearity factor NLF. I used MathCAD software to work out the fit. The plate resistance at a given current can be
calculated from the perveance.
This table summarizes the results, for both samples of each brand:
Tube
gain m
plate resistance rp, W
NLF
Sovtek 300B
4.085
542
0.146
“
4.079
563
0.141
AVVT AV302B
3.468
981
0.102
“
3.733
862
0.105
Svetlana 300B
4.056
666
0.112
“
4.018
645
0.116
Westrex 300B
4.034
631
0.117
“
4.006
625
0.113
The AVVTs were not a matched pair, but the others were. As can be seen, matching was extremely good for all brands. Figure 2 shows
a typical comparison of the measured current against the current from the model above; the fit is quite good. All tubes were run at 5.00
v rms on the filament except the AVVTs which were run at 4.80 v rms, at the suggestion of the retailer who so kindly loaned them (Tony
Bombera - thanks, man!).
0.1
0.08
Predicted current
0.06
Measured current
0.04
0.02
0
80
60
40
Grid voltage
20
0
Figure 1. Predicted versus measured current
A few comments can be made on these results. For one thing, the NLFs are not as low as I had thought they would be. This might be
because of the AC filament heating, however - further study might reveal more about this. The AVVT tubes meet their claim of being
more linear than any other, but they pay for it by having lower gain and higher plate resistance. All are still well within the expected
tolerances relative to the nominal specifications of 3.85 gain and 700 ohms plate resistance.
A few things puzzled me in the course of making these measurements. One was the high distortion of the Sovtek tubes, and the
interesting results published recently in Vacuum Tube Valley where this tube was found to have better linearity if the filament voltage
is lowered. So I tried running these tubes at 5, 4, and 3 volts, with these results:
Tube
gain m
plate resistance rp, W
NLF
No. 1 at 5 v rms
4.085
542
0.146
… at 4 v rms
4.104
572
0.133
… at 3 v rms
4.151
653
0.114
No. 2 at 5 v rms
4.079
563
0.141
… 4 v rms
4.127
594
0.124
… at 3 v rms
4.090
737
0.079
It looks like I have confirmed Matt Kamna’s results - they do indeed get more linear at lower filament voltages. At 3 volts, the increased
plate resistance suggests that this is not really enough filament voltage - the emission is really being limited - but it looks like they’ll run
happily at 4 volts. Incidentally, I also listened to them in my amplifiers with about 2.8 volts on the filament, which is what I get in the
2A3 filament switch position - they sounded a bit cleaner to me, but seemed to run out of steam on loud passages. At 3 volts, the filament
glow cannot be seen even with the lights off!
The other puzzlement was that one of the AVVT tubes seemed to have limited emission. In order to get a good fit, I had to leave out the
highest current test conditions. Figures 3 and 4 show the two AVVTs.
0.08
0.1
0.08
0.06
0.06
Ib
Ib
Ibx
Ibx
0.04
0.04
0.02
0.02
0
80
60
40
Ec
20
0
0
80
60
40
Ec
20
0
Fig 3 - weak emission, Fig. 4 - strong emission
While I was scratching my head about this, I got an email from Jim Dowdy asking if I would be willing to measure the nonlinearity of
some AV300Bs that he had recently acquired. This was too good an opportunity to pass up, so I accepted gladly and he sent me ten of
them for testing. The tubes were branded Audio Note, and older than the AV302SL’s that were tested above. This turned out to be quite
interesting!
On first testing them, I found that 8 of the 10 tubes seemed to have the same emission limit “problem.” To clarify this problem, I tested
a Sovtek and a Svetlana 300B (both of which had been used for 20-100 hours). They showed no drift at all, and would draw 110mA at
80-90 volts with zero bias. When set to draw 150mA, reducing the filament voltage from 4.95 to 4.50 reduced the current to 145mA in
both cases. Using the worst of the AVVTs, I found saturation at around 100mA - could not get more current no matter what the plate
voltage. The saturation current varied with filament voltage, as follows:
Vfil
Isat
4.90
4.80
4.70
4.60
4.50
115mA
105
93
84
75
This confirmed that the problem was saturation. The current is proportional to the filament voltage to the 5th power, very closely. By
now, I was getting worried - was there something wrong with these otherwise beautifully made tubes?
Talking to Jim some more, however, he mentioned that he has seen these tubes drift a bit in current for the first 24 hours. So I burned
two of them in, running a steady current of 70mA for a day and 150mA for another 12 hours on each. At the end of this treatment, the
“problem” was completely gone! They showed no limitations at 4.80 filament volts, and could easily handle the full 200mA that my
bench power supply is capable of.
The first round of testing, before the burn-in, was done with 2500 ohms in the plate lead, and 350 volts supply. This gave about 100v at
about 100mA with zero bias, ant of course 350 volts and zero current at cutoff, which was usually close to -110 grid volts. I did my curve
fit at -20, -40, -60, -80, -100, -110, and -120v bias, leaving out the zero bias point because it was affected by the emission problem.
Here’s the results:
S/N
mu
NLF
rp
1350
1398
1399
1418
1426
1445
1446
1456
1472
1484
3.51
3.76
3.45
3.72
3.58
3.72
3.72
3.69
3.66
3.67
.084
.136
.091
.123
.103
.128
.118
.152
.136
.124
903
798
965
779
880
825
811
836
804
815
I repeated the analysis leaving out both the zero and -20v bias points, but the results seemed to virtually identical, which I took to mean
that only the zero bias point was affected.
After burning in the worst two, I got these results:
S/N
mu
NLF
rp
1399
1426
3.67
3.75
.129
.130
851
776
Note that all the parameters are now very close to the averages of the other AVVT tubes.
I conclude that these tubes simply need a good break-in period, at least 24-48 hours, before they will perform properly. They are tested
at the factory for emission, but this is a pulse test which I suppose relies on the space charge more than the steady-state emission. The
equivalent steady emission seems to be 80mA in the factory test, based on my reading of their specs.
As for the parameters, the mu is on the low side (others usually measure about 4.0) but within spec and the plate resistance is on the high
side (I’ve measured 700 ohms or less) but again within spec. The nonlinearity factor is about the same as all the other 300Bs I’ve
measured.
As far as I can tell, the only limitation on these tubes is their reputation (along with the similar ribbon-filament KR tubes) for fragile
filaments, apparently due to the unavailability of truly high-purity nickel. (I do recall hearing somewhere that Westrex has a stash of
filament nickel from several decades ago). Assuming that slow starting with a NTC thermistor and tight control of filament voltage to
not exceed 5.00 volts will prolong the life, they should give fine service. Perhaps it would be wise to not turn them on and off too
frequently, though. In any case, new production from AVVT uses a new filament with a core of stronger metal, which is said to
eliminate the filament breakage problems.
new
strokes for
foreplay
by Doc B.
After a fairly long run with the Soul Sister
line stage prototype I decided I really should
get around to finishing one of the eXception
line stages for myself. Since this will be a
bit of a project, involving the dismantling
of the Soul Sister, and since I had given
my demo Foreplay to Mikey after nyNoise,
so I decided to whip up another Foreplay
in the meantime.
This would give me the opportunity to experiment with a few upgrades I had in
mind.
dress sexy
Although a lot of Foreplay owners have
spray painted the chassis plate, I had never
tried one this way. I was really pleased with
the black look of the powder coated B-Glow
prototype chassis plate, so I decided to go
for a black Hammerite. I did my trick of
spraying on the stuff in mist coats, allowing each coat to dry before applying the next
one, until I had built up a really nice pebble
texture.
pick ‘em up off the ground
Before I sprayed the chassis I enlarged the
RCA jack mounting holes to accomodate
the nice gold plated jacks we use with the
Glow kits. The holes were made large
enough to allow use of the insulators so I
could get my chassis to ground buss interface down to a single point, the center terminal of the terminal strip closest to the
tube sockets ( terminal 13).
Most of the assembly was done in the usual
manner, with the Anticipation upgrade,
natch. Our super cool 20.5 ga. high purity
long crystal copper magnet wire was used
throughout.
dont snub ‘em, unless they deserve it.
The first deviation was the incorporation
of a “snubber” ahead of the bridge rectifier. See Buddha’s Afterglow article in this
issue for a description of it’s function in a
far more coherent way than I can offer. This
filter blocks the reverse recovery spike of
the UF4007 rectifier diodes, keeping it from
reflecting back into the power transformer
high voltage secondary, and subsequently
the heater winding and the ;rimary, and
back into your other gear. Instead of connecting the red secondary wired directly to
terminals 4 and 5, a 10 ohm 2Watt
wirewound resistor was attached to each
terminal, and the secondary leads were attached to the free ends of the resistors. Then
a .01mfd, 2kV ceramic disc capacitor was
attached, one lead to terminal 4, one lead
to terminal 5, effectively shunting across
the output ends of the resistors. A second
.01 mfd cap was attached across the input
ends of the resistors, at the red secondary
leads.
Now the wiper ( center terminal) just serves
to shunt the signal to ground, but the signal does not pass through it on its way to
the grid. Cool, huh.
You can retrofit this to a stock Foreplay.
Just remove the wires that connect from
more central pins of the selector switch to
the right hand terminals of each of the pots.
Now connect a 47K ohm resistor from those
same selector switch pins to the terminal
2 of each tube socket. Don’t mess with the
other components connected to the pots.
choke them, but be careful
take the path of least resistance
The next trick was to convert the inexpensive 100K ohm volume pots to shunt operation. Thanks to Lynn Olson for this idea.
Normally the signal from the selector
switch connects to one end terminal of the
pot. The other end terminal of the pot connects to ground, and the center terminal
connects to the tube’s grid. The signal
passes thru the carbon track on its way to
the wiper ( center terminl) and the grid,
not necessarily the highest quality resistive
material.
Between the power switch the power transformer primary I installed one simplified
version of Buddha’s CMC filters to cut
down crap coming in on the power line.
See his article once again for the latest part
numbers. I use 12 mfd 250VAC caps I had
on hand and one of the CMCs that Buddha
sent to Smoothplate a few months back. It
was a shoehorn job, but I managed to get it
to fit under the chassis.
cap it off
I’m pretty careful to tell anyone who buys
a Foreplay that at $99 you’re not getting
the best coupling caps ever made. Duh. I
ended up leaving the great sounding
Hovland 2 mfd caps that George Wright
gave me in the preamp that I gave LaFevre.
So on this one I put in the 3.3 mfd Solen
caps we are now supplying with our
parafeed amps. They were on hand, and
ended up sounding surprisingly great for
the price.
In a shunt pot setup a 47K ohm resistor
connects directly from the output of the
selector switch to the grid of the tube. The
end of the pot which would normally be
connected to the selector is left unconnected
to anything, and the other two connections
stay the same.
use the biggest thing you can lay your
hand on
Tucker came over one day a while back and
braided me a power cord out of three 6’
lengths of 12 gauge solid copper wire, with
a big ol’ yellow Hubble plug on the end. A
total pain in the ass to braid, it makes your
hands ache for quite a while afterward. We
put it in in place of the hospital grade cord
I’d been using for my big VA5300 amp’s
plate supply, and the increase in bass punch
had us just staring at each other goggle
eyed( OK we was drinkin’ beer too). Kudos to Allen Wright for the original idea,
in his Cable Cookbook. Natch one of these
cords made it’s way onto the Foreplay.
Well, the result is a very nice preamp. I
lucked out to get a set of very unique Siemens ECC82s to use, which seem to have
nickel plates! They are really fast and clear
compared to my favorite 5963, although a
bit lighter in the bottom end. A pair of
TungSol 5963s proved to make a very nice
sound too, although unfortunately my pair
are hummers.
To damp vibration I used one of Andy
Bartha’s Whatchamacallits on the chassis
right behind the tubes, and a pair of IERC
tube shields. This really controls any microphonics.
Had a chance to compare this preamp with
the eXception at the last meeting. On Ed
Fallon’s system, which was a bit bass heavy
and mellow on the top end, I finally managed a one time miracle and beat out the
eXception, with the Siemens’ ECC82s upper frequency emphasis being to the advantage. A couple hours later on Tucker’s
eXemplars, the eXception with a 6N1P and
a nickel output transformer was the clear
winner, with lower distortion when driven
hard. Ditto on my own system with the
Lowther front horns and an eXception using the 5965 tube.
Considering the difference in cost and sophistication, I’m pretty proud that the Foreplay stood up as well as it did. Give some
or all of these mods a try.
UPDATE
Since writing this article I have developed
a new budget stepped attenuator. See the
next page for the story
Sweet Whispers
an affordable stepped attenuator kit from Electronic Tonalities
At last, a stepped attenuator for less than $150.
We have been asked many times over the past year for an upgrade to the stock potentiometers supplied with the Foreplay
preamp kit (something like ”Gee Doc, if anyone can do a stepped attenuator for almost free, it’s you”). Some Foreplay
owners have also expressed a desire for a more precise way to keep the dual mono volume controls balanced. So we
decided to put together an attenuator that fit with the ratio of price to performance that the Foreplay offers, no small order.
We eXperimented and came up with some minor compromises that would allow us to keep the price below $50 and still offer
a substantial improvement to Foreplay’s sound ( yes, even over the shunt pot scheme shown in this issue).
The difference in the cost of 23 position switches and 12 position switches is confounding. The best way to keep costs down
was to reduce the number of steps from the typical 23 ( of which about half are never used in the typical preamp) to 12
(eleven steps plus a “mute” setting). The current switch does not have a stop at the lowest setting (something we are working
on changing), but in practice this is not a big deal, you just need to stop turning when you hit the mute setting.
To keep the attenuation range flexible the step size was set at 3 dB per step instead of the usual 2 dB. This seems to be of
little consequence, we don’t find ourselves wishing for “in between” settings.
In the process of determining the resistor values, we realized that we could solve a problem faced by owners of sensitive
amps and speakers, that of ‘hair trigger’ volume controls that can’t be turned past about 9 o’clock before blasting. So the kit
comes with two sets of resistors -a set that will give 0 to -30 dB attenuation for typical medium sensitivity setups, and also
a set of resistors that will give -20 to -50dB attenuation, turning the Foreplay into essentially a unity gain preamp, similar in
gain to a passive preamp. You just build the attenuators whichever way suits your system!
Along with the mono switches shown in the photo above, we managed to find a reasonably priced two gang switch. And so
you can order an attenuator kit to upgrade projects that contain 100K ohm stereo pots too.
And another cool thing, they take about 20 minutes to build. You get a drawing that shows exactly where each resistor goes,
and they install in a manner very similar to the stock pots. In a Foreplay you will just need to do a bit of rewiring of the ground
buss once you remove the stock pots.
We said less than $150. Heck, didn’t we say less than $50? How about $40, for ether a pair of mono stepped attenuator kits,
or a single stereo attenuator kit. Doc takes care of your needs, baby...
Buddhafied
Afterglow part two
by John “Buddha” Camille
Several comments and questions have been
received from the field concerning the Afterglow modifications published in
CyberVALVE issue3.
This modification was intended as a minimum cost enhancement to fit Ken
Dangerfield’s needs of the time - not a
double throw down killer amp with every
trick known to man or beast.
Evidently the beasts who analyze various
circuit topologies with a background of
audio rumor, heresay and superstition are
hard at work saying it won’t work.
