Class A Push Pull - output transformer specs

[abbey road d enfer]

What can I say, comparing some clean stuff, the Class A 260v version is 1.2dB louder than the Class AB 325v version.

Class A has more gain than class AB, but less max output power. That's because in class A the output tubes are in the middle of their slope, when in class AB their are at the beginning.
It's easily noticeable on an amp with variable bias. Often, people tune them for maximum output when strumming their guitar, which tends to be at the maximum of the bias range. They often red-plate the tubes.

 

[CJ]

Get a dummy load and hook it to a 1/4" jack,

feed in a sine wave into amp and measure voltage across dummy load.

I have a table of values at work up on the wall,

5 volts = 12 watts,

25 volts = 50 watts

etc

Class A might have different harmonic content than AB, I don't know, I ain't no rocket scientist, I just stick in different components til it sounds good.

Don't forget that fixed bias class a might sound different than cathode bias class a. Maybe some compression on the self bias? Maybe some compression with a 5U4? Maybe some compression due to transformer DCR?

I don't know, what kind of weird amp did Hubert Sumlin use?

 

[CJ]

dummy load, stereo in case i get a weird amp in like a JC 120,

100 w at 8 ohms each, jacks on the outside are for a speaker load, Altec 15" on each side,

i can parallel those resistors if checking a 200 watt pig amp

transformer for isolating the output signal for the scope, sometimes the amp does not like a scope ground connected to the chassis, (oscillation)

V-AC for power and milliamp meter for doing a death bias, (shunting B+ to OPT primary)

 

[trobbins]

It's also worth going back to the loadlines in post #22 and appreciating what options you may have. The last plot shows an output transformer PP impedance of circa 4k5 ohm, but the low distortion section is likely for plate voltage swing between 75V and 400V, beyond which the peak of the waveform becomes increasingly more distorted.

To avoid that in vintage hi-fi amps, the PP impedance is increased to say 10k and the idle current reduced, and then the low distortion voltage swing extends to perhaps 50V to 450V, such that the loadline doesn't reach cut-off with zero current, and the total loadline effectively stays under the max dissipation curve you want to apply. One aspect to appreciate is that the actual loadline is not straight, but will extend out like an ellipse, so margin above and below the loadline allows for lower distortion operation.

So apart from tube selection, the next phase of 'design' perhaps is the OPT impedance you have on hand, and from that you can start to define the idle V and I point that achieves as much linear option in the tube curves before approaching tube regions where screen current rise starts to become a player.

 

[systemtruck]

It’s also worth going back to the loadlines in post #22 and appreciating what options you may have. The last plot shows an output transformer PP impedance of circa 4k5 ohm, but the low distortion section is likely for plate voltage swing between 75V and 400V, beyond which the peak of the waveform becomes increasingly more distorted.

To avoid that in vintage hi-fi amps, the PP impedance is increased to say 10k and the idle current reduced, and then the low distortion voltage swing extends to perhaps 50V to 450V, such that the loadline doesn't reach cut-off with zero current, and the total loadline effectively stays under the max dissipation curve you want to apply. One aspect to appreciate is that the actual loadline is not straight, but will extend out like an ellipse, so margin above and below the loadline allows for lower distortion operation.

So apart from tube selection, the next phase of 'design' perhaps is the OPT impedance you have on hand, and from that you can start to define the idle V and I point that achieves as much linear option in the tube curves before approaching tube regions where screen current rise starts to become a player.

 

[Wavelength]

That absolutely makes sense.. Yes. It outside of the typical tube amp topology. That 60-70% threshold is where the 2nd harmonic really starts to creep in and create the typical single ended distortion profile. But if you can do as you do and use a low Z driver and supply current to the grid then a single ended amp extends it's range quite profoundly.. but with that cost of a a grid choke. I like where you're going.. except for that 3-legged fuse of a LM317.

I use the LM317 as a current source. Also have done fixed with zeners or chained LEDS.

 

[Cqwet Dbdfte]

Bias class A vs AB unclipped would be very hard to differentiate without instrumentation. Adding "color' to sound would be best done at low level, stomp boxes etc, and let power amps do the heavy lifting within their normal operating range.

 

[Cqwet Dbdfte]

A Class A like DHT 300B SET amp is true class A. Pentode on the other hand is different because it has more than one thing drawing and giving current. I use a reactor follower and direct couple the cathode of the driver (usually a 6SN7 or smaller 12AU7) to drive pentodes into A2. What I was saying is not text book, but the way it works. Even a 300B when reaching 60-70% of it's bias will start to draw grid current. That current will excite the resistance to ground from the grid and rebias the tube. One of the reasons we use grid chokes as they have high AC Z and low dcr (less than 4K for a 4000H high nickel unit). Look at any pentode and you can see the plate vs screen current as the incoming signal reaches the bias point in SE. The amp on the left is a 5686 (small EL84) and in the protos I tried using a LM317L to set the bias current and of course that current screen + plate will change and as the screen draws more the plate would draw less. The same thing happens with self bias R/C. Now I could fix bias the amp and if nothing is in the cathode then the power supply current would still go north as the input signal reaches the bias point.

I've only seen grid bias via a choke in a DHT (zero bias) 811A amp, driven by cathode followers. Grid current, sure, no problem. Not a common topology.
The SE DHT 300B, a very early Western Electric purpose designed audio tube, would be a good choice for a nice amp, if you have very high efficiency speakers, like the old horns.
Every push-pull transformer suppresses even order harmonics, not a problem if low enough
If I had room for those high efficiency horns the SE amps would be good choice.

 

[Wavelength]

I've only seen grid bias via a choke in a DHT (zero bias) 811A amp, driven by cathode followers. Grid current, sure, no problem. Not a common topology.
The SE DHT 300B, a very early Western Electric purpose designed audio tube, would be a good choice for a nice amp, if you have very high efficiency speakers, like the old horns.
Every push-pull transformer suppresses even order harmonics, not a problem if low enough
If I had room for those high efficiency horns the SE amps would be good choice.

I have made 1280 SET High End Audio amps with various directly heated triodes from Telefunken RE604 to WE212 (50W SE 2x chassis, both 80lbs). I currently sell three WE300B SET amps, Triton Silver Signature 15W, Napoleon Silver and Duetto v5 available in silver or copper wound transformers. I have on occasions designed tubes for some companies. Pentodes don't draw as much current in the grid as directly heated triodes do. But then again, Pentodes have a lot less bias voltage than say a 300B does.
I have made 15W 300B guitar amps before. Ok for Jazz but they don't break up like a pentode does.
I have made some audio amps that go A2, like the 211 and the Telefunken RS241. Both of them used the reactor follower direct couple to maximize their power. I try and stay away from transmitter tubes like 811/572 and others like that for audio.
The grid chokes I use for cap coupled stuff are not meant for direct current. But they do work well for directly heated triodes. I have seen as much as 75uA of grid current on a 300B. If you use a 220K gird resistor that's going to be over 10V. If you use a grid choke with a dcr of 4000, that's only 0.2v giving you like I said about 20% more power.