LTP bias Q

[nickb]

Thanks RG for checking.

[Chuck H]

Thanks for taking the time R.G. You called it like it was but ran the sims anyway. Capitol research. I don't have the simulation skills (or programs) so it saved me bench time with a hot iron. And, to be clear, WRT this sort of work and research I think your time is more valuable than mine. So I really appreciate it.

Also... Kudos to teemuk for calling it right out of the gate (and before finishing his coffee )

[Chuck H]

On another note, Let's stop calling the zener across the cathode bias circuit the "Chuck H mod". Since I didn't actually invent it, it makes me feel too "Randall Smith" More research in the archives revealed that a member with the SN "Dutch" posted in September 2003 that he's been doing it for awhile. He got the idea from another members design, Carl Zwengel, who had been using a higher value cathode resistor and then trimming the voltage down lower with the zener for fully fixed bias. And, of course, there's not much new under the sun. It could probably be found in some dusty old tech manual somewhere.

P.S. We haven't seen Dutch since December 2007 and we haven't seen Carl since September 2006. I remember Dutch from the old BBS. Maybe I read the post then and plagiarized the circuit subconsciously. Carl did an amp builder thing for awhile. I bought some stuff from him when he got out of the game.

[Dave H]

The PR mod uses a pair of zeners to clip the PI negative swing a volt or so above the bias voltage. This minimizes coupling cap charging and allows a period of time for them to discharge while the power tubes are in cutoff. Brilliant! But since I use this circuit in "my" designs I've always wanted my own solution to the problem.

Try this in "your" designs - use a zener connected to ground with a diode from each grid to the zener. It could be the new "Chuck H mod", more elegant than thr PR mod as it does exactly the same job but with fewer components

[Chuck H]

But wouldn't that divert all the positive swing from each leg to ground? I figured that was why the standard diodes were reversed in series. I've seen the circuit proposed as you describe it, but never tried it. Next time I have an amp on the bench I can try it out.

[GainFreak]

What is the purpose of those two diodes (CR101, 102). Aren't they supposed to prevent grid conduction? Peavey is using them often.

XXX-power.jpg Photo by warmitag | Photobucket

[Dave H]

But wouldn't that divert all the positive swing from each leg to ground?

No, the diodes are reversed (as below). Only the negative swing is clamped.

[Dave H]

What is the purpose of those two diodes (CR101, 102). Aren't they supposed to prevent grid conduction? Peavey is using them often.

I don't see that they have a useful purpose. They may prevent grid conduction but now you have diode conduction which is the same as far as bias shift is concerned.

[Chuck H]

No, the diodes are reversed (as below). Only the negative swing is clamped.

[ATTACH=CONFIG]34674[/ATTACH]

Ah. I interpreted the circuit incorrectly in my head (happens). Should work. It's the same as the PR circuit but your single zener is doing double duty. So yeah, one less component. Might prove less clumsy looking on the layout too.

[Chuck H]

What is the purpose of those two diodes (CR101, 102). Aren't they supposed to prevent grid conduction? Peavey is using them often.

XXX-power.jpg Photo by warmitag | Photobucket

Hmmm.... Now THAT'S a regular diode on each PI output diverting positive to 0V. I don't imagine that's drawn correctly. Am I missing something about what this would do?

[teemuk]

I don't see that they have a useful purpose. They may prevent grid conduction but now you have diode conduction which is the same as far as bias shift is concerned.

Well, they certainly seem to clip limit before clipping via grid conduction so you're right, they replace grid conduction with diode conduction "which is the same as far as bias shift is concerned".

But we don't only concern bias shift. Grid conduction also causes screen grid to draw plenty of current, which would lead to all kinds of interesting things during overdrive depending on the screen circuit design in general. It certainly looks Peavey's aim is to prevent this associated surge in screen current draw by preventing grid conduction, while still preserving the bias shift characteristics associated to it.

Given that Peavey has handful of patents concerning crossover distortion's importance in "push-pull" amp sound (these patents spread across few decades) they might fall to the camp who regards slight bias shifting, crossover "fizz" and blocking distortion as essential guitar tube amp traits. Who knows...

[Chuck H]

Given that Peavey has handful of patents concerning crossover distortion's importance in "push-pull" amp sound (these patents spread across few decades) they might fall to the camp who regards slight bias shifting, crossover "fizz" and blocking distortion as essential guitar tube amp traits. Who knows...

That sounds right. They were ahead of the curve with their TransTube amps before digital modeling took over. The first time I used the PR circuit in an amp it was to reduce an awful fizziness. I remember thinking "Well that's not fizzy enough now." Which surprised me. I have the circuit adjusted to get a trace of crossover distortion. I actually need to rework the circuit in my personal amp now because of the new power tubes. I'm not getting ANY crossover distortion and I want a little. El84 amps never really "swirl" IME. It's a light swirly buzz that really does nothing for the tone that I can pinpoint except to say too much is bad and I miss it when it's not there at all.

[Dave H]

Grid conduction also causes screen grid to draw plenty of current, which would lead to all kinds of interesting things during overdrive depending on the screen circuit design in general. It certainly looks Peavey's aim is to prevent this associated surge in screen current draw by preventing grid conduction, while still preserving the bias shift characteristics associated to it

Thanks, I was only thinking of bias shift. I didn't know about the screen current surge.

A quick simulation shows 30mA peak screen current without diodes but only 12mA with diodes.

[Dave H]

I'm not getting ANY crossover distortion and I want a little.

Put a 500k log rheostat in series with the zener of the circuit in post #37 and adjust to taste. Use the ccw and w terminals of a 500k log pot to make the rheostat. It works fine in LTSpice so it's worth a try. You could measure the rheostat when it sounds right and replace it with a fixed resistor.

[loudthud]

Another interesting trick when a PPI Master Volume is used is to clip the positive peaks to ground (fixed bias) or to the cathodes (cathode bias) at the top of the pots.