Suggestions for Revisions?

[kt88pppamp]

I created a MKIII driver board with the attached circuit. It performs with excellent distortion characteristics < 0.3% with around 20dB of feedback at 1k. It could do better I suppose if the 6SN7's are matched. I am having some issues with the first revision though. They are:

- Occasional oscillation without input attached. It seems pretty stable otherwise.

- Potential oscillation without loudspeaker connected (too afraid to try it).

Questions:

With this configuration, what type of network and where should I place it to help get better ringing performance with the square wave, and is it really necessary?

If possible, what should I try to obtain the distortion numbers I am achieving with less feedback (if possible)?

Also, what other general suggestions do you have for this circuit?

Schematic errors:

- Nodes H and I reversed.

- Bias network resistors are of different values (experimentation).

[Geek]

1) Try a grid snubber on your input tube.

2) Get a resistor in series with your zero-pole comp cap of 180p and re-Spice for cutoff characteristics. As it is, it's creating a Push-Pull-Plate-to-Plate (PPPP) oscillator. That's a Jones RF oscillator only with the reactive device between the anodes, rather than grids

Cheers!

[kt88pppamp]

Should I add anything else between the two plates of the first stage plus a series resistor for square wave tweaking? Would moving the cap plus network between the grids of the first and second stages be a better idea?

[DeathRex]

Need to reverse the grids on the KT88s, it's messing with my head

20db seems a lot for UL, can you add some KT88 anode to middle driver anode local feedback?

[kt88pppamp]

Anode to anode feedback is making my phase splitter stage clip sooner in spice. What is the best way to tweak so distortion is reduced, not increased? It happens in real life too, as I used it in one of my amplifiers.

http://www.tubecad.com/march2001/index.html

[Geek]

Should I add anything else between the two plates of the first stage plus a series resistor for square wave tweaking? Would moving the cap plus network between the grids of the first and second stages be a better idea?

Nope, leave it where it is. Just a resistor in series with your 180p cap.

Though in practise, you'll find say, 82p to 120p in series with 10K to 47K will be optimum.

Cheers!

[kt88pppamp]

The headroom seriously seems to decrease on the phase splitter with partial feedback. Any tricks on working with this if I were to use it? The amp needs to swing about 25 volts peak for 60 watts out. For 150 about 40 volts (with my PPP amp).