On 13 Jun 2004 23:23:05 GMT, ohawker (Andy
Evans) wrote:

Some discussion of voicing DIY tube amps:

Rumor has it that the Scott company brought professional musicians in house for
auditions when they were voicing the 299B.

I doubt VAC will supply us with their schematics. Besides, their forte is
really in the extensive voicing with parts selection, listening.

I wouldn't bother the -3dB point, just listen. Sometimes for 100ohm cathode
resistor 100uF bypass cap is fine, sometimes 220 and sometimes the 470uF, even
if you use the same time of cap. Use cap value as amp voicing method. If your
amp sound a bit too thin, you may tweak it by a larger cathode cap etc.

(from designer Joe Curcio) Many of us (me included) have wasted time and money
blindly combining the "acclaimed best" of similarly flavored components only to
become dissatisfied with our efforts. And let me add that the power supply
choices, from line filtering thru rectification, DC filtering, bypassing,
regulation, and supply & ground returns, each contribute as much or more to the
recipe. It's easy to see that with all of these contributing elements, it is
very easy to get lost in the spectra of system voicing that can range from
bright and irritating to dull and lifeless.

It's the system synergy thing, one more time. It seems to me that power amps
with "neutral" sonic signatures, like the H/K Cit 2. and AVA FET-Valve that I
own, should be mated with line stages that are SLIGHTLY euphonic (warm) in
character. Net gain should be LOW as the "brutes" are fairly sensitive. I'm
thinking in terms of a 12BH7 (6GU7) in SRPP feeding a 3:1 step down trafo. The
trafo at the O/P reduces net gain, lowers O/P impedance, increases O/P current,
makes the unit non-inverting, and breaks up ground loops. I might as well go
"whole hog" and choke load the SRPP. The opinions of other "inmates" would be
appreciated.

I generally agree with your voicing suggestion but why on earth SRPP?

My Pentode PP amp does not like teflon caps in my preamp, they make the amp
sound thin and edgy. With CCS on the output it got even worse. This is the
complete opposite of my experience with these caps with single ended amps. I
think voicing is very big with an amp this transparent. I now use Jupiter caps
in my preamp - sound very good, in fact about the best I have heard.

I've recently been experimenting with the capacitors in a phono stage. It's a
relatively simple ciruit and has one coupling capacitor and one output
capacitor.

I put Auricaps in both locations and got nice deatail at the expense of some
midrange sweetness. I put Jupiter caps in both locations and got nice sweetness
at the expense of some detail.

I have no place for electrolytics in my amps (with the possible exception of
Blackgates or Elnas) except when absolutely necessary, such as DC filter
supplies for DHTs. Even in DC filament supplies I use Blackgates or Elnas,
Schottky diodes, nice filament chokes and no resistors. I don't know about
mixing caps as you suggest, they might act as mutual filters. You might try
bypassing a Jupiter with one of the small value Russian surplus teflons, that
might give you the warmth and detail you desire. I'm not usually a fan of
bypassing caps in the signal path, but I've heard a couple of amps that sound
great this way. I know of a couple of builders that swear by bypassing a warm
cap like a Jupiter with a teflon. I do like bypassing caps in the power supply,
like an oil with a Solen, and bypassing the Solen with a teflon. I've been
voicing my system for years now.

Hi Ken - can linearity simply be determined the 'old way' by examining the
inverting stage's plus and minus voltage ac-loadline swings around the
operating point on the plate characteristic curves? - is there any point
otherwise besides trial and error to get a specific "voicing"?

I think the Foreplay is the most bang for the buck in the audio world. The nice
thing about building it stock, and adding upgrades later is, that you learn how
certain mods effect the sound. This really helps with voicing projects on down
the line, which I'm pretty sure will happen

I built a fairly straight-forward PP 6B4G amp using Electraprint autoformers to
split phase, based on inmate Scholl's work. 5842 driver. Fixed bias on the
6B4's, LCLC power supply. Vintage Dynaco P-782 power trans, scavanged Sansui
1000A output iron. VERY good sound; I took them to a Chicago Bottlehead
meeting, and they embarrassed some pretty expensive competition. That said, it
has a couple of weaknesses. Coupling cap is critical to voicing the amp. Also,
I always had a bit too much "lushness" in the mid-bass. I ended up putting in
Electraprint interstage transformers; sounds better to my ears (tho I miss that
false "lush").

I drove a dual 211 in long tailed pair, PSRejection mode for some years using a
6SN7 pair as Mu Follower and listened to fast and clean sound.
I then upgraded it to a 2A3 working into a Tamura NC14 IST, 5K/5K. The change
in sonics was tremendous. The interesting bit was that I ran this on the right
channel and just the 2A3 with 5K OPT on the left to ensure that the voicing was
the same. I haven't found anyone who could tell the difference yet.
Upgrading the fixed bias to 195 volts runs an 845 quite nicely except that I
did not like the sound and went back to the 211. Your earlier comments re
drivers not being as good as the real sound of a 300B stand up very well. Once
you get the driver stage right, big drive, substantial current, the tube sound
vanishes and the power tube becomes not much more than an impedance buffer with
a small sonic contribution.

I've used Rikens, carbon comps like the Allen Bradleys, cheapie Radio Shack
carbon films, Philips and NTE metal films, and 'no-name' metal oxides in my DIY
amps.

My personal preferences for resistors are in the order above. The Rikens are
best IMO, with good detail yet a nice warm sound, carbon comps and the RS
carbon films are about equal, the Philips and NTE metals are okay but too
"cold" sounding for many circuits (though they are nicely detailed), and
'no-name' metal oxides are completely unacceptable due to their very cold and
grainy character. IMHO the most important resistors in common cathode circuits
are the plate load resistors and the gridstoppers. For a nice warm sound, I'd
go carbon (Riken, RS carbon film, or carbon comp) there. However, you can try
the Philips/NTE metal films as plate load resistors if you want a more detailed
sound. I will admit to using a pair of 150 ohm Philips ECG metal films as plate
loads at top of the cascode in my 6DJ8/6922 amp to get a little more detail, so
they are indeed useable in voicing an amp that's a little too warm and fuzzy.

In trying to address the limitations of the original Joe Curcio designed Daniel
and preserve and improve on its strength we tried many tube (and solid state
for that matter) "gain blocks" including, SRPP, cascade, cascode, MU Follower,
differential amplifier, as well as other more unconventional topologies and
catalogued their signature "voicing" colorations. Many circuits whose
theoretical and measured performance are exceptional failed to sound musical
and realistic losing depth and skewing spectral balance. Some sounded outright
harsh; similar to second generation solid state circuitry. A few were
outstanding in one area (i.e. dynamics) while mediocre in another (i.e. image).
As a result, we have developed a portfolio of "select" topologies that are
applied to meet the requirements. We have realized excellent results with anode
followers, optimized cascodes, and a unique triode / Junction FET hybrid
deployed in our Daniel II reference preamplifier. It's measured specifications
are unparalleled but more importantly, it provides a complete sense of realism
both in spectral balance and holographic imaging.




=== Andy Evans ===
Visit our Website:- http://www.artsandmedia.com
Audio, music and health pages and interesting links.

The moral of this tale is that if you want to make sure nobody can
disagree with you, just say stuff that is devoid of any meaning.

d
Pearce Consulting
http://www.pearce.uk.com