Thus I will add a very slight embellishment
to the theory of operation (editor’s note:
this embellishment has come to me in the
form of 34 handwritten pages - this is truly
the short version where Buddha is concerned). I will leave it to the real experimenters to charge into the books to find
out more about the whys and wherefores.
This modification did little to the basic signal circuitry - a few tweaks here and there.
A dual ground was added to make the amp
compatible with a follow on total system
design concept for low noise ( 100dB SNR
vs. the 50-60dB SNR of most tube equipment). A minimum of filtering and line isolation was added to the power supply. Everything in these changes was driven by the
parts availability at the time and the cubic
volume available. I consider the amount of
filtering and line isolation now provided
as a bare minimum, totally brute force and
very non esoteric. Esoteric spelled any other
way is m-o-n-e-y.
The power supply included a small percentage of solid state stuff to increase Ken’s
familiarity with the three legged fuse world.
The overall circuit is also safe in that it does
not require a good scope, function generator nor various analyzers to debug.
Experimenters should use the mod as a
baseline, not an end all. Most parts values
are not sacrosanct and in the main were
determined while kitting the amp from various store shelves. Still this amp will outperform 95% of the amps out there, especially the store bought variety. Experimenters may try different techniques, parts and
concepts at will - that is what valvedom is
all about.
Parts Availability
Several parts listed from Tanner Electronics are no longer available. Local manufacturers and foreign travelers frequently
wipe him out of the good audio stuff.
10 uF 250VAC line cap
The super small 10 uF 250VAC across line
rated caps were bought out by a local fabrication shop. A near exact replacement is
DigiKey pn P9408 @ 10 for $39. Tanner
carries a 10 uF 220VAC tubular for $1.99.
This tubular does not have the form factor
of the 9408 nor is it rated for across line
service, but is still legal since the primary
circuit is fused. This capacitance value is
also not magic, see the FL2 theory.
Relay K2 was spec’d with a 9V coil vs. the
actual 10V coil. Ask Tanner for the deer
feeder relay.
3K32 20W 2A3 cathode resistor
This mil spec chasis mount resistor of
around 3333 ohms biases the 2A3 to a little
over maximum rated plate dissipation,
about 15.5 Watts vs. rated 15 Watts, assuming a 425V B+. To maximize tube life increase the valus of the cathode resistor to
4K. Use Ohmite type 270/ DigiKey pn
L25J4K0. This value drops the plate dissipation to around 13.5Watts. Another option is to use a 5K adjustable power resistor to set the plate current. Use Ohmite/
Memcor style AR25/Mouser 588-AR255K. Dont forget mounting brackets. the
adjustable resistor will allow exact setting
of plate curent and thus plate dissipation
to your local operating conditions.
UF4007
Tanner’s has had a run on his UF4007 diodes. These diodes are available at both
Mouser, pn 625-UF4007. $0.25, and DigiKey, pn UF400DICT , $0.74. Am testing a
new Phillips diode, BYM26E, that may be
a cheap replacement.
.01uF 6kV disc cap
Parts Additions/Comments
The .01uF 6kV caps are no longer available at Tanner’s. He has a quantity of
.006uF 3kV disc caps that will work well
in this application. Also see PN Reverse
Recovery Filter.
MOV addition
K2 errata
Somewhere along the way two protective
MOVs were deleted from the power supply
drawing, sheet 4. MOV1 should be added
in parallel with the fourth 10 uF cap following CMC3. MOV2 should be added in
parallel with the primary of T-2. Use
Panasonic pn ERZ-V20R201/Digi-Key pn
P7322 ($0.72).
Contrary to popular belief, I add MOVs to
the primary circuit after all filtering for
several reasons. One, the line filter reduces
the slope of the dv/dt rise from line borne
impulse transients thus allowing the MOV
to clip earlier in the transient cycle. Two,
placing the MOVs next to each transformer
permits tighter clipping of the inductive
kick back of the transformer upon shutdown. This kick back transient is the usual
cause of solid state failures on the down
stream side of the transformer. Also, see
On MOVs.
CMC change
The Panasonic common mode choke
(CMC) special for CMC 1 thru 5 is spec’d
at 4.7mH/4A on the crates at Tanner’s. I
noticed that the CMC pn ELF-18D850B is
now spec’d at 3.3mH/4.2A in the latest
Digi-Key catalog. The CMC will probably
meet both specs but I wanted to eliminate
the confusion. Tanner also has a TDK pn
472Y4R0t CMC rated at 4.7mH/4A. Both
CMCs are less than a buck a piece. I prefer
the TDK.
Relay K2 substitution
The specified relays out two sets of contacts have been used to switch secondarty
circuits at voltages up to 400VAC above
ground. However, this has only been done
in conjunction with primary power switching, done here with K1. The contact point
spacing may not hold off stepped up line
transients if the transformer were constantly
energized as in some other applications.
Circuit changes or substitution of another
relay may cause problems so check it out.
Rather than dragging out my hi-pot supply, I kludge up an AC supply and let it run
for several hours per contact while doing
something else. This technique allows possible carbon tracks or other insulation failures to develop. Blow the relay here, not
after it is installed.
Discharge Diodes
Delete the discharge diode D1 between C2
and C3. This diode was originally used to
discharge C3 in order to prevent magnetization of the OPT core duriong shutdown.
However it has been found that resonant
effects can develop significant AC voltages
with some secondary loads on the OPT.
Thus diode D1 may become forward biased
during part of the AF cycle causing clipping. Diode D1 is replaced by a realy circuit that will prevent OPT core magnetiza-
tion from both charge and discharge currents to C3. See Demag Prevention.
Add a UF 4007 discharge diode between
C1 and C2, see the drawing. This diode
speeds the discharge of C1 during shut
down. I mentally miscomputed the time
constant of C1 (1500uF) and the cathode
resistor (3320ohms) at around 2 seconds,
vs. 20 seconds (25 actual). My power of
ten neuron dropped a digit. This fix will
decrease cathode stripping in V1.
Bypass Cap C3
Change the voltage rating of C3 to 600630 VDC. This change will provide a
slightly better safety factor against low frequency resonance effects and the resultant
high voltages that can be generated across
C3. the experimenter should be aware thet
the load on the secondarty of the OPT has
large effects on the amount of voltage generated on the primary.
A momentary disconnect of the speaker
load will cause extremely high voltages to
be generated by the series resonant circuit,
more than enough to break down the insulation of C3 and possibly the insulation in
T1. Of course this failure mode will occur
only if the load is removed when a low
frequncy near resonant signal is being amplified.
It is also concievable that a speaker having
low frequency resonant Z peak in the 10 to
30Hz region may cause the same problem.
It might be wise to measure the AC voltage across C3 with your normal speaker
load while playing a series of low organ
pedal notes or a test record, to see what
happens. Fortunately for most pocket books,
most SE speaker systems are resonant at
much higher frequencies, mitigating this
problem.
This resonance problem does not affect systems that use very large value of capacitance for C3 where series resonant effects
are well below the magnetic capability of
the OPT.
Manual OPT core magnetization
prevention
Add a manual magnetization prevention
circuit, see drawing. This circuit adds a
realy and manual switch to short the primary winding of the output transformer
during turn on and turn off cycles. The idea
is to prevent magnetization of the OPT core
by the virtually DC charge and discharge
currents of C3.
I found that these DC currents caused measurable ( visible on the scope) magnetic
saturation effects within the nickel OPT in
a tuner I was developing for bottlehead
Bruce Nilson. An automatic demag circuit
was developed for the tuner but it is too
complicated for this basic amp. Thus the
manual circuit is recommended. It will prevent core magnetization and should provide faster “warmup”. I found that it took
almost an hour for the tuner’s MagneQuest
B7 line transformer to “clean up” with a
20Hz signal at normal output - 1VRMS.
At full output , 5.5V RMS, the transformer
still required several minutes to demagnetize.
effective. The FL1/FL2 combination reduces line borne noise fairly effectively in
a relatively small volume and is dirt cheap
for the isolation provided.
Second, the network is also a rough approximation of a linear transformer that
reduces the AC RF line impedance significantly. For the experimenter with CAD capability, FL2 can be optimized using modern network filter theory algorithms such
that the output impedance is a sub multiple
of the input impedance. AC powerline RF
impedance normally runs in the 100-150
ohm neighborhood This impedance can be
halved or quartered without too much
trouble. the L’s and C’s can get pretty big
volume wise, however. While you’re at it,
put a notch at 180Hz where the worst noise
side band usually exists.
Why did you do that’s?
FL2
The network consisting of CMC 1 thru 3
and the four 10mfd capacitors serves several functions.
First, the network serves as a low pass filter that starts rolling off at several hundred
Hertz. FL2 is effective to well over 50 kHz
where the line filter FL1 starts becoming
Thirdly the filter FL1/FL2 smooths rectifier current impulses that travel back down
the power line to low level stages in the
audio system.
To be facetious, we are trying to simlulate
a kilobuck line cord here. A general rule of
thumb should be intuitively obvious to the
most casual observer - if you can hear the
effects of a line cord you need work on your
power supply.
Schottky diodes for BR1
Using Schottky diodes for low voltage high
current supplies virtually eliminates the
noise generated by pn diodes. The often
recommended high current bridges commonly called out for filament/heater supplies generate seious amounts of reverse
recovery noise. These high energy noise
transients are virtually impossible to filter.
Since Schottky diodes have essentially no
minority carriers they produce very low
level diode noise spikes. these low level
spikes are readily removed such as is done
with CMC4.
Schottky diodes also have roughly one half
the voltage drop of pn silicon diodes, thus
greater headroom is usually available to
down stream regulators.
RLC filters for HV rectifiers
The pn silicon diodes used to rectify the
high voltage B+ power generate a significant reverse recovery spike ( actually a
damped wave with a very high initial
pulse). The specified UF4007 diodes from
Tanner’s have a pretty clean recovery spike
- comparable to HEXFREDs. In fact in several installations I have found the UF4007
easier to filter than the HEXFREDs faster
recovery spike.
At any rate, brute force RLC filter networks
on each side of the rectifier diodes attenuate these spikes significantly - usually down
to the millivolt level. The filter formed by
CMC6 and companion capacitors further
drop the 10-20mV spike train down into
the noise.
NOTE: the filter/diode networks should be
assembled using extremely short leads, RF
style, preferably on a perf board with a
ground plane. In this supply all HV parts
from K2 thru the third 470mfd cap can be
mounted on a proto board available from
Tanner’s. Ask for the TI puddle board. Of
course the smoothing chokes L1 and L2 are
chassis mounted. Twist and shield leads to
the board from T1, L1 and L2. Shields
should be returned to the ground plane that
in turn is returned to the chassis through
3/8 inch stand off spacers. All small parts
should be mounted on the solder side of
the board and thus shielded between the
ground plane and the chassis. Also, ground
plane clearance holes should be countersunk slightly with a 1/4” metal drill (120
degree rake) where HV leads pass through
the board. Deflux the board and shoot with
several coats of clear Krylon for added
safety. ( Dang, Boss, and this is the quick
and dirty method???!! - B.)
The L ( inductance ) portion of the above
filters is provided by the use of 10 ohm
wirewound resistors available from
Tanner’s. Alternatively a 1/2W carbon film
or carbon composition resistor may be close
wound with #34 magnet wire. This method
is much less effective than the small
wirewounds. Be sure the first color code
stripe is double width, signifying
wirewound, when purchasing.
B+ discharge circuit
B+ may be conveniently discharged very
quickly (<2 seconds) by utilizing the center normally closed contacts on K2.
NOTE: observe the schematic and tie the
150 ohm resistor from the B+ buss to relay
lug 2, not lug 8, which should be grounded.
The primary reason for this circuit is to
rapidly discharge the HV buss in order to
reduce cathode stripping as the filaments/
heaters cool after shut down. Tubes are not
a quarter each anymore and the total additional cost of this mod is around 25 cents.
In addition the discharge circuit will probably be exercised a hundred times during
the debug and will no doubt save several
skinned knuckles.
Lastly, a bleeder that draws adequate current (10%) will dissipate another 4.5W of
under chassis heat ( 40Kohm resistor) and
take 5 minutes or more to drain the HV
buss to safe levels.
C4S diode
Why place the back biased diode across the
C4S for V1? Probably not really required
for this particular circuit but it has been
found necessary in other circuits, especially
RC coupled stages. I recommend using the
diode on all C4S installations. Here’s why:
The C4S has been tested and used by the
author and others in various installations
and has been found to be very robust. However some field installations have popped
the driver transistor several times. This
high Beta very small structure transistor
does not like back bias currents of more
than a few milliamperes. This current can
be generated by several mechanisms. The
most likely mechanism in this installation
is arc back between the plate- grid structure of the 2A3. Arc back is self clearing
and usually goes unnoticed but happens
periodically, especially in new tubes. Small
particles of filament/cathoe coating and
other foreign particles and debris break
loose and may be propelled by electric fields
to a point where the arc occurs. this arc
back biases the C4S and the transistors avalanche. This avalanche current is too much
for the fragile high gain driver transistor.
Another failure mode occurs when the C4S
is used to feed the plate of an RC coupled
stage. The charged coupling cap back biases the C4S on shut down and once again
the avalanche current may pop the three
legged fuse.
Other failure modes are out there, including hameous fistus, but the UF4007 should
be fast enough to correct the problem.
L1 and L2 filter chokes
A swinging choke is not necessary at L1
since this application uses a capacitor input filter. thus the output voltage has already soared to its maximum values. Also,
the current load of this amplifier does not
change appreciably - alleviating the reason
for a swingin choke, even with a choke input filter. Critical inductance for a choke
input filter is about 7Hy for this amp.
The specified chokes have more than adequate inductance and current capability
and could serve in future mods. The
100ohm DCR is an important factor here
due to low B+ overhead.
Why not regulation?
Voltage regulation was deemed unfeasible
in this mod for several reasons:
1
The power transformer plate
winding does not have adequate voltage
output (headroom) for a series type regulator.
2
The plate winding does not have
an adequate current capability for a shunt
regulator.
3
Inadequate room on Ken’s chassis.
4
Experience and test equipment
availability. A first time layout on virtually
any regulated power supply I have ever built
has always oscillated. Other folks seem to
have the same problem as several regulator kits and commercial “store bought”
amplifiers I have checked also oscillate or
go into a limit cycle mode. One popular
solid state kit regulator has had a limit cycle
oscillation at between 90mHz and 210 mHz
on the three systems I have been asked to
look at. The limit cycle mode,
manifested by a 1V to 10V RF signal riding
on the B+ DC voltage, can only be dtected
with the proper equipment such as a good
fast scope. Cleaning up the limit cycle removed the hard edge on the amps involved.
A popular all tube regulator making the
rounds also hs a limit cycle pulse mode at
several hundred kHz. The pulse train rides
on parts of the 120Hz output. The 120Hz
output of a volt or so was due to the extremely low bandwidth of the basic regulator circuit - DC to about 10Hz. It is hard
for an error amplifier with a 10 Hz
bandwidthto correct the 120Hz hum input
to the V-reg.
To be truly effective, a regulator circuit
must be faster than the load that is regulated. If your amp rolls off at 40kHz, a good
rule of thumb dictates a 400 kHz bandwidth
for the regulator - this is hard to do. Add a
wideband noise requirement of less than
100 microvolts and a low dynamic impedance and the problem really gets hard. To
get the gain/bandwidth for these two requirements is at the limit of the state of the
art for solid state, virtually impossible with
valves. This hard to do problem might explain the complaint of most folks that my
power supplies are much too complicated so they press on with their hum, hiss and
turtle slow systems.
