Measuring Output Class

[lowell]

Can someone please tell me how to see if a power amp is operating in Class A push-pull? I modded my 5E3X2. The plate voltages are at 300v and there is roughly 32ma on each 6V6.

[stokes]

That is kind of in between AB and A.More like a "hot" AB.Without starting the whole"what constitutes true class A" discussion,I'll just say class A is closer to a 90% idle dissipation.

[lowell]

Ok so 42ma is what I'm aiming for? Also, I scoped it at 42ma and there is still cutoff... I'm guessing this is because of the PI drive signal. Is this right? If so, how shall I adjust drive to compensate?

[ETR]

Check the voltage at the 6V6s cathodes. If it gets much higher when you crank the amp up you're most likely in class AB

[lowell]

ETR,
Interesting... would you mind enlightening me on why?

[Robert M. Martinelli]

Can someone please tell me how to see if a power amp is operating in Class A push-pull? I modded my 5E3X2. The plate voltages are at 300v and there is roughly 32ma on each 6V6.

The class of operation should be stated under working conditions, as it refers to the signal over its 360 degrees. Whatever the amp, to be in true class A the tubes should be on at all times ( 360 degrees ). This means setting the quiescent point of operation ( bias ) in such a way that the tube never cuts off when the signal swings.

Hope this helps

Best regards

Bob

[tubeswell]

Assuming you have good 6V6s in there, if the 6V6s are cathode-biased, you can set them at 46mA per tube if you are only running 300V on the plates. Cathode biasing provides automatic adjustment to bias - depending on plate voltage (as the plate voltage increases the voltage drop across the cathode resistor also increases, which increases the bias) so it is perfectly safe. I think they will sound better at 46mA per tube

[Old Tele man]

Class A = 100% plate dissipation at idle.

Class B = 50% plate dissipation at idle.

Class AB = between 51%-99% plate dissipation at idle.

[Enzo]

Class B 50% at idle? I thought class B each tube was biased right at cutoff. Only conducts on its own half of the cycle.

COnducts 50% of the waveform in class B, while class A conducts 100% of the waveform. That isn't dissipation though.

Class AB conducts 51-99% of the waveform.

[lowell]

Guys I appreciate the response. However I know what Class A is and am familiar w/ biasing etc... I'd simply like to know how to measure if it IS in class A or not. Can someone tell me how to do this? I was using a scope yesterday and saw that the amp is still NOT in class A as there is a scoop on top of the voltage waveform for each side of the OT. Which I guess answers my question of how to see if it is in Class A or not. So now I'm curious as to how I get rid of this scoop. I'm guessing it's something to do w/ the PI...?

[Enzo]

On both positive and negative peaks? That would be clipping, and more likely from the PI stage.

[ETR]

ETR,
Interesting... would you mind enlightening me on why?

It's something called bias shift that doesn't occur in class A. If you check the cathode resistor in a typical bypassed preamp stage you'll notice that the DC voltage doesn't really change between idle and operation. That's because it's also a class A stage. If you're in class A with a shared cathode resistor in an output section neither tube goes into cutoff for the full unclipped output. If neither tube ever turns off the current through the cathode resistor (and the voltage drop across it) will remain pretty much the same. That's because as one tube is conducting more the other tube is conducting equally less current. Of course there is usually some slight variation since both halves are never perfectly balanced but you shouldn't get much fluctuation of DC across the cathode resistor if you're really in class A. There's other ways to check but they're more complicated and take more than a voltmeter to check.

[Steve Conner]

This argument is only valid if the tubes were perfectly linear. In practice, the presence of even harmonic distortion causes the B+ current draw to increase with drive, and the cathode voltage to rise accordingly, even in Class-A.

RDH4 calls this "rectification" and says that it can be significant with beam tetrode tubes.

I guess it would be easy enough to tell the difference between the gradual current increase caused by rectification, and the sudden rise that would happen as you came out of Class-A.

[ETR]

This argument is only valid if the tubes were perfectly linear. In practice, the presence of even harmonic distortion causes the B+ current draw to increase with drive, and the cathode voltage to rise accordingly, even in Class-A.

RDH4 calls this "rectification" and says that it can be significant with beam tetrode tubes.

I guess it would be easy enough to tell the difference between the gradual current increase caused by rectification, and the sudden rise that would happen as you came out of Class-A.

The B+ current shoudn't change much in class A until the output waveform gets clipped, which is probably where the "rectification" starts. It will change a little due to imbalance and nonlinear curves but it's usually very slight. Once you get into clipping all bets are off.

[lowell]

Good info guys and thanks. So constant current/voltage drop across Rk is result of Class A operation. So how much change in current/voltage drop across Rk is to be expected when deciding operation is no longer in Class A? Are we talking couple ma?