• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Push-pull 807 amplifier without global NFB

P-P pentodes need feedback to work well. Ultralinear or triode strapping is a form of feedback. Local instead of global, but feedback nonetheless. There are complete feedback-o-phobes, and global feedback-o-phobes. The first kind balk at triode strapped EL34’s, which can sound quite good even in P-P with no GNFB. Make no mistake about it - there IS feedback.

A low power SE pentode (with zero feedback) can have a pleasing sound, even if it lacks “realism”. An old 50C5 output stage in an old radio or phonograph. They don’t sound “bad” at all. But go to Push pull without any feedback and heaven help you, unless you’re playing an electric guitar thru it. Might be “ok” if you bias it up to class A, but nobody ever did that because the whole idea was to get the improved power/efficiency.
Class A triode PP is exactly what I am suggesting here. It offers all the benefits of SE but eliminates many of the issues. It's the base assumption that PP amps can only be built with pentodes which is the problem here. Vacuum States PP amps are an exercise in what can be achieved if you reject that assumption. The sort of logic that is applied to SE designs is equally applicable to PP designs it's just that in the early days of valves power output was always the driving factor.

Shoog
 
Last edited:
The amount of global negative feedback is 12dB. Did I calculate correctly? The open loop gain is 12dB higher than the closed loop gain.
What I measured is that the THD is about 0.2% up to the point where visible clipping occurs. But the main argument for GNFB is the increased DF (=lower output impedance) that helps damping the bass resonance of my speakers.
 
If you wish to use modern multiway speakers then you need to get the amp output impedance down to say at least 1/3 of the speaker impedance. If you don't then you get a bass hump when the bass cone resonates. Pentode mode is high impedance, UL is better, UL with NFB better still. I think its probably this that changes the sound rather than the THD.
 
Every speaker gives a bass hump at resonance - multi-way or not. The worst “offenders” are speakers with a high Qms. The higher it is, the larger and narrower the Z peak. Mechanical damping tends to reduce it, but Qms of less that about 2 isn’t very common in woofers. And not in high efficiency ones for sure.

Multi-way speakers have more than one impedance peak(or the two you get with a reflex box) - you usually get a broad impedance peak at or near each crossover frequency. And multi driver (multi-midrange especially) have impedance dips often as low as 2 ohms. Overall, this adds up to coloration in more than just the bass when driven from high Z sources.
 
The main reason why i would use gNF in this case is because the 807 has a poor distortion profile with lots of higher order harmonics. Without feedback I would expect them to sound quite nasty.

Shoog
The 6L6-oids definitely do make ESSSSSS-loads of h5+ While developing the Vixen design, I ran it open loop. The Twin-T test showed a residual that looked like a sawtooth at three times the fundamental frequency (1100Hz). Listening tests showed no obvious distortion, but there was something as irritating as finger nails on a black board. It was worse with some program materials, but always there. 807s def need help in the form of lNFB and gNFB. The designer of the type (Schade) recommended feeding back 10% of the plate signal to the grids (~7dbv local NFB) gNFB finishes the clean-up.

Zero feedback Push Pull amps can be the best of both world - but there are almost no examples of them around so most people assume that all PP amps have to have loads of feedback to work - which is a very wrong assumption.

Shoog
How about this (attached)? I wondered for quite awhile why the designer selected 6BQ6s for the finals in this no gNFB plate modulator. You'd expect to see 6L6s for this power class. The 6BQ6-oids have no audio pedigree; there is no mention of any use as an audio final in the spec sheet. After doing some audio loadlines, it still wasn't obvious. The estimated THD came in at about the same level as the 6V6-oids. Anyway, I decided to do the Le Renard design based off this modulator project.

Then it bacame obvious: open loop, the Twin-T test showed a residual 3X the fundamental, its ampulitde was higher than what I saw with 807s, but it was a slightly distorted sine wave, suggesting mostly h3 with a trace of h5. Listening open loop produced no pentode nastiness, just an overly "aggressive" sound that gNFB tamed easily. The no NFB plate modulator woukld sound good enough for ham work, and without all that harmonic distortion going over the air. You can consider the 6BQ6 (and likely other TV HD finals) as 6V6s with 2X+ the output.

.
A low power SE pentode (with zero feedback) can have a pleasing sound, even if it lacks “realism”. An old 50C5 output stage in an old radio or phonograph. They don’t sound “bad” at all.
Interesting you should mention that. The spec sheet for the 50C5 gives two sets of plate characteristics: one for the most common uses you mentioned, and a second with V2K= 90V that are more linear. The manufacturer didn't want their 50C5s blamed for how gawdawful most things that used them sounded
 

Attachments

  • mod_2x6bq6_2.jpg
    mod_2x6bq6_2.jpg
    248.8 KB · Views: 217
Interesting you should mention that. The spec sheet for the 50C5 gives two sets of plate characteristics: one for the most common uses you mentioned, and a second with V2K= 90V that are more linear. The manufacturer didn't want their 50C5s blamed for how gawdawful most things that used them sounded
I never though they sounded particularly “bad” when used at the 1-watt level. But I did like the 50EH5 better. The g2 probably sat at 90 in most of those apps, since the plate might have been 100 or 110. If you go cranking the power you are better off with the bigger bottle 50L6.

Interesting that 6BQ6 modulator runs the g2 at 150. It does seem to be a sweet spot for running sweep tubes as audio finals - at half the B+. It can make for a very simple power supply - just need a CT power trafo and full wave bridge. I’ve run really hot cathode types (DQ5) down at 120 even (regulating it down a little further) and gotten good sound out of them. What I’ve found is that when you don’t need as much current as you could potentially get you run the screen down and it has better sound - especially less crossover distortion. With 6L6 oids and their cold 0.9A cathode, you need every milliampere you can squeeze out of them by running the screen up at B+. Then comes all that high order IMD.
 
Very few people build zero feedback 300B PP amps, but I think if they did and compared them to a 300B SE amp - the PP would win every single time.
People will tolerate very low power in SE designs but never think to take advantage of low power Triode PP amps.

My first serious amp design was a 807 PP Tabor clone using 6AU6 drivers and high levels of Schade feedback. I used a VR tube to supply the screens. It sounded OK and I still have it boxed up in storage in my workshed. Might pull it out and have a listen some time. I moved onto similar designs to the 6BQ6 amp you shared but using 6080 as the finals. That really was a fine sounding amp and I seriously am thinking about rebuilding it.
 
Last edited:
  • Like
Reactions: wirewiggler