At any rate, getting back to test equipment,
get yourself a surplus TEK 7704 or 7904
scope and at least a 7A26 vertical plug in
(200MHz). 7904 mainframes are around
$200 in this area. This setup will get you
through 90% of the V-reg development
problems. That 5 mHz scope might get
through most audio amp problems but will
not be fast enough for the power supply.
Hopwever, that hard sounding 6922 preamp
might also be oscillating at 400 mHz...
Grid stops
The majority of amplifier stage oscillations
can be cured with a grid stop resistor. High
Gm tubes ( over 2000) love to oscillate and
at some frequency virtually any layout will
become a tuned plate, tuned grid
(TPTG)oscillator, see a 50’s ARRL Handbook. The simple expedient of placing a
grid stopper at the grid lug of all tubes will
usually stop the TPTG oscillation effect.
When I say at the grid lug, I mean at the
grid lug.
If you are of the screen grid valve
persuassion, a grid stopper should be used
on the screen grid lug, especially if the
screen is returned directly to ground
through a bypass cap. Also, do not take the
shortcut of tying other leads to the grid lug.
The only part soldered to the grid lug should
be the grid stopper.
Plate stops ( not used here)
On several recent occassions I have observed VHF/UHF RF oscillations between
200 and 400 MHz in commercial 6922 voltage amp stages. The layouts were ideal for
an Rf oscillator, with relatively long (>1
in.) leads or traces to both the grid and the
plate of the 6922, a really hot tube for RF
work ( for which it was designed).
The first amplifier was stabilized with a real
stopper. The real grid stopper was kludged
into the circuit trace next to the grid lug
between the original film grid stopper ( ~1.5
inches away) and the tube. Remember, a
lot of film and bulk foil resistors go inductive above 50-100 MHz. This circuit them
became unconditionally stable over the audio cycle and sure sounded better.
The second amplifier calmed down with a
grid stopper placed into the signal lead at
the tube’s grid lug. It did not oscillate continuously, however the circuit would break
into a very low level RF oscillation during
a portion of an audio signal - a most common problem, dude. this oscillation was
killed with a 51 ohm 1/8 watt carbon comp
resistor between the plate lug and the “noninductive” plate resistor.
For starters, there is no such thing as a noninductive resistor. At some frequency the
so called non-inductive resistor will start
showing a significant inductive reactance.
If the tube and associated circuit has sufficient gain at that frequency it will oscillate, especially a high Gm frame grid tube.
This greatly simplified explanation is meant
to alert experimenters to the fact that many
things are going on in even the most mundane circuit. Layout and parts selection are
critical and every change should be evaluated. fortunately a critical ear can detect
when a change makes the sound better or
worse, but a new circuit may be discarded
out of hand for a subtle reason that may
only be betected with proper test equipment.
Note: the above part cycle oscillation only
occured when the plate swing went over
10V p-p. the spurious oscillation was of a
very low level , only 100 millivolts or so at
the frequency of around 150 MHz. This
appeared as a very slight thickening of the
CRT trace during the most negative portion of the 10V p-p, 1 kHz plate swing. A
high pass filter was kludged out of a 1 pF
cap and a 10 microhenry RFC at the tip of
the scope probe to better oberve the VHF
oscillation, i.e. to get rid of the 1kHz signal. The oscillation would also be killed
when the scope probe was connected to the
affected stage. This phenomenon occurs
frequntly, so the search for spurs can be
conducted by holding the probe near, but
not touching, various terminals. Of course
the scope should be set to one of the more
sensitive positions e.g. 5mV/div. with the
X10 probe during spur searches. Never use
a X1 probe due to bandwidth limitations.
The lessons here are that virtually any active circuit can amplify and oscillate at the
same time, or oscillate at several frequencies, or any combination of the above. The
famous hard sound of the 6922 is a typical
example - it is usually oscillating. Also,
suspect those very neat layouts where all
parts are neatly aligned with leads or traces
meandering all over the place. those
nanohenries and femtofarads turn into nice
VHF-UHF tank circuits.
Speedup resistor
Why the 82 ohm resistor in series with the
diode across K2’s coil? In the old days this
resistor was called the speedup resistor the physics of which I won’t go into. As
used here, this resistor speeds the release
time of the associated relay. this reduces
contact arcing to some extent and in this
case to 1/4 cycle or so of the 60Hz supply
due to pulsating DC nature of current flow.
I generally use a resistor value that approximates the DC resistance of the relay coil.
This value halves the value of back EMF
current flow upon shutdown, reducing the
relay armature magnetic hold force significantly. The down side is that the back EMF
voltage spike generated is twice the relay
coil voltage vs, approximately 0.7V. In this
case approx. 16V. The 16V pulse is no problem for the PN2222A that is good for 60+
Volts. The 16V pulse might back into the
8V buss and damage the 555 but this pulse
is clipped at ~ 12V by the 12V TVS. This
is a side benefit of the TVS whose primary
purpose is to clip line borne transients on
the 8VDC buss. A tertiary purpose of the
TVS is to clip the shutdown transient pulse
from K1.
Easy Tweak
( editor’s note: gasketing the bell ends as
described below will void the warantee on
MagneQuest products. We are not saying
it won’t work, just that you are risking dropping and stressing the unfastened core far
worse than stresses described within this
next section - proceed at your own risk)
Permalloy, mu metal and other high nickel
( Ni) cores must not be stressed mechanically for optimum results. High tech cores
are enclosed and floated in silicon grease
or very compliant elastomers such as silicon rubber. The enclosure prevents winding stress to the core material after heat
treatment thus maintaining the magnetic
qualities of the core metal. The experimenter can relieve future stresses by proper
mounting techniques. If the transformer is
rigidly mounted to a chassis differential
expansion rates will keep the core under a
constant state of stress during operation.
This tweak will prevent these stresses from
building to typical levels. This tweak will
also produce audible results on silicon steel
cored transformers.
Mod Procedures
1
Remove the bell ends of the transformer. If the assembly bolts are steel, replace them with threaded brass rod.
2
Fabricate two rubber gaskets taht
will fit between the bell ends and the core.I
use the soft neoprene gasket material sold
at hardware and auto parts stores.These
rubber sheets are .020 to .030 inch thick
and are dead soft and pliable. The idea is
to prevent any machanical contact between
the bell ends and the faces of the core ma-
terial.
3
Cover the transformer assembly
bolts/rods with shrink or vinyl tape over
the portion that passes thorugh the core.
Prevent mechanical contact between these
bolts/rods and the core material.
4
Reassemble the transformer, ensuring that lead wires and gasket material
are properly positioned. Tighten all nuts
finger tight.
5
Draw all nuts down evenly using
a nut driver. Draw down should be done in
several stages using crisscross sequence
like the auto wheel lug torquing method.
Final torque should only be 6-8 inch
pounds.
6
Degrease the transformer assembly and shoot with several coats of clear
Krylon. This clear coat will stop the slight
shock one will receive from the core or the
bell ends during operation. The shcok is
caused by the now totally isolated core rising to around B+voltyage due to electrostatic forces. Older paper insulated transformers should have a drain wire ( small,
stranded, tinned copper) inserted between
the rubber gasket and the core. The drain
wire should be returned to the amplifier
chassis ground.
7
Mount the transformer using nylon hardware and shoulder washers. Drill
mounting holes so that several thousandths
of lateral play exists between washer shoulders and the chassis. Add plain washers as
necessary between shoulder washers so that
several thousandths of vertical play exists
between each shoulder washer set and the
chassis. The transformer should be free to
move several thousandths both vertically
and laterally afetr all nut/bolt sets have been
torqued to 8-10 inch pounds ( 8-32 hardware). After final assembly and debug, the
nut/bolt sets may be RTV’d for long term
security.
On MOVs
I have been using MOVs on every piece of
gear built or overhauled since 1975 when
my house in Tampa received part of a lightning strike to a nearby pine tree. Most of
the stroke went directly to ground through
the tree as evidenced by the complete debarking of the 10 insh caliber tree. Part of
the stroke went over to the house through
several dozen roofing nails and thence into
the house wiring. My brand new (1 month)
solid state Heathkit TV and every piece of
ham gear using FETs were damaged. Also
every motor in the house including the AC
compressor plus water heater and stove
heating elements were zapped.
While at Graingers buying new compressors for the AC and refrigerator, plus all
those other motors, a prominent display of
GE Thyrite surge suppressors caught my
eye, so I bought several. In addition GE
MOVs were starting to appear in electronic
parts houses, so I got a hand full of those. I
placed a Thyrite across the electrical service entrance, across the breaker panel, the
Ac compresor and the AC airhandler. 20
joule MOVs were placed across all other
motors, all appliances, and all electronic
equipment.
A year later alomost to the day my ham
antenna mast took a direct hit - a portion
of which got into the houe wiring and vaporized the innards of the power meter.
Also had a lightning ball bounce between
the AC grills in the living room, so plenty
of energy entered the house. The only damage suffered during this second hit was to
the masthead preamp where the helicoil
front end filter was vaporized along with
that new $40 FET.
Since that time I’ve treated each new house
the same way, thyrites and MOVs across
everything. Four or five years ago the
present house took a hit ot the chimney
mounted TV-FM antenna. the only damage suffered was the vaporized balun between the antenna and the masthead amp.
A coax spark gap suppressor even saved
that super lo noise transistor in the front
end of the mast head amp. I have yet to
experience any of the ills suffered by several writers in recent audio rag articles.
Even if i did suffer a MOV failure or two I
would really be happy. That six bit sacrificial lamb went down fighting and savd me
a thousand bucks of bitching and moaning
while fixing damage.
The gradual degradation of MOV devices
has also not been noticed. Maybe this is
due to the dual layers of protection offered
by the Thyrites and several dozen MOVs
across the line at any one time. Initially I
used to measure the breakover voltage on
MOVs installed in test equipment undergoing calibration. No observable increase
in breakover had occured in many samples
so I quit the practice 5 or 6 years ago. I still
check the MOVs across the fridge occasionally and no deterioration has been noted.
This compressor motor produced up to 6
kV pulses on each shutdown.These spikes
are reduced to about 1kV with the MOV
protection.
GE Thryrites are code in the US and
Canada and are available at the nearest GE
supply. They are called “Traquell surge arrestor” with pn 9215ECB001 at $60 - 70
each. These deviced are easy to install in a
convenient knock out or can be suspended
by the elads in breaker boxes, etc. Keep
leads short and straight, one black lead to
each side of the 220VAC line and the white
lead to the neutral block.
MOVs should be connected across the line
and from both line and neutral to ground
in major appliances and other high dollar
equipment. For the small stuff, place one
across the line, line to neutral.
On inaccessable gear such as the fridge and
microwave place the 3 MOVs across the
back of the wall receptacle.
the EXO-50
“What this country needs is a good 5 watt 845 amp”
Mike LaFevre, MagneQuest
Here’s a neat design idea Paul Joppa whipped up in an evening, after an e-mail from Doc. The basic premise was to do an 845 amp
using the PGP 8.1 power trans, and the EXO-50 parafeed output transformer. Here’s what Paul came up with “With a full wave bridge and a choke input filter you can get 630VDC minus the choke losses, at 60 mA. This is workable, a 5800 ohm
load is optimal so the 5K EXO-50 should work well. I recommend leaving the PGP 8.1 CT alone. The filter choke needs to be pretty big,
at least 15H, to get proper performance at full current, and at that inductance it needs to be rated for the full AC also, probably 100mA
or more. The caps need to be rated for 1kV peak, of course (660 VAC).
If you do it as cathode bias (69VDC) you drop to about 525VDC across the 845. The operating point is very good - it only dissipates
32W ( those $40 a piece Chinese 845s will last forever! - B.) and you’ll get 7.5W output.
Pushing driver plate voltage down to 200VDC and running it at 4 mA, you will get into 10K Rp territory, 12AT7 types could work.”
Doc drew up a rough and ready schematic, which is shown on the next couple of pages.
This is an untried circuit, but as usual we have a pretty high confidence level in it’s performance. What we would like to do is invite your
questions and commentary on the Bottlehead Forum of Audio Asylum. We will use the Forum to offer up refinements on parts values
and info on things like the availability of that unamed MagneQuest filter choke...
EXO-50 parallel feed output transformer 5K primary, 16,8,4, ohm secondary, 20 watts, M6 and Permalloy “Pinstripe”, $550 the pair
EXO-04 plate loading choke 50H, 60 mA, 317 ohms DCR, $198 the pair
BCP-16GC grid choke!!!!!H, no DC, $80 the pair
PGP 8.1 power transformer $165 the pair
C4S active load kits $35 the pair
Large Richard
a Parafeed 845 Se amp
Power SupplySchematic
10VAC filament transformer could be Mouser PN 553-VPP105600 (Magnetek10VAC 5.6A)
Filter caps could be Mouser PN 5987-660V30 30 mF 660VAC C-D motor start cap
15H 75mA choke from MagneQuest TBA
Use double UF4007 to handle voltage peak
See Buddha article in this issue for info on reverse recovery spike filter ahead of full wave bridge
Large Richard
a Parafeed 845 Se amp
Amplifier Schematic
Try Mouser PN 5987-660V2.0 C-D motor run cap for parafeed coupling cap, or other 2-3 mfd 1kVDC rated cap.
MagneQuest BCP-14 with 100Kohm 5W wirewound resistor may be substitued for C4S load and 62K WW resistor. Don’t leave out the
100mfd cap!
50 ohm 10Watt pot may be available from All Electronics, www.allcorp.com, PN POT-50C
Use your favorite flavor of wirewound resistors. Ohmites are very alright in Doc’s book.
da’ basics
Basic design considerations for
plate loaded triode stages
Here’s an very cool reparte that occurred
between Bart Shepard and Paul Joppa, after Bart requested some enlightenment on
the Sound Practices “Joelist”. Note here
that Bart is very humble in his asking for
help, but that he has obviously studied hard
to figure which questions to ask - he’s not
just throwing cookbook circuits together,
but rather he’s working to grasp the knowledge tht will let him design his own fantasies - dass’ what I’m talkin’ about!!! - B.
Hi guys,
I have a couple of really basic design
questions that keep bugging me and
thought I’d ask them here to see if any of
the very knowledgeable guys on the list
could sort me out.
I am interested in the design of a basic
plate loaded triode stage. As I understand
it a constant current through the tube will
provide for minimum distortion and
maximum gain. Hence all the fancy totem poles and CC sinks and sources
around.
— yes; though power supply isolation is
another benefit.
I also understand that these effectively
provide a very high impedance load.
— yes
some 60dB isolation. That’s why John
Camille pushes for those high impedances.
I know several people who claim to hear
the difference; my own direct experience is
not enough to make any claims.
Does the noise of the load R become a
problem? If so, in what circumstances?
Is this just an issue for pre-amp stages?
Surely not for high voltage swing driver
stages.
— yes, it does. Most bulk resistor materials (carbon, sputtered metal) exhibit some
noise voltage proportional to the voltage
across the resistor, in addition to the thermal noise common to all resistors. The
higher the voltage, the higher the noise;
consequently it cannot be ignored in driver
circuits. Solid metal resistors (wirewound,
bulk foil) show much less of this behavior.
