Strange HF distortion

[Jeb-D.]

My latest breadboard experiment (triode wired, Class A1, 6CA7 PP) is exhibiting some strange distortion at high frequencies (the onset starts around 20-30kHz)The rising slopes of the sine wave have a wiggle, but the falling slopes look ok. Kind of looks like this.



The distortion increases with amplitude and also increases as frequency goes higher. Adding feedback makes the onset of the distortion occur sooner. The compensation scheme or values used when the loop is closed seem to offer little or no change in this regard.

Any ideas what is going on here? Some kind of OPT limitation? I've never seen this on my previous builds.

[Globulator]

Originally Posted by Jeb-D.

My latest breadboard experiment (triode wired, Class A1, 6CA7 PP)

I think your description of the circuit is a little brief.
Schematics and scope traces at all audio stages would help

[pedroskova]

Is it your output stage or your driver stage causing it? Oscillation?

[DF96]

HF resonance in the OPT? Perhaps triggered by crossover distortion?

[Jeb-D.]

Sorry, I thought someone would recognize this type of distortion right off the bat. I'll post some more details when I get a chance. Right now, the schematic is in my head. Each channel is pretty much 2; 2-stage SE amps strapped back to back via input and output transformers. The output stage is cathode biased 5k plate to plate load, 380V B+. The driver of each 6CA7 is 1/2 6BN11 wired triode common cathode. The phase splitting is done by the input transformer.

The signal coming from the drivers looks clean, but I should probably check for any kind of phase error or change in amplitude difference between the two halves. However, it seems like a transformer parasitic causing the problem.

[tubelab.com]

Sorry, I thought someone would recognize this type of distortion......However, it seems like a transformer parasitic causing the problem.

I have seen distortion like this and it went away when I swapped the OPT's. Does the maximum power output drop at high frequencies too? Mine did.

[Jeb-D.]

Turns out that there is a phase shift going on between the two driving signals.
They are 180deg out of phase (as is desired) for frequencies below 20kHz. But once it hits about 20kHz, one of the signals phase starts shifting earlier than the other.

Tracked this phase shift down to the input tubes grid. So it seems the input capacitance of the two sections on the input tube is significantly different (a look at the 6BN11 suggests this may be plausible). Or there is some kind of imbalance between the two halves of the input transformers secondary.

[kevinkr]

Originally Posted by Jeb-D.

Turns out that there is a phase shift going on between the two driving signals.
They are 180deg out of phase (as is desired) for frequencies below 20kHz. But once it hits about 20kHz, one of the signals phase starts shifting earlier than the other.

Tracked this phase shift down to the input tubes grid. So it seems the input capacitance of the two sections on the input tube is significantly different (a look at the 6BN11 suggests this may be plausible). Or there is some kind of imbalance between the two halves of the input transformers secondary.

Both are quite possible.. And I suspect in practice this might not be that much of an issue.

You might be able to pad up the input capacitance on the lower of the two sections with a gimmick between the grid and plate (grid and ground?) if you suspect that might be the cause.

[Jeb-D.]

You might be able to pad up the input capacitance on the lower of the two sections with a gimmick between the grid and plate (grid and ground?) if you suspect that might be the cause.

Thanks for the suggestion. That's what I'll try to do when I get a chance. It's a little surprising this type of issue isn't more common. There are a lot of "balanced" designs floating around out there.

[DF96]

If you have two input transformers then there is no reason why their HF resonances will coincide. Same for output. Near resonance you get a very rapid phase shift. This is why people normally use one combined transformer for both phases.