Crossover Distortion During clipping

[JJman]

I had crossover distortion on the output of a home build with a 5e3 output but with a LTPI. I read up on it and it’s caused by the ramifications of driving the grids somewhat positive which charges the coupler for a period of time. The bias is momentarily pushed down I think while the coupler discharges. (It took a while for me to wrap my head around it.) That bias push-down increases the cutoff period and results in the crossover notch. I resolved it with lower grid leaks to reduce the drive somewhat. I had read that reducing the couplers was the proscribed way to resolve it but wanted to reduce the drive some anyway and didn’t need to reduce the .1 couplers.

Now I’m looking at my ’71 SFDR which has BF component values. It has a small notch when the output is clipping. (The PI clips later on the volume control.) I’ve read some say that it’s normal for Fenders to do this because of the .1u couplers but there’s plenty of bass to spare so I tried the recommended reduction on the couplers to .047 then .022. No improvement. I’ve also read that reducing the amount of NFB can eliminate the phenomenon and will try that next. I don’t think reducing the drive with lower bias resistors would make sense here.

Anyone ever notice this on Fenders and “fix” it?

 

[DT7]

That's normal for a properly biased push-pull (Class AB) amp, and there's no getting completely away from it. The whole Mesa-Boogie "simul-class" concept was an attempt...but only partly successful (depending who you talk to).

 

[RussB]

Every amp I've put on the 'scope looks like that waveform when pushed hard

 

[JJman]

Forgot to mention the grid stopper resistors. After all, they are there to "stop" the grids from conducting current which is the root cause of the chain reaction that leads to the notch. I've read I can go to 10k for them w/o losing high end in a guitar amp. I'll probably give that a try to see if there's an improvement.

 

[RussB]

you're trying to "fix" what ain't broken

 

[pdf64]

'the grid stopper resistors. After all, they are there to "stop" the grids from conducting current'
My understanding is that they are called grid stoppers because their function is to 'stop' high frequency oscillation, by forming a low pass filter in conjunction with the grid-plate capacitence.
I guess they might mitigate the grid clamping action a little, but the bias will still inevitably pump up under heavy overdrive, thereby causing crossover distortion.