That’s why Mills wirewound are typically
called for as plate loads. Also the voltage
capability of the resistor is a limit. Most
resistors are rated for 200v or so. So a
“good” plate load might require several
resistors in series, if operating from a 1000v
power supply.
The next limitation I am aware of is the
resulting high Zout for the stage. This
leads to my next question.
— the output impedance is the plate resistance in parallel with the load resistance;
it doesn’t go up more than 30% or so relative to a conventional resistor-loaded circuit. Often the CS loads are used with an
unbypassed cathode resistor, however,
which raises the effective plate resistance
by (nearly) mu. That can be a problem.
My first question is what limits just using a very large plate load resistor? The
usual limitation quoted is the high B+
voltage required to provide the desired
plate voltage. But in , for example, a 211
or 845 power amp, high B+ is not a problem.
How do I calculate the maximum Zout
allowable to correctly drive the
output load (e.g. should Zout be < 10%
following stage Zin??) and in fact how
do I calculate what this load is when it is
the grid of a following stage?
— depends on how high an impedance you
want. Just looking at the plate curves and
thinking of distortion, a load impedance of
2-3 times the tube’s Eb/Ib gets you most of
the way. But for power supply isolation, that
only buys you 10dB or so. Very high load
impedances can be achieved - I think the
CCCS is at least 10 megohms - that gives
— Usually the most important factor is the
grid capacitance. There are two limitations,
frequency response and current capability.
Frequency response is the driver source resistance (see above) interacting with the
capacitance. Most designers seem to think
that a -3dB frequency of 100kHz or more
is desirable. For current capability, I think
it was jc morrison who said that the stand-
ing current in the driver stage should be 5
times the rms current drawn by the capacitance at 20kHz. Both of the above criteria
seem to be quite conservative.
If I design to not have to drive the following stage to extremes I should be
able to avoid grid current, shouldn’t I?
So shouldn’t this load be not much lower
than the grid resistor for a self biased
stage?
— Only very few tubes have a max. grid
resistor so low as to be important here. The
type 50 is an example; it specifies a maximum 10k grid resistor as I recall.
Do I need to take into account Miller effect as an additional load? If so, how? (
add Cgp & Cgc then times gain??)
Should this reactance be > 10 times the
load on the grid?? or does plate load only
affect this re frequencyresponse?
— Certainly; see above; it’s the dominant
load usually. Strictly it’s Cgp times gain,
plus Cgk, plus stray capacitance depending on layout. It’s usually good enough to
use Cgp times mu.
Does the nature of the following stage
matter? It seems that output stages running high currents need better drive than
low current VA stages. Why?
— Basically, output tubes are physically
large and have highter capacitances. Plus
they need lots of drive voltage, which increases the current into their capacitance,
thus increasing the minimum standing current in the driver.
Is this due to a need to be able to supply
drive current? If so, why, if not driving
into grid current or does grid current
start at biases well above zero volts??
— as noted above, it’s the drive current into
the grid capacitance. There probably is
some advantage to being able to drive into
class A2 on transient peaks, but that’s a
whole different can of worms. (Incidentally,
*some* grid current shows up when the
grid is within 2V of the cathode, depending on details of the materials...)
Or is it due to a need provide a low impedance load on the grid for some rea-
son????
— no
Is a low driver Zout only required for its
ability to deliver drive current or are
there other reasons?
— As above, the frequency response is another consideration.
What appear to be trivial basic questions
soon seem to get pretty involved to me
and I haven’t found the real answers easy
to come by in any of the introductory
texts I have looked it. A lot of this stuff
seems to be skimmed over, an infinite
grid impedance not requiring any current!!
— yes.
Sorry if this stuff is real basic but perhaps there are others like me who get OK
results but don’t really deeply understand the issues and your responses may
help them also.
Cheers,
Bart ( not too proud to ask :-))
Entwined
interconnect cable kits from Electronic Tonalities
pre-cut twisted pair 20.5 guage High purity continuous cast long crystal solid copper
conductors with a high temperature insulative coating, pre-cut polyethylene outer jacket
and Vampire Wire RCA plugs
$39 .5 per meter pair, $49 per 1 meter pair, $59 per 2 meter pair
VALVE
the magazine of astounding sound
THE PENTODE ISSUE
A Novel Tranformer-Coupled 6V6
Stereo Amp
A Doc B. Modded Zen 84B
Twelve Ways to Parafeed
What I Wished I Would Have Known
About Dual Volume Controls...
Heavy Metal and more...
volume 6
number 5
1999
“If it’s musical refinement you’re after, this is the
address you’ll be looking up.” Peter Breuninger,
Listener, Autumn 1999
“...the Paraglows are very good music-makers: insightful and immediate, with excellent pacing and
pitch accuracy. They drive Lowthers beautifully
well— ditto Galantes...” Art Dudley, Listener, Autumn 1999
“In one word, GLORIOUS!!! … After only a few
hours of break-in, they presented a soundstage from
room edge to edge, with at least five feet more of
depth than I’ve heard before. Even on digital, I
could make out hall sounds never heard before...The
mid range was almost palpable, with voices taking
on that three dimensional quality of a living body.
Bass down to 50 Hz. was tight, controlled and
THERE. Measurement showed it to be flat down to
at least 10 Hz ... Highs extended out as far as I could
hear.” Bill Gaw, EnjoytheMusic.com, November
1999
Paraglow— Currently inhabiting the listening rooms
of Lowther America, Avantgarde Acoustics,
MagneQuest and Listener Magazine.
www.bottlehead.com
VALVE
the magazine of
astounding sound
Chief Editor and Publisher
Dan “Dr. Bottlehead” Schmalle
Managing Editor
Brad “Just fell off the turnip truck” Brooks
Chief Administrator
“Queen Eileen” Schmalle
Resident Smart Guy and Technical Editor
Paul “Braniac” Joppa
Resident Hot Iron and Dr. B’s Bodyguard
John “Smoothplate” Tucker
Big gun OEM advisor
Michael “Airgap” LaFevre
Contributing Editors
David “Full Track” Dintenfass
“Crazy Eric” Lenius
VALVE
P.O. Box 2786
Poulsbo, WA 98370
by phone: 360-697-1936
business hours: 9-4:30 PST, Mon -Fri
technical queries: Thurs. 12-4:30 PST
fax: 360-697-3348
e-mail - [email protected]
Editorial and Advertising Inquiry
1428 Roberts Ave.
Nashville, TN 37206
by Email: [email protected]
by Phone: 615-228-2792, evenings CST
Let it ring...
VALVE in no way assumes responsibility for
anyone harming themselves through exposure to the
contents of this magazine. We believe electrons flow
from minus to plus, and that they can kill you along
the way if you’re not careful. Vacuum tube audio
equipment operates at potentially lethal voltages.
Always treat it with respect.
Many ideas published in this magazine are untried,
and involve the use of potentially dangerous parts
and tools. In attempting any idea or project
published herein, you assume total responsibility for
your actions and any harm caused to yourself or
others.
This publication is produced as a service to
the audio community and is wholly owned
and published by Electronic Tonalities. The
intent of this publication is to offer ideas to
inspire and educate audiophiles in an effort to
increase their understanding of the audio
equipment that they use and cherish. Blatant
copying of the circuits published in this
magazine for use in commercial products
shows a complete lack of original thought.
Don’t just copy and distribute stuff without the
author’s and publisher’s consent, OK?
editor’s
thing
Season’s Greetings!
I am really excited by the opportunity to edit my favorite
audio eXperimenter's and DIY
magazine. Under Doc’s guidance, VALVE has developed into
a valuable resource for audio
eXperimenters and Hi-fi enthusiasts around the world. As the
new editor, I am excited by the
challenge of continuing to offer
exciting technical features.
VALVE will continue as a tube
enthusiasts, experimenters and
DIY journal, just like when it
started back in 1994. It has
been a formula for success
since the beginning, so why
change now? You can see for
your self—the VALVE CD-ROM
Archives are now available.
Was VALVE ahead of it’s time?
Maybe. I think that it was innovative even in it’s humble
beginnings as a simple club
newsletter— get the back issues and check it out.
Look forward to a few small
changes. Beginning January
we will be publishing VALVE
quarterly. We will also be revamping the VALVE web page,
and we will be posting additional goodies there in-between
issues.
I will be eXperimenting with
various formats to enable you
to use this document easier.
PDF format is great, and we will
continue to distribute VALVE in
PDF format. I will be looking at
ways to make it work even better as a screen document, so it
is easier to read at work (on
breaks only!) Do not have your
BAN up and running yet
(Bathroom Area Network)? So
print it out. That way you can
read it on the bus, or while
waiting in the car at the mall
while your “dear heart” shops.
Either on screen, or as a
printed document, your
choice. We will constantly be
tweaking.
We will be searching out more
advertisers of DIY kits and
parts to include in VALVE. Our
ad costs are reasonable, heck,
they are dirt cheap! Look forward to new sponsors in the
coming issues.
Thanks to everyone for your
patience while we produced
this latest issue. Please drop
me a line now and again with
feedback on how we are doing.
Thanks to Dan, Eileen, Paul,
Dave, John T., John C., Mikey,
Eric, and all others who have
contributed thus far to make
VALVE what it is. They have
set the bar high. I thank Dan
for the opportunity to continue
his fine work. Happy New
Year everyone, and Don’t let
the blue smoke out!
Without further adieu, we
bring you:
THE PENTODE ISSUE
On the cover:
Cameron Etezadi’s 6V6 amplifier.
Read more about how he designed
and constructed this screamin’
blue amp and power supply.
Keep those equipment Glam shots
coming!
new
classics from
First Impression Music
FIM XRCD 022, The Artistry of Linda Rosenthal
Linda Rosenthal, violin, Lisa Bergman, piano
A collection of timeless violin solo favorites, by the most
talented student of Jascha Heifetz with dazzling skill, dynamics and colors. You have to listen to believe. A must
for demo.
FIM XRCD 2000, Messiah by George Frideric Handel
Yvonne Kenny, Paul Esswood, Martin Hill, Magnus
Linden
Members of the SWEDISH RADIO SYMPHONY ORCHESTRA
THE STOCKHOLM BACH CHOIR ANDERS OHRWALL.
Conductor
Recorded live at the Adolf Fredriks Church in Stockholm,
Sweden, on February 13 & 14, 1982
#189 16149 Redmond Way
Redmond, WA 98052
http://www.fimpression.com
Tel: 425-868-5326
Fax: 425-836-9061
A NOVEL TRANSFORMER-COUPLED 6V6
STEREO AMPLIFIER
by Cameron Etezadi
ABSTRACT
Certain types of music or certain types of speakers require more power to
drive them to acceptable volume levels than a relatively inexpensive DHT or
other SE configuration can produce. The inexpensive solution to this issue
is construction of a push-pull amplifier, with the resultant change in sonic
character such a topology brings. This article details the construction of an
inexpensive and effective push-pull amplifier, offering excellent sound at
increased power levels. Some interesting design tricks are employed; these
may be successfully utilized in other single ended or push-pull designs. The
result is an amplifier with an almost flat frequency response from
approximately 18 Hz all the way to 24 kHz, 0.58% THD at 1 W, and no
shortage of detail or bass, yet with a total finished cost of less than $600,
well within the budget of the average do-it-yourself builder.
INTRODUCTION
When the Whammodyne! speakers were first published in VALVE, a copy quickly
became my main reference pair of speakers. They offer remarkable efficiency, extreme
detail, and have a reasonable frequency response. These speakers were bi-amplified with
identical pairs of homemade 6GW8/ECL86 single ended amplifiers, using parts
scavenged from old phonograph amplifiers generously provided by both my mother and
George Wright. This setup worked well for a while, but the strained three watts per
channel provided just could not handle the heavier passages of more modern,
contemporary music, nor could the meager power supplies and output transformers keep
up with the dynamics of the more refined classical or jazz pieces. Given these
shortcomings and a slim budget to spend on audio equipment, the design of a cheap, yet
reliable and high fidelity push-pull amplifier was sought that could remedy this situation.
The limitations and drawbacks of push-pull designs are obvious, yet the more reasonable
prices of the output transformers, coupled with the higher power levels available from
such configurations make it the logical choice, if the price to performance point for this
project was to have been reached.
DESIGN CRITERIA AND GOALS
This amplifier had only two major goals; these seem to be the common goals of all
VALVE projects: good sound and a price as low as possible.
With these two goals in mind, the search began for parts to use. One of the easiest ways
to bring the price of any project under control is to scrounge. The Puyallup, Washington
hamfest was a great help here; for those in the eastern part of the country, a spring trip to
the Dayton Hamvention would be in order every year. A suitable old TV transformer and
a heavy duty military surplus choke inductor of 5H at 300 mA were found, each for $5.
Several chassis were either bartered for or purchased on the surplus market for under $10
each. Inexpensive NOS tubes were also procured for $1 each or less. Once the junk box
was reasonably stocked, the amplifier layout could begin.
First, to keep cost at a minimum, two power supplies would be impossible. Each would
require a chassis, a large heavy transformer, and a matching choke. Finding more than
one transformer of a matching pair was not possible. That meant the stereo supply had to
be shared between the two audio channels. Monoblock amplifiers provide better sound
quality, though, as the two audio channels do not couple to each other through the power
supply. To ensure adequate decoupling in the single supply, the power supply for this
amplifier is regulated, with each channel being routed through a separate series pass
element and utilizing an independent voltage error amplifier. Adding additional regulator
parts is cheaper than adding another transformer and choke, unless one happens to have a
matched set of two in his junk box. Furthermore, there are additional benefits in a
regulated supply, specifically in the area of supply impedance, which make it worth
considering. A regulated supply will also help remove a few extra decibels of hum from
the circuit, leaving more headroom for the music. Hum was a major problem in the
supplies of the 6GW8 amplifiers, and great pains were taken to eliminate it from this
amplifier.
The only downside to a regulated supply is a lack of the sonic effects of “tube rectifier
sag” upon the circuit. This effect is often desirable, and many of the projects presented in
VALVE are often the subject of rectifier swap tweaks to change their sonic signature.
However, the intent of the project is to hear the amplifier, and not to listen to the power
supply for the circuit.
Scavenging my inventory for suitable tubes to use in this project, valves were chosen that
were either in current production, large surplus supply, or easily replaced with solid state
components, should the need arise. Also, the search was limited to the use of octal based
tubes, exclusively. Most of the pentodes and tetrodes and beam power tubes that are
suitable for audio output are based in an octal fashion. Excellent preamplifier tubes are
also available in this size. Keeping the size uniform throughout the project permits a
single size of socket, a single size set of hardware, and a single hole saw to bore the
chassis out, thus saving several dollars overall. The 6V6 tube was picked for audio
output. This tube was common in early guitar amplifiers, but was largely replaced with
the larger 6L6, EL34, or 6550. Few guitar amplifiers past the late 50’s or early 60’s were
designed with this tube. It retained some popularity in hi-fi circuits, but almost always
ended up playing second fiddle to larger, more powerful tetrodes and beam power tubes.
The 6V6 is rated for more than 14 watts in push-pull, fixed bias operation, which is more
than enough to drive the Whammodynes!, and should certainly suffice for the much less
efficient speaker projects that are planned.
