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.
It's the split-load design that was popularised by McIntosh. I believe Audio Research are still keen on it. What the diagram doesn't show is that you need an awful lot of swing from the driver stage. A common way of solving that problem is to use the output anodes as bootstrapped HT supplies for the drivers stage. It works, but it makes stability very much harder to achieve.
Good point. As the HT voltage rises, the current through the anode load resistor rises, so it appears to the valve as a reduced value of anode load resistor, causing a steeper loadline and increased distortion. There's just no free lunch...
While I am not able to argue your assertion in the electrical sense, this article is from the Proceedings of the IRE July 1954 while the McIntosh circuit was published as least as early as 1949. This would seem to make it unlikely they are the same. Though I have difficulty deciphering the schematics in the McIntosh paper I do not see the capacitive coupling of the cathode windings with the plate windings as shown in the scheme I posted. True? Are there other differences? Enlighten me.
The capacitors allow the transformer to be made without bifilar winding as in the McIntosh design. The caps provide the close coupling between sections to overcome xfmr leakage reactance. The text refers to this and that it would be cheaper to build.
It's exactly the McIntosh circuit except for the 'bridging' caps. Both McIntosh and Amemiya were interested in reducing crossover distortion in class B amps by AC coupling the cathode winding of one tube to the plate winding of the other. Mac did it by specifying bifilar windings and Amemiya used a cap and a presumably cheaper transformer.
In any case, it's not clear what the tight coupling buys you in class A. In other words, a class A amp without bifilar windings might work equally well with and without Amemiya's cap.
What could be some advantages in Class A operation? Better high frequency response with simple transformers? Since you are driving the difference between the grid and the cathode rather than the grid and ground is a greater output swing available without reaching into the grid current region? Mosfet source follower seems more complicated than a cap. Are there significant advantages or is it just way of doing things?
Is Class AB with quality to rival Class A available with proper tube selection?
I haven't seen any proof, whether analytical or controlled listening tests, that class A is superior to properly designed AB.
The unity coupled output stage, whether used in A or AB, has a much lower source impedance (hence better damping) than the same tubes used as pentodes, UL, or triodes. Distortion is also considerably lower.
MOSFETs are just a different way of doing things. There may be some practical advantages, but I haven't built anything like that so can't really say what those advantages might be.
I think both circuits are 2 x gain of cathode follower. I don't know where the term unity gain got started referring to the McIntosh. You can get even higher gain by using a lower turn winding on the cathode section than the plate section. See two posts up, about 3 x gain using UL taps on xfmr for cathode connections.
I don't believe the Quad II was bifilar wound or had the caps, but maybe someone knows for sure. Might have been even lower than 20%, maybe 15%. Probably most DHTs have low Gm compared to the pentodes, so will require even higher driving voltages. You will probably need a higher impedance primary for the DHTs too, making drive voltages worse yet.
We had our submersible well pump changed recently, and the guy said he had worked on some wells where they held him by his feet while he worked in a 3 foot diameter hole in the ground. This might be a suitable environment for this DHT undertaking! But always a first time.
Heh, heh, are you suggesting I throw the DHT idea down a deep hole? I was just asking, and while I'm at it what about a direct heated pentode?
As a matter of interest, why would the DHT require a higher primary impedance?
Don't think the Quad had caps, don't know about the bifilar, doubt it though. I have a Bell amp that has a tertiary winding for the cathodes. And come to think of it I modified a Stromberg-Carlson for cathode load by bringing the secondaries back to the cathodes, original ground side and 16 Ohm taps with 4 Ohm tap now grounded. In fact I've been using it for my midrange amp for the last week or two and I think it sounds pretty good. It uses 6N7s as dedicated drivers and I think they can handle the extra drive required, but I have never measured the amp.
RDH4 seems to think all tube types and all classes of operation benefit from unity coupling and/or balanced load, whatever you want to call it.
"are you suggesting I throw the DHT idea down a deep hole?"
Don't let me discourage you. I was just being a little off the wall.
I'm sure this could be made to work with the lower % cathode feedback levels. Just that at 50%, like the McIntosh, with all the other factors weighing somewhat against also, might be a "commitable" offense. I think Patrick Turner recommends 12% to 15% or 20% for cathode windings. He does a lot with this type of design. I would strongly recommend taking a tour through his extensive web site.
I'm no expert on this balanced load area or DHTs either. Just my impression that most of the DHTs or DHPs tend to operate at high voltage and low current like transmitting tubes. No doubt there are some exceptions like certain voltage regulator tubes. 6AS7, 6C33c come to mind. But notice how low their Mu are, once again leading to high drive voltage.
Triodes in general need higher drive voltages than pentodes too, since the negative feedback from the plate has to be overcome. Adding the cathode feedback winding effectively is lowering the Mu further. In addition, triodes need hefty plate voltage to get emission and power. The pentodes always have the fixed screen voltage to get good emission and power. (although operating into the low plate voltage region, below screen voltage, leads to distortion from excessive screen current)
(electron emission from the cathode depends on the electric field strength present from the grid and the plate. The opposing plate swing subtracts some from the grid field swing, leading to the Mu as the ratio of effects. The cathode feedback winding also subtracts some fraction of the output swing from the grid - cathode field, so is a lowering of Mu effectively. So adding a CFB winding to a triode is somewhat redundant really. CFB is really a design for making pentodes behave like triodes. A 20% CFB winding makes a pentode act like a triode with Mu of 5, 50% gives a Mu of 2)