What’s fun about knowing Don and Thom is they have different ideas that bounce off each other. Way back when, Thom had some pretty gruesome-sounding amplifiers, and in my usual blunt and ham-fisted way, I said as much. I may be the son of a diplomat, but I’m not always diplomatic when it comes to audio.
Hey Thom, try this, and showed him the original Reichert SET schematic. You can skip the DC heating, because the one I reviewed for PF magazine had an AC balance pot and it sounded fine. You can improve it further by using TWO separate B+ supplies, one for the input+driver and the second for the output tubes. And use damper diodes instead of the usual 5AR4 or 5U4. And if you really want to get fancy, use LC coupling instead of RC coupling for driver/output interface. You just need a big audio-grade choke (not power-supply drek) and a decent coupling capacitor. Plus, you’ve got two good Magnequest single-ended output transformers laying around, why not put them to work?
Thom built each amp on a wood plank ... literal breadboards ... and they wiped out the overpriced audiophile amplifiers (no, I’m not telling which one). Thom sold the audiophile amps and never looked back, improving the prototypes over the years, until they became the NiWatt amplifiers they are now, many many refinements later. I’d give hints every now and then, sometimes in a heavy-handed way, but every improvement took it further, and 90% to 95% of the improvements came from Thom, who has a very sharp ear. A win-win, from my perspective.
And it’s a lot of fun working with Don, who has the nerve to restore Citation amplifiers, which are notorious in the industry as one of the most complex and eccentric designs of the Golden Age. Stu Hegeman was seriously out there, compared to everyone else. I would feel faint just looking at the underside of those things, never mind trying to correlate the schematic with the maze of point-to-point wiring used back then. It would be like troubleshooting a 1963 RCA color TV with 28 tubes and weird, complex setup procedures. Complete respect on my part.
I’m frankly amazed Don took on my amps. They are NOT for beginners, with demanding layout requirements, and no relationship to Golden Age amps at all. There are no points of familiarity except to seriously obscure amps of the 1930’s. One of the dirty secrets of the high-end biz is that most medium to high-power PP amps are warmed-over Golden Age designs, with a few extra regulators or cathode followers thrown in here and there. Full respect to Don for even looking at the schematics. Back when I started this in the late Nineties, I had people laugh in my face when I told them what I was doing. The big thing then was single-ended EVERYTHING, and as a speaker designer, that just felt wrong.
For one thing, tubes (and transistors) work by varying their resistance. That’s all they do, there’s no little elves inside that give them special properties. With a SE amp, that varying resistance (powered by B+) drives an output transformer, which faithfully reflects it down to the speaker. The transformer is fully passive and has no special properties; it multiplies current and divides voltage, about 28:1 or so. (The current/voltage multiplication is the same as the turns ratio, but the impedance ratio is the SQUARE of the turns ratio.)
I don’t feel comfortable about the speaker being driven by a varying impedance ... that means the damping factor is merely an average value over time; at any one instant, it can be anywhere. If I want a steady, constant impedance, what can I do?
A unique property of Class A push-pull is nearly exact symmetry between the two tubes. When the resistance of one tube goes up, the other goes down, with a precision of a percent or so. It’s a exact see-saw action, and it is unique to Class A push-pull triode. The curved grid lines you see in tube manuals (for SE circuits) nearly exactly cancel out, leaving parallel lines.
As far as the speaker is concerned, it is being driven by a low-value resistor, not a wildly varying source impedance, thanks to the precise complementary action of the two tubes. The biggest mismatch you are likely to see is 5% or so.
Is this true of Class AB? No. When one tube cuts off, which is typically around 2 to 5 watts, we’re back in the single-ended situation again. A Class AB amplifier has three regions of operation: upper tube ON, both tubes at once, and lower tube ON. Only in the small central region is there true Class A operation. If you are not careful, the sharp cutoff associated with Class AB transitions can even generate ringing in the output transformer.
It gets worse with pentodes. The grid lines of pentodes have higher-order curvature, so the complementary action does not fully cancel, so the summed grid lines are wavy and not straight. This was understood when pentodes replaced triodes in the late Thirties and Forties, but feedback was always used to straighten out the mess. But it was controversial at the time, with triode fans holding on to their beloved 45’s and 2A3’s (300B’s were not for sale to the public, and were never used in consumer electronics).
So there is only one class of circuit that has a constant, unvarying output impedance: Class A push-pull triode. All the rest (tube or transistor) require feedback to synthesize that impedance. In fact, of the famous Western Electric 300B theater amplifiers revered by Japanese collectors, only the single-ended model (the 91A) uses feedback. The two push-pull 300B models do NOT use feedback; Western Electric used a very unusual circuit called the Harmonic Balancer, which had been completely forgotten by the Fifties, and in following years. It wasn’t re-discovered until John Atwood and I referenced it in Vacuum Tube Valley magazine in the late Nineties, sixty years later.
This is why I never joined the SET bandwagon, but wasn’t interested in Golden Age pentode amplifiers, either. (Yes, you in back, I see you raising your hand. The paired 6V6 beam tetrodes in the driver stage of our amplifier are triode-connected, which gives a very close approximation to a 45 triode. Plus, the 6V6 pair remain in Class A push-pull under all conditions, including heavy clipping. The interstage transformer re-balances and sums the error terms before the drive signal reaches the 300B grids.)
Don mentioned the subtler aspects of the Karna Mk II’s, or Statement 300B amplifiers. This is just one of them.