For output, the first choice was to try a pair of Brooklyn B-15 transformers, which looked
like a good, cheap way of getting the sound from the tubes to the speakers. The circuit
needed an ultralinear output configuration to reduce the distortion in the amplfier and to
simplify the output tube hookup. After some consultation, MagneQuest would not build
the B-15 with ultralinear taps. Instead, one of the first production pairs of MagneQuest
MQ-565 output transformers was obtained. They are at about the same cost and offer
great performance for not a whole lot of money.
Ease of use of the amplifier was also important, so a cathode-bias arrangement for the
6V6 tubes is employed. Re-biasing the tubes with each tube swap was not an appealing
thought. Also, fixed bias causes even more complexity in the power supply, as a bias
supply must then be derived.
In the driver for a push-pull amplifier, some method of phase inversion must take place,
as the signals applied to the grids of the output tubes must be 180 degrees out of phase.
Various schemes for phase inversion have been developed, both active and passive. The
active methods all suffer from a voltage gain of less than unity, since they employ some
form of cathode follower output, unless an additional gain stage is employed. That
additional gain stage is usually an RC coupled amplifier, although it may be direct
coupled. The major disadvantage of this arrangement is the need to balance the phase
inverter carefully. Another possibility exists for doing the same task – a transformer
coupled stage. Antique Electronic Supply makes a PT-157 nickel core interstage
transformer specifically for driving push-pull grids from single ended output stages.
Transformers are remarkably easy to use and need no balancing. The tradeoff is limited
bandwidth, as well as saturation of the nickel core if it is overloaded with DC current.
The bandwidth limit cannot be circumvented, but the DC saturation can be prevented by
not placing any DC on the primary leads of the transformer. Thinking back to basic
electronics, just about any inductance can be replaced with an active load. Most octal
triodes are dual triode units, and hence provide can have one triode used as an active
element for loading the lower, voltage amplifier stage properly. Here, a mu-follower
configuration provides the proper loading for the 6SN7 tube chosen to be the voltage
amplifier. The output of this stage is then transformer coupled (with a DC blocking
capacitor in series with the primary lead of the interstage) to the grids of the second stage.
The PT-157 has been explored previously in VALVE articles. It is a cheap part (at under
$13 each) yet sounds surprisingly good. Placing it on an oscilloscope and subjecting it to
a square wave, however, is a little bit disappointing – there is a lot of overshoot on the
rising edge of the peak, and the frequency response is very clearly frequency limited,
even in the audio range. However, it can be significantly improved by loading the output
down with resistors. Subjectively, the transformer sounds great and the nickel core is
really fast when listening to complex passages. With the cost coupled with reasonable
performance, at least subjectively, this part looked like an ideal choice.
Now that the amplifiers’ tubes were selected, and a rough idea of operating points
located, the design of the power supply could be completed. Given the heavy choke,
large power transformer, and the complexity of the additional regulation circuitry, the
power supply was built on its own separate chassis. This arrangement permits later
substitution of either amplifiers or power supplies, should the need arise. It also allows
further physical separation of all the units to reduce magnetic coupling between
transformers wherever possible. Lastly, it allows a much smaller chassis for the whole
unit, which means already owned chassis units could be used.
For the heater supply, there were two choices – AC or DC. DC won out easily. In an
amplifier like this one, there is hardly a good reason not to use a DC supply here. The
parts count is small and it squeezes a few extra decibels of hum out of the amplifier. It
provides a low voltage to drive any solid state components on outboard amplifiers, and is
so easy to provide with the three terminal regulators that are available. Here, the circuit
uses the LM338, which is a five amp, three terminal adjustable regulator, to provide this
voltage. With a large 10 gauge steel chassis, heat sinking this part was not an difficult
task.
For the B+ power supply, the circuit was originally designed using the high gain 6SL7
dual triode in a cascode configuration as an error amplifier for each half of the 6080
series pass regulator tube (a 6AS7G would also work here.) The reference voltage was
supplied from a 0D3/VR150. It could just as easily have come from a zener or string of
zeners, but the beautiful purple glow of a VR tube is not to be missed. Zeners are also
noisy, but they do yield a much more stable reference voltage than a gas tube. However,
after building a copy of this circuit, the cascode error amplifier could not be stabilized
enough to keep the circuit from self-oscillation when the output voltage was lowered to
an operating point usable with the 6V6 tubes. As a result, the circuit was switched to a
variant of the regulated power supply available in the “GE Essential Characteristics”
book (see references at the end.) The gist of that circuit is almost identical to the one
previously used, except that the relatively complex cascode error amplifiers are replaced
with much simpler pentode based (6SJ7) ones. Working the numbers through, after the
fact, for the old and new circuits revealed operating points that were “impossible” for the
old circuit and flawless, safe operating points in the new. Power supply design, even in
the case of a regulated linear supply, can be tricky, and unlike much audio design, where
trial and error based on reasonable assumptions will often work, there is simply no escape
from pushing the pencil across the page.
Lastly, the HV supply rectification is done by high voltage, high current silicon diodes.
With a regulator in the circuit, one is simply not going to hear a sonic signature of a
vacuum regulator. Save a couple of dollars and the space on the chassis and go with solid
state. Increased reliability and lifespan of the solid state components, plus a lower
voltage drop across the rectifier comes as a bonus. A well designed regulator does not
permit the voltage to vary much from the designed value; it also serves to control much
of the ripple. Ergo, the surplus voltage afforded by the solid state diodes is useful in
providing overhead for the regulation drop. The variance in output voltage as a function
of current, or “sag” that is inherent in a tube regulator, would be masked by the series
pass element and error amplifier, unless this circuit was malfunctioning. In fact, load
regulation in this supply is on the order of 3VDC over the design range of output current,
and ripple should be less than 5mV, if good construction practice is followed. For
smoothing prior to the regulator, an LC filter configuration provides a peak DC of 90% of
the transformer’s AC voltage; a CLC configuration would yield a peak of 140%, instead.
The regulator will make up for the slightly lower ripple reduction from the filter with one
less half-section.
THE POWER SUPPLY CIRCUIT
Figure 1 shows the power supply unit. The important bits of this circuit will be discussed
by subcomponent. The circuit itself was built with point to point wiring for the HV
supply and on perfboard for the heater supply. Scraps of perfboard were used for the
voltage divisor network on the HV supply feedback loop, however, to facilitate mounting
of small PCB-style ½ W trimmers.
Figure 1 – Dual channel regulated B+ supply, 200V – 400V @ 125 mA/channel, with single 6.3 VDC @ 5A
heater supply
The simplest and easiest part to start with is the DC heater supply section, which is in the
upper left hand corner of the diagram.
The DC heater supply is filtered and regulated. The LM338 is a three terminal adjustable
regulator housed in a TO-3 package. The third terminal is the case itself, and it is the
“adjust out” line, so mounting this part must isolate its body from the chassis. Mica
washers and TO-3 mounting kits are fine, but be sure to heat sink everything well and use
silicone grease. The part is capable of five amps of current, but will generate quite a bit
of heat. Forced air is not necessary, but stuffing this part inside the chassis would be a
bad idea. This is also not a low-dropout regulator, so if you have the five volt winding on
the transformer, or can squeeze more than the requisite 6.3 VAC out of a filament
winding, it will appreciate the extra overhead. It is a linear regulator, however, so this
will burn off as heat.
The extra capacitor across the adjustment potentiometer serves to reduce ripple, and is
not required, but given the price-for-performance improvement for a 15 cent item,
include it. The two additional diodes (1N4001 or something similar) serve to protect the
regulator from the discharge path of the hefty (1000 microfarad) capacitor on the output
filter side and the much smaller ripple-reducer capacitor. The 0.01 microfarad ceramic
capacitors are a design habit. They may be left out if so desired. There is no proof that
they make a difference one way or the other, but my drawer is full of them, so they get
used a lot as extra bypass capacitors. By not using these parts, there is a risk of blowing
out the regulator under certain power-down scenarios. The regulator is not a cheap part
(about $10) and sometimes hard to come by. A simple CRC filter could be used, instead,
but would be a bit noisier than the regulated supply here, yet would still offer a major
improvement from AC heaters.
On the high voltage side of the supply, there are a couple of interesting pieces. Most of
this circuit comes almost directly from the General Electric Essential Characteristics
guide, which most tube enthusiasts should have as part of their reference library.
The input to the power transformer has a thermistor current limiting device in series with
the primary of the power transformer. This part is available from Mouser and is 10 ohms
cold, 0.1 ohms hot. It will limit the inrush current nicely, and prevent turn-on surges. A
major problem with the initial power supply design was blowing fuses prior to insertion
of this part. Part of that problem stemmed from the initial CLC filter design, which
basically presents a dead short to ground for the first half-cycle of the AC waveform,
until some charge has built up. Needless to say, some sort of current limiting is good
here. Note that the audio tubes will never have high voltage applied before they have
warmed up, because the output voltage is dependant on the regulator’s conduction, and
the 6080 is a slow tube to warm up. Also, the LM338 has current foldback, so input
surges are limited to the heaters of the audio power supply automatically, regardless of
whether the thermistor is there or not.
On the secondary side of the power transformer, there are small resistors in series with
the rectifier diodes. These help to quiet the spikes in the transformer. Before inserting
them on a hint from George Wright, the power supply diodes would routinely fail in a
catastrophic and often spectacular manner. The transformer in this project was designed
for an old tube television set, and as such, is more of a match for a tube rectifier. Solid
state rectification really needs these resistors to be happy.
The regulator feedback mechanism is pretty self-explanatory. Matching resistors or tight
tolerance items are unnecessary, although use metal film resistors wherever possible in
power circuits. They retain their value over time, unlike carbon composition resistors,
and are thus better suited to the strain and demand of power supplies. Remember, also, to
watch the voltage breakdown ratings of whatever resistors used in the project, as the raw
voltage in this supply is around 400 volts, more than many resistors can handle.
There absolutely must be a separate filament supply for the power supply tubes. This
supply has to be floated above ground; failure to do this simple action will exceed the
cathode-heater breakdown voltages for many of the tubes in this supply, and will lead to
undesirable quantities of smoke being emitted from the device. Two 1K resistors, as
shown, coupled to the 150V reference source provided by the VR tube will more than
suffice. In addition, a small capacitor across the VR tube will shunt some of the highfrequency noise inherent in these tubes to ground. Keep it less than 0.1 microfarad, or this
type of tube is prone to oscillation.
The output capacitors on the power supply (after the cathodes of the 6080, in the lower
right portion of the schematic) are simply electrolytics, of about 10 or 20 microfarads at
450V. Use a much higher rating here than the supply is worth, as the supply voltage is
variable, and component tolerances, settings, and errors in the first pass at building this
supply may result in voltage higher than a “marginal” 350V capacitor could handle
safely. Also, the cost difference is almost insignificant between the two. It is a small
price for a little safety. Do not waste expensive polypropylene capacitors prior to the
regulator – an audible difference will not be heard.
The meters in this project are surplus, PM-89 style, and are easily obtained from Antique
Electronic Supply or from All Electronics Corporation. They are wildly inaccurate, but
give a nice ballpark reading to ensure everything is healthy. An AC ammeter would also
be nice on the input side of the transformer, and might be a nice replacement for the DC
one. The DC ammeter reads >1A all the time, which, according to a Fluke multimeter, is
off by a factor of ten! The voltmeter is within about 10% of the actual value. Higher
quality meters would be a nice investment, or that part of the circuit could be left out all
together, as it is essentially set-and-forget. The meters mostly serve to point out the tube
age in the power supply. As the 6080 ages, it will likely cause the output voltage to drop,
and that should signal replacement. Do not use these meters for actual adjustment of
working voltages – set those with a high quality DMM.
On the voltmeter – 400VDC units are difficult to come by. Here a 100VDC unit is
employed and the faceplate remarked by scanning and editing it, then cutting out a highquality laser printout, and pasting it on with 3M spray adhesive. If the voltage divider is
used to provide a sample voltage for this meter, it becomes more critical to match the 47k
voltage divider resistors to avoid introducing more inaccuracies in the measurement;
close matching will keep the errors to a minimum. Use high quality resistors if any
accuracy is to be expected.
On layout, be careful to space the 6080 (or 6AS7G) well away from other components in
an area of the chassis where good air circulation can be obtained. This tube gets quite
warm and needs to remain relatively cool.
All the output voltages and grounds are run to Jones style jacks on the rear panel. Use
six-pin jacks, wired with the center two pins as ground, and ensure that, while the plugs
are polarized, excessive force in inserting them the wrong way would not produce high
voltage on the heater line, nor the converse. This actually makes it reasonably safe to
plug and unplug these items while “live”; at the least, it is not much more hazardous than
standard mains plugs on the wall.
Figure 2 - Jones socket wiring
(top view)
The power input jack to this unit is a standard three-prong unit (IEC), but features an
inline RFI/EMI suppressor. These units are available surplus for as little as $1.50, and
are a worthwhile upgrade for any and all equipment featuring this style of connector.
THE AUDIO AMPLIFIER CIRCUIT
The audio amplifier circuit is fairly straightforward for a transformer-coupled amplifier.
The schematic appears on the next page (figure 3).
Figure 3 -Transformer coupled audio amplifier
Each amplifier connects through a private “umbilical” cord terminated in a six-pin Jones
plug that mates with the jacks on the main unit. Each Jones plug is wired identically, as
illustrated in figure 2, so that the side of the power supply used is irrelevant. The jacks
on the power supply only bring out one B+ channel to each socket, and it is a different
one between left and right sides. The umbilical cords feature snap-on ferrite core chokes
obtained from All Electronics Corporation, as an added reduction measure against any
high frequency noise that might have been introduced in the cable run. The cable itself is
also shielded four-conductor Belden wire, with the shield grounded only to the power
supply chassis ground bus.
Note that the heater supply is completely isolated from the ground of this unit. This
reduces hum; the filament supply is referenced to ground in the power supply -connecting it here to the chassis would only cause a ground loop. This setup also allows
the use of an AC heater, should a low cost power supply be built, instead. However, the
“on” lamp, a blue LED, is fragile, and should be either eliminated or diode-protected
(another 1N4001 in series would do) in this case. Note, too, that substituting another
color of LED for the blue would require changing the dropping resistor value, as blue
LEDs have extremely high voltage drops across them.
As far as the audio circuit is concerned, the number one noise elimination step to be taken
is the utilization of a star ground in the construction of this unit. This technique should
be a standard construction practice, anyway. Pick a good, centrally located point and
solder the ground to the chassis, if it is possible. If the chassis is aluminum, use the
mounting screws of the terminal strips to make the connection.
Starting at the output end of the amplifier, the tubes’ operating point is set by the cathode
bias resistor. Here, it is a 280 ohm cement unit. The operating point that should be
aimed for is approximately –20V of grid bias at 315VDC B+. The 280 ohm cement
resistor has to supply the current for both output tubes. The maximum output current is
about 100mA, combined, at peak AF input voltage (for the pair of tubes). This gives a
grid bias of about –28 volts, a bit high. However, normal drive on these tubes uses much
less current, and the more reasonable zero-signal usage of 70mA across two tubes
combined leads to about –20 volts on the grids. This value is not terribly critical, and is
worth experimenting with. Note that a very large bypass capacitor could be used here –
on the order of 4700 microfarads at the appropriate voltage. This value would serve to
maintain that 20 volt bias during high power transients. A large value is necessary to
prevent the time constant of the cathode resistor-capacitor circuit from delaying the
transient response. Individual resistors, which would be a good upgrade and reduce the
need for a matched pair of tubes, would be of a much higher value and could benefit from
bypassing by smaller electrolytics. If the cathode resistors are split, these capacitors
become necessary to prevent the introduction of negative current feedback, which would
reduce output impedance and increase drive requirements. Alternatively, connect the two
cathodes together with a non-polar electrolytic.
The 27K resistors on the secondary of the interstage coupling transformer serve to load
the secondary down, as well as to load the grids of the 6V6 tubes. These resistors
improve the sound of this transformer as well as its frequency response.
The most complex part of this entire amplifier is the input stage. S.E.X. amp kit builders
might recognize the mu-follower configuration used here. In the case of this amplifier,
not only does give increased linearity of the lower stage from the active loading of the
upper stage, but the nickel core interstage now no longer needs to carry DC, as it would if
it were to provide the load for the plate. Honestly, the nickel interstage simply lacks the
impedance to be used well as a direct load of a voltage gain stage.
The following bit of theory, paraphrased from Paul Joppa, allows effective design the
input stage:
Good operating points for this configuration have Eb/Ib = 4 * rp or potentially more. Since
the interstage transformer’s impedance is fairly low, this should be kept fairly low.
rp for a 6SN7 is about 7700 ohms, so an operating point of about 150V at 5mA, per triode
section is about right. Remember that they are stacked one above the other, for a total
drop of about 300 volts.
With cathode bias, one figures that:
0.7 * Eb / mu
should provide the bias voltage. Mu is 20 for this tube, so, plugging in from above yields
5.25V of bias. With ohm’s law, that gives the cathode resistors as:
R = E/I = 5.25V / 0.005A
or 1050 ohms. 1.1K is a standard value that is close enough in a 5% tolerance.
For a mu-follower, figure on about a 20% drop at Eb across the current sensing resistor.
So, the plate resistor for the lower tube, which is really the current sensing resistor,
should have, from ohm’s law again, a value of:
(0.2)(150V) / (0.005A)
or 6000 ohms. 6.2K is again a close enough standard value to use here without worrying.
An arbitrarily picked value of 470K provides a nice grid resistance. It is possible to go as
high as 1M here, but consider the input impedance of the amplifier when choosing this
value.
The grid coupling capacitor should then be chosen based on this value. Paul’s formula
lists 30000/R, where R is 470K, or about 0.06 microfarads here. Just use 0.05, which is
close enough. The input capacitor was initially the same, although it was later replaced it
with a much nicer Solen 2.2 microfarad unit left over from the Whammodyne!
construction.
Lastly, the DC blocking, or coupling capacitor should be as large as possible to allow the
bass to really play though. At a bare minimum, it should have enough reactance to match
the transformer inductance at the load impedance, which, for the transformer used, is
about 6 kohms, as shown with the load. This translates into about 30H, or 6 kohms at
32Hz. So, the coupling capacitor should be:
C = 1 / (2 * pi * f * X)
where f is the frequency and X is the reactance. Using the supplied numbers, this works
out to 0.82 microfarads. 1 microfarad should be fine, and, finding a great surplus deal on
1 microfarad, 1 kilovolt polypropylene units, this was exactly what went into the
amplifier. Note that this value must be rated for the full power supply voltage, at least,
just to be safe.
PERFORMANCE DATA
The most important characteristic of any amplifier naturally is how it sounds. Before
heading into the subjective listening tests and commentary, here is some real performance
data on the frequency response and the spectral output of this amplifier. Measurements
were obtained through a HP 35665A Dynamic Signal Analyzer, courtesy of Dave
Dintenfass.
IInput (V) Output (V) THD % 2nd harm (dB) 3rd harm (dB) 4th harm (dB)5th harm (dB)Pwr out (W)
0.1
0.22
0.283
0.5
1
1.3
2.83
3.6
5.9
9
0.17%
0.58%
1.07%
4.38%
14.80%
-58.4
-57
-52
-50.7
-35.8
-59.5
-45
-39.6
-27.2
-16.9
-71.3
-72.5
-67.6
-56.3
0.21125
-72.4 1.0011125
-66.8
1.62
-58.9
4.35125
-37.2
10.125
Table 1 – Performance data, 1kHz sine wave input, 8 ohm resistive load
Table 1, above, presents the data that was taken in regards to the harmonics and distortion
of the amplifier. Performance is excellent at low power levels, but degrades somewhat at
higher levels. However, performance is still acceptable for the normal listening range.
Here are three possible tweaks to lower the THD at higher power levels. First, the
cathode bias resistor of the 6V6 should be lowered slightly, to raise the grid voltage,
relative to zero volts (bring it from a full-signal –29V to closer to that magical –19V.)
Second, this test used NOS Ken-Rad tubes. There was only one pair of Ken-Rads (the
other channel sports RCAs), and they are not matched. Close AC and DC matching,
especially in this common cathode resistor case, will improve performance somewhat.
Lastly, at high power levels, the circuit may exceed what the tiny nickel interstage can
handle. A heftier transformer here would drive up cost, but improve fidelity quite a bit.
On the other hand, Whammodynes!, with just one good watt of power, perform
fantastically, as long as the transients are handled gracefully. The recovery of cathode
bias certainly affords that luxury.
The harmonic data is certainly more
telling in graphic form, and Figures 4
and 5 illustrate clearly the relative
dominance of the odd order harmonics,
as is characteristic of a push-pull
amplifier. Note, however, that the
even order harmonics are not
particularly heavily suppressed – a
result that is surprising, unless one
refers to my earlier comment about not
matching the tubes well. A nonbalanced match actually benefits in this
situation, as the even order harmonics
are quite pleasing musically.
Figure 4 - 1 Watt Spectral Output,
1kHz Sine Input
In the ten watt version, above, it is
clear that the amp is beginning to
distort, and the numeric data shows
clearly that the third harmonic is only
about 17dB down
from
the
fundamental.
Frequency response is excellent for
this amplifier, which is a tribute to the
output transformers used.
The
frequency
response
of
these
transformers is actually given by the
manufacturer at full rated power, not
the low power levels that may other
companies use. As a result, the graph
of frequency response is identical for
all the power levels tried. Just one plot
which is representative is included.
Figure 5 - 10 Watt Spectral Output,
1kHz Sine Input
Figure 6, below, shows the transformer’s –3dB points are 17.8 Hz, as well as 25.0 kHz,
more than adequate for all listening, and certainly more than adequate for
Whammodynes! (good, unfortunately, only to about 30 Hz).
Figure 6 - Frequency sweep, 2 Watts output power
The scale of Figure 5 is 1 dB per vertical division, so the spike at 30 Hz is not as dramatic
as it appears on the graph – only about 2 dB above baseline, and it is centered exactly at
30 Hz.
The –1dB points for this amplifier occur at 20.1 Hz and 20.4 kHz, which are effectively
the upper and lower edges of what is considered to be the full audible spectrum.
Obviously, this statement is not the exact truth, but most speakers fall into this range,
anyway.
The damping factor for this amplifier is 0.38, which is a little lower than might be
expected. This value could be increased with some negative feedback, or, alternatively,
speakers that work well with low damping could be used instead. This low number is
indicative that this amp functions more as a current source than a voltage source, and as a
result, with the proper speakers, can have quite low loudspeaker distortion.
SUBJECTIVE PERFORMANCE
The amp itself has replaced a bi-amplified single-ended setup and is now the main
listening reference. Construction on the unit finished in April, and it has been listened to
extensively and burned it in for quite a few hours ever since. The system setup used is
either a turntable (a mish-mash of homemade upgrades to a cheap table, always in
progress) or a Marantz CD-67, feeding a Carver Silver One Reference Preamplifier (a
prototype), into these amplifiers, and out to the SuperWhammodynes!. The system is
fast, responsive, and does not lack the bass punch that push-pull configurations often
seem to suffer from. Highs seem to be a little bit suppressed. At the last meeting, Dan
agreed, noting that the signature of the regulator tube was evident in the audio, perhaps
contributing to that sound and feeling.
The amplifier was also tested as a subwoofer amplifier when paired up with a set of horns
in the Electronic Tonalities workshop. It performed admirably, bringing out a nice, full
bottom end and having plenty of power to spare driving the 96dB subwoofers against a
100dB+ horn system. In addition, it was auditioned alone through a two-driver + tweeter
vented box, using MCM’s heavier 5-inch aluminum drivers and the Whammodyne!
tweeter. The performance was similar to that of the Whammodynes! at home, it and
really shone nicely.
TIPS, TWEAKS, AND UPGRADES
There are quite a few areas of this amplifier that could be upgraded or improved upon.
For those building the setup, or inspired to build pieces of it (since its design is so
modular), here are some suggestions:
! Upgrade the interstage transformer. This piece is likely the first and foremost
limiting component in the entire amplifier signal chain, and more metal is almost
surely equivalent to better sound and power handling.
Wirewound adjustable
! Install separate cathode resistors for the finals.
potentiometers might even be better, here. Bypass them with polypropylene or
electrolytic capacitors if so desired. This eliminates the need for DC balanced
tubes to be used as finals, although AC balancing may or may not be something to
play with.
! Remove that input capacitor from the grid of the lower half of the mu-follower.
One less component in the signal chain is always better. Be careful, though, to
only use gear that is safe to couple this way.
! Adjust the connection of the primary of the interstage transformer. There are
eight possible configurations for this transformer’s hookup. Each of the following
has two locations in the circuit:
o DC blocking capacitor may be as shown, or it may be placed at the “high”
side of the transformer, prior to the audio signal passing through the core.
o The “high” side of the transformer may be connected either to pin 3 of the
6SN7GTB, for cathode-follower, low-impedance drive, or to pin 5, for
high-impedance drive.
o The “low” side of the transformer may be returned either to the cathode of
the lower half of the 6SN7GTB (pin 6, as shown), which provides a
minimal amount of feedback to the circuit, or simply connected directly to
ground.
Each of these permutations will make a difference in the quality of the audio from
this amplifier, and moving this part around is one of the most interesting
experiments to undertake.
! Employ a global feedback loop from the output of the transformer to the cathode
of the mu-follower. Either voltage or current feedback could be used to further
reduce distortion, at the expense of changing the frequency response of the circuit.
This change may or may not be advantageous, given that the low-power distortion
of this amplifier is very low.
! Replace the 6SN7GTB tubes with 6SL7GT tubes. The operating parameters of
these two tubes are different, but they will actually use the same component
values to achieve good operating points. The 6SL7 will provide somewhat more
gain than the 6SN7 used here, but the circuit should perform just as well. Again,
it is a matter of individual listening preferences. The 6SN7 is usually preferred in
listening tests, however. Miniature dual triode tubes could be used as well in the
same configuration, provided the operating points are recalculated.
! Replace the hefty tube-regulated power supply with a solid state equivalent.
There is little question that tubes look impressive glowing on the stereo rack.
However, the reality is that the supply electronics could be shrunk to a much
smaller package through the use of high-voltage Darlington pass transistors, opamp dc error amplifiers, and the like. The hefty choke might not even be
necessary – a CRC filter and good solid state regulation should do just as well (is
this heresy in VALVE?) (Nope. Ed) If there is interest, write to the e-mail
address below, and discuss some solid-state, high-voltage supplies built and used.
The look of the tubes, especially the 0B3/VR90, and the ease of visually checking
that the regulator is operating right are both advantages of tube-regulated supplies.
However, the size of the power supply is quite unwieldy, except as compared with
some of Dan’s wild amplifiers.
! Replace the LM338 with a smaller regulator and a pass transistor. More parts, yet
easier and cheaper replacement. One could even use a three terminal regulator
and two forward diode drops in the ground lead to achieve about 6.3 VDC.
LM7805 IC chips, even the five-amp versions, are significantly cheaper than the
LM338 used here.
! Replace the meters’ voltage divider chain with a single resistor and a meter rated
full voltage, if more accurate readings are desired. Ensure the resistor has a
breakdown voltage rating capable of handling the power supply maximum output
voltage.
! Polypropylene output capacitors on the output of the series pass elements of the
power supply, as well as on the decoupling resistor in the amplifiers themselves
would be a nice upgrade and may make a minor sonic difference.
! A triode-pentode switch could be installed in the amplifiers. Use a 100 ohm, five
watt resistor between the plate and the screen if this option is installed. The sound
should change quite nicely. A small, 0.01 microfarad capacitor between screen
and ground could be installed here, as well.
ACKNOWLEDGEMENTS
Much deserved thanks and a show of appreciation go to Dave Dintenfass, Paul Joppa,
George Wright, and Dan Schmalle for parts, theory, feedback, and analysis throughout
the construction of this amplifier and power supply. Without their help, it would not
have been possible.
CONCLUSIONS
This amplifier plays everything from classic jazz albums and great symphonic
performances to modern alternative rock. With such a selection, it is hard to find an
amplifier set up that can deliver in all cases, and it may make sense to keep several sets
handy for different types of music. However, this particular 6V6 amplifier works
amazingly well, and seems to be a very acceptable and pleasing compromise.
As always, comments,
[email protected].
questions,
and
suggestions
may
be
delivered
to
REFERENCES
My father was an electrical engineer and went to college in the late 50’s and early 60’s.
Most of his engineering textbooks have been more than valuable in designing vacuum
tube circuits.
systems:
The following books might prove useful to those designing similar
Gottlieb, Irving M. Regulated Power Supplies.U 4th ed. New York: TAB Books,
1992.
Horowitz, Paul, and Winfield Hill. The Art of Electronics. 2nd ed. Cambridge:
University of Cambridge Press, 1989.
Kempton, R. G., et al. Essential Characteristics. 10th ed. General Electric Co.,
1963.
Millman, Jacob. Vacuum Tube and Semiconductor Electronics. New York:
McGraw Hill, 1958.
RCA Receiving Tube Manual. Harrison: Radio Corporation of America, 1964.
Seely, Samuel. Electron Tube Circuits. 2nd ed. York: McGraw Hill, 1958.
Products by:
Allen Bradley
Alpha Core
Alps
Black Gate
Caddock
CAIG Labs
Cardas
DH Labs
Elna Cerafine
Golden Dragon
Grayhill
Hammond
Holco
Hovland MusiCaps
Hubble
Jensen
Kimber
KR Enterprise
LCR Caps
Lowther
Mills
MultiCap
Neutrik
Noble
OCSL
Ram Labs
Roederstein
Solen
Sorbothane
Sound Care
Sovtek
Svetlana
Tesla
Vampire
WBT
Western Electric
Wima
Yamamoto
and many more...
Tube Linestages
Passive Linestages
Phonostages
Single-Ended Amps
from 2.5W to 25W
High Quality High Fidelity Kits and Assembled Products
Largest Selection of Components and Accessories
The Most Informative Catalog in the Business
AC Line Filters
Mods & Modules
VISA
MasterCard
Visit Our New On-Line Store
http://www.welbornelabs.com
Celebrating 11 Years of Excellence 1988 - 1999
Our 1999 Catalog & Design Manual features loads of kits, parts, books, software, schematics, fun projects and the usual information.
It’s just $10 U.S. and Canada, $18 International, refundable. Visa and Mastercard accepted. International orders are welcome.
Welborne Labs
P.O. Box 260198, Littleton, CO 80126 USA
Phone: 303.470.6585
Fax: 303.791.5783
Brainiac's
Today's Topic: Ugly Ears (don’t ask!) - nation stack of 1.875 by 1.5 by 0.625
an interstage transformer
inch and mounting holes about 2.375
inches apart (when they’re not all bent
This unit is available for $13.95 from to heck! … don’t ask). It weighs 8 oz.
Antique Electronics, and is one of the on my kitchen scale.
small class of cheap interstage units
available. There are several in the $10 This particular unit was given to me by
to $20 range, made for old radios usu- Reid Welch; it’s the same one that apally. Nothing fancy, but their small size pears in his Glass Audio article of sevmakes for low leakage inductance even eral years ago. I’ve since given it to Tim
without interleaving and they can some- Lollar who is building an experimental
times be used with decent results. They P-P 6B4G amplifier. Jerry Cottingham
are usually designed for SE drive of has discovered that if you carefully repush-pull output tubes, but this one move the paper wrapping and expose
offers either push-pull or single-ended the secondary wires, the center tap can
drive. Generally they are not optimum be separated so that the two secondaries
for single-ended output. In my experi- are accessible. This would permit sepaence these transformers are often very rate bias voltages in push-pull, for ininaccurate in their specifications; they stance. I have not tried this, so caveat
should be measured before a circuit is emptor!
designed around them. This one is attractive because it has very good high All the data below is for single-ended
An occasional review of some nifty
...an occasional review of s frequency extension and will handle drive, ignoring the primary center tap.
chunk
of
iron
(usually
cheap!)
chunk of iron (usually chesubstantial DC current, though it is
quite demanding in drive requirements.
by
byPaul
PaulJoppa
Joppa
It is in a U-channel frame, with a lami-
Heavy
Metal
Low frequency data:
Turns ratio:
Primary:
Secondary:
DC current:
1CT:1.5CT; specified for 7k CT:15.8k CT impedance
230+248 ohm DC
402+306 ohm DC
Inductance with 12v/60Hz excitation is 7.9H at 0 mA, falling to 4H at approx. 90 mA. I would rate
it at 45 mA, where inductance is 6.8H
Peak output: 58v + 58v at 40 Hz.
High frequency data (resistance for best available square wave at 5 kHz):
push-pull:
with a 7k source and 15k + 15k loads, it looks adequate, though there’s still 1.5 cycles of ringing
around 60kHz - but the bass will be worthless, maybe -3 dB at 150 Hz (I didn’t even bother to measure it). With a 1700 ohm source, loads of 5k + 5k are needed to get a nice square wave. There’s
only a tiny amount, about 1%, of high-Q ringing at around 250kHz. In this case, bass goes down to
70Hz at -1dB, 30Hz at -3dB. The highs extend to -1dB at 45 and 60kHz, -3dB at 75 and 87kHz. By
far the best high end of the cheap units I’ve measured!
single-ended: You can’t stop the ringing at all. The best compromise was with a source impedance of 700 and a
load of 14k, but it still rings for 4 cycles at around 70kHz. Very similar results are obtained at 1700
ohms source and 20k load. The ringing is a little lower in level but lasts a bit longer if you reverse
the primary leads and load the secondary with 5k (source 1700) but that’s a pretty heavy load for
little benefit. Still, it does provide -1dB at 16kHz and -3dB at 40kHz.
This is pretty good performance, if you can adequately drive it - say, with a type 45 or a 2A3 tube! More seriously, some of
the television vertical output tubes like a 6EM7, 6DN7, or 6CK4 might work well here, or maybe a triode-wired EL84. In a
single stage, possibly a 5842/417A or a 7788 (triode-connected) might work. It could, just barely, drive a pair of 2A3’s in
push-pull to 40 Hz if the driver uses the full 45mADC capability. A 5842 at the “book” value of 150v and 26mA will run
out of steam at 33 + 33 volts peak output at 40 Hz or 58 + 58 volts at 70 Hz due to the low primary inductance.
Hurry! This offer must end December 31st, 1999!
Hurry! This offer must end December 31st, 1999!
Z-glow? ParaZen?
A Doc B. Modded Zen 84B
By Doc Bottlehead
Intro by Brad Brooks
Who else would you expect to take a perfectly great little SE pentode amplifier, and
soup it up by parafeeding it? Don’t put
anything past old Doc B. This will teach a
guy about leaving anything on Doc’s workbench unattended!
Why did he do it? Well, a number of readers asked him to—Doc aims to please.
From deep within Doc’s newly remodeled
dungeon, this editor obtained the following
drawings and photos. What’s this all about
anyway?
Check out these mods:
• C4S constant current source
in the driver stage
• Parallel-fed output stage
• Reworked the ground buss to
a single grounding point at the
input jack
• Power transformer was isolated from the chassis and the
core was grounded at the input
jack.
• Tube rectifier out, ultrafast
soft-start diode rectifiers in.
• Diode reverse recovery spike
filtering.
• Schottky Diodes in heater
supply!
Doc uses the existing output
tranny primaries as the parallel
-feed plate choke. The secondary windings are connected
as a sort of regenerative feedback loop.
In addition, the driver tube
plate is loaded up with a C4S constant current source. Not only a current source,
this little board helps the 6N1P dual-triode driver operate at close to the theoretical
operating points, as it sees effectively a
These measurements were made after
very high plate resistance. Another benemodifications were completed:
fit of using these cool C4S boards is that
they further isolate the driver from power
• Power - 2.29 watts RMS ~
supply noise.
10%THD 1KHz 8 Ohm nonWith additional components added to the
chassis, Doc helped the thermodynamic
issues by creating a “heat drain” where the
rectifier tube used to be. This, and spacing the base plate out with some hardware
helps cool the innards.
•
inductive
Hum & Noise < 20mV
Tube operating points
6N1P
200V Plate
10mA
-2V Grid
Man, I can only imagine how much of
Doc’s knuckle skin was left inside these
little cuties...
SV83
300V Plate
38mA
-10.3 Grid
How does it all sound? Try it and see! Be
sure to send us the results of your hotrodded gear.
son????
— no
Is a low driver Zout only required for its
ability to deliver drive current or are
there other reasons?
— As above, the frequency response is another consideration.
What appear to be trivial basic questions
soon seem to get pretty involved to me
and I haven’t found the real answers easy
to come by in any of the introductory
texts I have looked it. A lot of this stuff
seems to be skimmed over, an infinite
grid impedance not requiring any current!!
— yes.
Sorry if this stuff is real basic but perhaps there are others like me who get OK
results but don’t really deeply understand the issues and your responses may
help them also.
Cheers,
Bart ( not too proud to ask :-))
TWELVE WAYS TO PARAFEED
By Paul Joppa
A recent thread on the Bottlehead forum asked about using parafeed
with a cathode follower output tube (a 6V6. ed). The attempt to
describe the circuit failed, partly because words are a poor substitute
for a picture, but also because there’s lots of ways to connect a
parafeed output. In fact, there’s six ways to do it for a regular anode
load, and another six for cathode followers - a total of twelve ways!
Figure 1 shows the twelve connections. How does one choose between
them? Here are some considerations:
1.
2.
3.
4.
5.
6.
DC voltage at the transformer primary. This is an important
consideration when using transformers that are not rated for high
voltages, such as line output transformers and multifilar wound
interstage transformers. The primary may be grounded (safest), at
cathode potential (reasonable if the cathode is close to ground, but
DC coupled amps may have high cathode voltages), or at plate or
supply voltage (needs a transformer rated for high voltage).
DC voltage across the coupling capacitor. At audio frequencies, this
is the capacitor voltage. At low frequencies however, some signal
voltage may appear across the capacitor in addition to the DC
voltage. The best way to estimate this is to simulate the circuit;
however a reasonable estimate is that the signal voltage peak is no
more than the B+ voltage. The cap should be rated for the sum of
signal plus DC voltages.
In some connections the signal loop is confined to the tube, cap,
and transformer. Others may include the power supply or the
cathode bypass caps, or both. Incidentally, if the loop does not
include the cathode bypass cap, that cap can often be eliminated.
Some configurations are said to just sound better, for reasons more
subtle than the obvious circuit topology. This is a matter of
experience, and our experience with parafeed is still limited, so
listening to it yourself is still the most reliable approach.
In the cathode-follower circuits (G) through (L), the bias resistor
and bypass cap can be eliminated by choosing the choke DC
resistance to be the right value to set the bias. In all cases, a fixedbias arrangement can be used, eliminating the cathode resistor and
cap.
Generally, the plate or cathode loading choke can be replaced with
a current source provided it has enough voltage compliance.
Figure 1. Twelve Ways to Parafeed.
B+
B+
(A)
B+
Circuit
(B)
B+
(G)
B+
B+
(C)
B+
(H)
(D)
B+
(I)
A
B
C
Primary DC
voltage
Ground
Cathode
B+
Capacitor DC
voltage
B+
B+, minus cathode
Low
D
B+
B+
E
B+
B+, minus cathode
F
B+
Low
G
H
Ground
Cathode
Cathode
Low
I
J
K
B+
Cathode
Cathode
B+, minus cathode
Cathode
Small
L
Cathode
B+, minus cathode
B+
B+
(E)
B+
(J)
(F)
B+
(K)
Signal loop
includes:
cathode bypass
power supply and
cathode bypass
cathode bypass
power supply and
cathode bypass
power supply
power supply and
cathode bypass
power supply
power supply and
cathode bypass
(L)
Comment
safest
said to sound
good
said to sound
good
safest
What I Wish I Had Known
About
Dual Volume Controls
Before I Purchased...
By
Tom McDonald
cussion of volume control ergonomics. It is a kind of discussion that many people might not
have read before. I started ergonomic studies when I was a
computer programmer who
wanted to create programs that
were simple to use.
Ergonomics of Dual Volume
Controls
Over the last twenty years, every pre-amp or integrated amp
that I had ever used had a
single volume control and a
separate balance control. I was
familiar with this system and
did not want to change. But
some stereo equipment uses
dual volume controls, one for
each channel, and does not
have a balance control.
I finally purchased the Foreplay pre-amp kit with its dual
volume controls. I thought I
would hate this system, but I
loved it, for the most basic of
ergonomic reasons. So what
are ergonomics?
Ergonomics, as defined by the
latest Webster Dictionary, is:
“An applied science concerned
with designing and arranging
things so that people and things
interact most efficiently and
safely - called also human engineering.”
The rest of this article hopes to
communicate the point that the
ergonomics of dual volume
controls seems more normal,
organic, natural, logical, and
human. What follows is an industrial engineering type dis-
The words "control" and "knob"
designate the same control element, they are used interchangeably below.
If we dissect the logic of the
volume changing process, we
find that with dual volume controls the volume is raised by
turning clockwise with two
hands, one hand on each control
knob, or alternating between the
knobs with a single hand.
If the left ear hears too much
loudness (compared to the right
ear), the left arm and hand respond by turning back counterclockwise. Left side is always
left side, right side is always
right side, a clockwise turn always raises volume, a counterclockwise turn always lowers
volume. The mind only has to
think which side of the body to
respond with. The arms and
hands and ears function normally
responding to their own side of
the body. This is classic, simple
ergonomics.
Ergonomics of Stereo Volume
Control and Balance Control
But with a single stereo (dual
channel) volume control and a
separate balance control, the operator many times is trying to
turn in opposite directions and
thus can be confused.
As the operator raises both
channel's volume with a single
volume control, by turning it
clockwise, if the left ear hears
too little volume (compared to
the right ear) then he/she must
turn the balance control counterclockwise toward the left channel to raise the left side volume.
The normal clockwise rotational
direction to raise the volume has
been reversed to a counterclockwise direction by the balance control.
The ergonomic problems here
are that the arms/hands must
respond in two different directions for the same function of
raising volume. Clockwise for
the volume control and counterclockwise for the balance control.
Reversing the scenario, if the
right ear hears too little volume
(when the left ear hears correct
volume level) then the balance
control is turned clockwise
which is the opposite of the
left-ear-too-quiet scenario. So
the balance control causes confusion by operating in different
directions depending on which
ear has the too-little-volume
problem.
In addition, as the balance control raises one side's volume, the
volume of the other side is
lowered even if the operator
doesn't want to lower it. Usually
this means that the operator responds by raising or lowering the
overall volume higher and then
adjusting the balance again to get
the channels balanced at the new
higher volume level. A complex
adjustment cycling problem begins caused by the sometimesclockwise sometimes-counterclockwise operation of the balance control and the subsequent
need to readjust the volume control.
So dual volume controls are simpler to use mentally, although
they sometimes require more
physical effort, since you must
always turn two controls to raise
or lower the overall volume
level. Once a balance control is
set you may only have to use
only one control, the volume
knob, to change the overall volume level. Some sound systems
need a different balance control
setting for low, mid, and high
volume levels due to speaker performance changes at different
volume levels. In such cases the
stereo volume control and separate balance control also require
the adjustment of two knobs as
does dual separate volume controls. So the work level in this
case is similar.
Probably the simplest mental/
physical process instructions for
using the stereo volume control
and balance control combination
is as follows:
1) Turn up the volume knob
while listening for the loudest channel. (I think that it is
easier to detect which channel is loudest than which
channel is quietest.)
2) Turn the balance knob to
"send" the extra volume
away from the loudest
speaker.
3) Raise or lower volume and
adjust the "send" until the
overall volume level is
reached and each ear hears
the same volume level.
"calculate" the true volume level
by looking at two knobs is an
irritation; it is also easy to forget
to do.
A Verses B Channel Comparisons
I must admit the points made
above are subtle, but subtleties
are what high-end audio and ergonomics are all about. After
adding up many fine points a
general bias is produced of either
overall simple or overall complex operation. Even though ergonomic considerations vary it
can be generalized that thesimplest-is-the-best for most
people. Obviously, ergonomic
discussions can continue ad nauseam based on the needs, likes,
and dislikes of the operator.
Another advantage for dual volume controls is the ability to
attach different speakers to each
channel and then easily do a
speaker A verses speaker B type
sound comparison. Each speaker
has its own identity in a separate
volume knob. This system also
helps
compare component
changes in one channel verses
the other channel. This may not
sound like much of a help but in
truth it is much less work than
with a volume control and balance control.
With a stereo volume control and
a separate balance control the
operator must mentally calculate
by first viewing the stereo volume knob's position then looking
to the balance knob to see if the
channel in question is raised or
lowered by the balance knob position. When your mind is already busy checking speaker
room position, looking for reflective surfaces that may be affecting the sound, and considering
the circuit and/or part change
that you made, the added need to
But with dual volume knobs, that
is a single knob for each channel,
a single look is all that is needed
to know a channel's true volume
level
Ergonomic Summary
Overall I've found that dual volume controls are not much of a
hindrance and are actually a significant convenience in many situations.
Volume 1 - 1994 issues - $12 *** Introductory Special Until Jan. 5th - $10 ***
a Williamson amp; Dyna Stereo 70 mod bake-off; converting the Stereo 70 to 6GH8's; a QUAD system; triode input Dyna MkIII;
MkIII vertical tasting; smoothing impedance curves; Altec A7; Ampexes Nagras and ribbon mikes; Triophoni, a 6CK4 amp; audio at
the 1939 World's Fair; books for collectors and builders; V. T. vs. R. M. A. cross reference; FM tuner tube substitutions; Big Mac
attack – the MI200; 6L6 shootout; a vintage "audessey"; more FM tuner mods; vintage radio mods; Heathkit rectifiers; PAS heater
mod.
Volume 2 - 1995 issues - $15 *** Introductory Special Until Jan. 5th - $13 ***
Rectifier shootout, tube vs. solid; FM 1000 recap and meters; single ended 10 amp; triode output W- 4; Optimus 990 - speaker for
SE?; star grounds; tuner shootout; Living Stereo, vinyl or CD?; World Audio SE integrated; firin' up - smoke checking; Brook 12A
schematic; 6C33 vs. 3C33; Heathkit power transformers; 6B4's + Magnequest = SEcstasy; W5 mods; triode operating points;
Dyna restorations; Marantz 7, 8 and Scott LK150 impressions; hackable vintage gear; Quasimodo - PP 805 amp; restoring a Scott
340 in 75 minutes; a dream system for 78's; cartridges and styli for 78's; Restoring a Lowther, Part 1& 2; easy tube CD output
hack; 6ER5 phono preamp; 304TL & 450TH SE operating points; hypothetical DC ESL amps.
Volume 3 - 1996 issues - $15 *** Introductory Special Until Jan. 5th - $13 ***
Single Watt, Single Tube, Single Ended, an amp for Lowthers; the Vintage Speaker Shootout of 1996, QUAD vs. Lowther, vs. A7;
the Voigt Loudspeaker, the Single Ended eXperimenter's kit; cathode coupled SE 6AS7 amp; how to build the Superwhamodyne;
refoaming AR woofers; mesh plate tubes; rebuilding QUADS; QUAD amp filter surgery; single gain stage amps; the Brooklet, and
Brookson, choke loaded PP 6080 amps; transformer coupled PP 6DN7 amp; the Iron Maiden; Building the Lowther Club Medallion;
the TQWT, a tapered pipe enclosure.
Volume 4 - 1997 issues - $15 *** Introductory Special Until Jan. 5th - $13 ***
the Whampipe/ Hyperwhamodyne; weird interconnects; winding your own SE output transformer; Tapered Quarter Wave Tubes;
battery bias; onetuber 417A and 437A amps; DAC attack; 6BL7/ 211 SE amp; pro sound speakers at AES; 46 plate curves; what’s
all this about parallel feed?; parafeed line stage; C. W. horn divided by two; Svetlana meets Brooklyn; parallel feed SE 811A amp;
parafeed 2A3 amp; Lowther fixes; Altec vs. the competition; VSAC 97 program guide; VSAC 97 photos; Andy Bartha’s cool
speaker cables; Paul Joppa’s 6DN7 driver stage; S. E. X. kit schematic revealed; an Edgarhorn builder’s story; direct coupled active
loaded parafeed 45 amp; Brainiac’s S. E. X. changes; VSAC 97 seminar notes; tweaking the one tube 6DN7 amp, Lowther drivers,
and the Wright preamp; 300B S. E. X. amp conversion; mini monitor for 300B amps.
Volume 5 – 1998 issues - $15 *** Introductory Special Until Jan. 5th - $13 ***
Blues Master part 1; Blues Master part 2; Strapping SE amps for More Power; Ron Welborne’s Parafeed 6EM7 Amp; Cool Technical Stuff from Brainiac; Brainiac on Load Impedance and Triode Operating Points; Blues Master Corrections; Meatloaf – an ineXpensive 94dB Speaker Recipe; Hotrodding the FM-3; A Visit with Brian Sowter; Load Impedance and Operating Points – Part
2; Soul Sister, a Parafeed Line Stage; Alan Douglas – When Tube Testers Disagree; Eric Barbour SV572-10 Line Stage; Load
Impedance and Operating Points Part 3; A Turntable Odyssey, Part One; A Doc B Phono Pre(liminary), CD Treatments; Jim
Dowdy 112A Parafeed Line Stage; VSAC 98 Photos!
Available Soon
Volume 6 – 1999 issues
*** Introductory Special Until Jan. 5th! ***
*** All Volumes 1994 through 1998 on One CD-ROM! Just $55 ***
Prices do not include shipping and handling.
Please allow 3-5 weeks for delivery.
Bottlehead
Reports
This month we feature reports from the
Valve group in Washington state, and
the Tennessee Tube Enthusiasts. Rumor has it that there have been a
couple of meets in Vancouver, British
Columbia—perhaps we will be able to
include a report from them in future
columns. Also, check out the report
from the Boston area group on the
www.enjoythemusic.com web site.
Sounds like a lot of cool fun...You got
something going on? Send me your
reports and photos—Brad.
Valve Meeting, October 1999.
Acoustical Magic -- Kirkland,
Washington.
Quest, our reporter in the field,
reports:
“Here's a brief report on Last Sunday's VALVE meeting @ Acoustical Magic (N. Kirkland)
“Last Sunday the weather was cool
but sunny. As usual I was wearing
my smile on my face an joy in my
heart to attend Doc's VALVE
meeting at Acoustical Magic.
“Being the first to arrive there
(aside from A.M.'s Gil and
Dorothy (Dorothy Harwood,
owner) who were there to open up
the shop at 11:00am) I swiftly
introduced myself and took a short
break in their largest showroom,
listening to their demo gear with
my own software. Shortly afterwards Doc showed up in his little
pickup truck with a blue tarp over
some "secret" stuff . Underneath
was Doc's newly made Straight 8
with blackish "fakestone" speckled
finishing on it. According to Doc
he had these speakers finished last
night and it hasn't been completely
"run in" yet. He was confident that
the sound will get better as the
days went by...and I believed him.
Along with his gizmos, tools,
Foreplay and ParaGlows he also
bought a pair of KR 2A3 along to
play with.
“John "Smoothplate" Tucker
came along with his Sovtek 6B4G
version of ParaGlow and his usual
"bag of tricks": his own power
supply, DAC, preamp, wiring,
etc...”
[Quest reports that typical Valve
meeting activities went on: Coffee drinking, swapping around
various combos of DIY gear and
listening to your favorite tunes.]
Quest continues…”Among us sat
a sneaky little guy that didn't introduce himself until the midst of
the listening session... Guess
who? It's Vinylly! He silently
brought over his Thailand-made
all chromed 300B SET amp to
audition: This baby consists of
single 300B per channel, driven
by 6SN7s and some small MT
tubes (ECG6189W you say??).
We played some Jazz vocals and it
struck me with the sound that is
typical with 300Bs: smooth, nonfatiguing soundstage with a strong
authoritative sense in vocal presentation -- female vocals came
out effortlessly with the finesse
that is known mainly to DHT
Single-Ended configuration.
“Halfway into our fun-filling sessions came Hopper accompanied
by Doc with 2 white boxes of
unknown content, mumbled
something like "...quick! stick'em
in while they are still warm!!" "
What was that?" I wondered. As
soon as Hopper pulled out a tube
for that box and proclaimed
"...VV....2A3." Did I ever noticed
that, and everybody was stunned
just by the looks of it! Physically
speaking those VV 2A3 looks
more like a VV-52 to us. The
glass envelope is twice the height
of the typical ST tube and it's one
and a half time bigger too. Of
course the contents inside is proportionally bigger as well. Doc
swiftly pulled out his KRs and put
in those big Muthas. Wow! exclaimed Richard: "....these 2A3
sounded very good indeed!"…
[I am excited myself to hear these
2A3s as well…there seems to be
no lack of good new 2A3s out
there for every budget…]
“Thanks to Acoustical Magic for
their place and various treats, Doc
and Smoothplate's various gears,
Vinylly for his super 300B SET
(boy that baby sure is heavy!!).
Best of all, Kudos to Hopper for
his VV 2A3. Those tubes are a
blast!!
“We had so much fun that afternoon. I always leave Seattle with
a smile of content in my face :^>”
Tennessee Tube Enthusiasts,
October 1999. Phil Sieg’s house.
Knoxville, Tennessee.
Phil’s got the scoop:
“Well, it's really the Tennessee
Tube Two right now. Brad Brooks
drove over to knoxpatch this
weekend with his stock Foreplay
and ParaSEX. We listened to the
T-1s for a while with the modified
Atmas driving full range - the
baseball playoffs are interfering
with finishing the 45s - so Brad
could get a feel for the "sound" of
the system.
Then we plugged in the Foreplay
and ParaSEX monoblocks. Everything else stayed the same. First, 5
parafed Watts will drive the T-1s
full range with relative ease. There
was less extension and air on top,
but the tonal quality of the
midrange was right there. Lester's
clarinet and Roy's trumpet
(Laughin' to Keep from Cryin')
had authority and body, and that
"bright" Eldridge sound came
across faithfully. Bass was not as
deep as with the Atmas, but it was
taut and well-controlled. Duke's
piano (This One's for Blanton)
didn't have quite the sonorous
body as it did with the Atmas, but
the attack and decay of the notes
was all there. Same with Ray
Brown's plucked bass.
So what does $500 worth of electronics driving a pair of $13K
speakers prove? First and foremost
is that circuit topology is by far
and away the most important factor in design of an amplification
stage. Tube rollin' and boutique
parts will only get you so far if the
topo is screwed up. Second, there's
a whole lot of pricey gear out there
that would slink off, head hung in
shame, when compared to this little combo.
Did they best the Atmas? Well, no
(and I think Brad agrees). But the
Atmas are an exceptional pair of
amps. However, my sonic memory
tells me that the combo is every bit
as good where it counts as my old
ARC VT-150SEs (and I cringe
inwardly as I think about the price
tag on those).
[Phil and his wife Tina took me
out to a well known pub in
Knoxville near the college. Never
have I ever seen so many people
wearing orange, the team colors
of UT, and it wasn’t even game
day! Man, this town sure loves it’s
football…]
Brad took some pictures, including one of my work-in-progress
45s, which will eventually show
up in VALVE. And he shared his
thoughts about what was next for
his Foreplay/ParaSEX combo, but
I'll let him talk about that. We
talked about other designs (we
have a shared interest in the KR
PX-25 and talked about building
ParaGlows that could accommodate a 2A3 and the PX-25). And
we discussed how we can grow the
Tennessee Tube Two. He going to
try to set up something at Nicholson's Hi-Fi in Nashville soon.
Then maybe something in East
Tennessee again before Christmas. All-in-all, a very good day
with good ideas for the future.
Anyone in the area reading this
who would like to participate,
please e-mail.”
[Sorry bottleheads, no pictures of
the meet this time, but I promise
to put them in the next issue.]
Valve Meeting, November 1999.
Rancho Tonalities -- Poulsbo
Washington.
Our tireless reporter Quest sends
this report:
“This time instead of driving to
Rancho Tonalities alone by myself
I dragged TonyD. along with me.
We managed to hit the Edmonds
ferry terminal for the 10:50am
ferry and arrived around 11:45am,
and before I reached into that evil
Dungeon door I was so familiar
with I heard some laughs and
music pouring out from the basement....It wasn't until I walked
down the stairs that I realized that
there was a huge gathering this
time.
“Eager to introduce myself I
waded through the crowd and
talked to many other guys on the
show: Paul Joppa, Cameron, Doc,
Tim, "Willy the squid" Tony , Ed,
John, Tony Glynn(Yes! Tony
Glynn of Lowther America was
here!!),Aaron and many others
around. Looking straight down
the end of the basement I saw the
already broken-in straight-8. It
sounded way smoother than our
first encounterings in Acoustical
Magic. Tony immediately pointed
out that the sound is much more
coherent than what he thought it
would be, and the bass? plenty for
all of us. We sat down while guys
took turns sitting on the "golden"
spot. From a distance I saw Doc's
B-Glow playing with John's eXception and his own DAC + a
Marantz CDP(?) playing music.
This time we opt for mostly vocal
stuff: Ella Fitzgerald, my little
favourite Charlie Hayden, etc.
The sound was nothing short of
stunning as I would say…”
[Quest describes the goings-on
that are typical of a Valve meeting: Laughing, playing music,
laughing, swapping out the latest
DIY wonders, listening to music.
And, as usual at a Valve meeting,
a little surprise… ]
“While 1/2 of the audionuts
stucked to their chair on Doc's
basement,
others
secretly
"sneaked" upstairs to take a look
at Doc's secret weapon: the Beveredge speaker system. This odd
looking speaker comes as 2 planar
structure and 2 "coffee-table" like
subwoofer. According to Doc the
array of ribbons on the planar
structure plays down to 80Hz before the subwoofer takes over.
With built-in amp (some sort of
SEPP using 36KD6?) on each planar structure base all we could see
is the RCA interconnect wirings
and that's all. When Doc played
some music through the system it
sounded stunning! A full-range
system that has the power to rock
the house and and clarity is everything you'll always wanted, and
more.....
“We hanged around Doc's upper
level as the rest of the people
carefully sneaked into the basement: Lehman(?), a friend of
Wendell in Japan showed up for a
bit of fun, and so did Jay. They
played Ry Cooder and V.M.
Bhatt's "A Meeting by the River"
with steel guitar and sitar and the
sound was just marvelous: Ry's
steel guitar has all the body and
weight that is appropriate to the
soundstage, without being overly
"dry" nor "shrill" and definitely
not edgy at all. With V.M.'s complement into the sonic picture everything was just perfect. (If it's
because I have to drive home later
that day I would have ask doc fo
a glass of wine :^>
“Queen Eileen was kind enough
to prepare some "num-nums" for
us (big applause for Eileen!!!) and
brought home a big pot of Starbucks coffee. While the folks upstairs chomping down the dip and
the spread we talked about many
other things in the audio industry,
people and dogs (Doc just brought
home a big dog ...it is a blend
between boarder-collie and ???)
[Pony, Perhaps?]
“Later with Tony Glynn and a few
others in the kitchen area some of
us went down to the basement
afterwards and discovered that the
crowd now played Charlie Hunter
and Leon Parker's Duo, with
Aaron sitting in the "golden chair"
quietly listening to the tone and
the color produced by the audio
picture. Lehman (sorry bud for I
might have spelled your name
wrong) started talking about his
recent problem with the hum from
his SE-1. Eager to jump into the
bunch I followed him and John
Tucker to Lehman's bug (yes! a
classical 1974 bug!) , picked up
his amp and later John took off the
bottom plate and took a quick look
into it. He advised Lehman to have
his amp reworked and instead of
using 6SN7 try something else .....
“As dusk came little by little some
of us left Rancho Tonalities and
went home. Instead of intensive
audio session this meeting was sort
of casual gathering in nature:
warm, friendly folks such as Ed.
John Tucker, "Willy the Squid's
warm smile, Richard and many
others make us feel like one big
family. (TonyD: are you coming
next time :-)
“Kudos to Doc and many other
favourite guest like Tony Glynn,
Paul Joppa and John Tucker,
Cameron and Richard, Jay and
TonyD., Tim and Ed, "Willy da
Squid" and Lehman (I hope your
hum problem will be over soon).
“Also a big thanks to Queen
Eileen's dip and other snacks!
Take care!”
Well, that’s all of the club reports and
going’s on for this issue. I look
forward to future reports, with photos
of all of the fine projects and smiling
faces enjoying everyone’s company
and the cool music. Send them to me
via my e-mail:
[email protected]
Until next issue...
Entwined
interconnect cable kits from Electronic Tonalities
pre-cut twisted pair 20.5 guage High purity continuous cast long crystal solid copper
conductors with a high temperature insulative coating, pre-cut polyethylene outer jacket
and Vampire Wire RCA plugs
$39 .5 per meter pair, $49 per 1 meter pair, $59 per 2 meter pair