PI Balance Tests

[Littlewyan]

I did some testing last night on my 6V6 PP amp to check the PI balance. Been reading up on this recently and thought I would share my findings.

My amp is 6V6 Cathode bias, 220K power valve grid leak resistors, LTPI with 100K plate resistors, 1M grid leaks, 470Ohm bias and a 27K tail (22K + 4.7K NFB resistor) with 4dB of NFB from the speaker output. With my scope I ran the amp up to just starting to clip and found the PI was fairly balanced, only very slightly out, you can just about see the positive clip on one side. And as expected, once I pushed it very far into clipping it went way out of balance. Scoping the speaker output it seems it's pretty well balanced, both sides start to clip evenly.

I then tested with an 82K plate resistor on the inverting side with a 22K tail resistor and found the PI was still only slightly out, except in the opposite way and a bit more so. You can see in the scope traces that one side clips earlier than the other and more so than with the 100K/27K tail configuration. I did find however that once pushed into heavy clipping the overall the imbalance was reduced a little bit, but it was still very much out of balance and I don't think this is worth the change. And scoping the speaker output with this configuration you can see it no longer clips evenly.

I've attached screenshots of the scope traces showing the PI output just starting to clip and then the speaker output just starting to clip. My conclusion, I'm sticking with 100K Plates and 27K tail resistor. It's a more balanced configuration and I can confirm that the 6V6s do clip before the PI, so to my mind they are being pushed hard enough by this setup. I didn't attach screenshots showing heavy clipping as I ran out of time and we know this causes a heavy imbalance anyway!

Caveats:
- My PI valve may not be balanced, I didn't try others, so it's possible a different valve could produce different results. Although I believe Martin Manning has tested this and found it didn't seem to make much difference to AC balance.
- My 6V6s are running in Cathode Bias with 340V on the anodes, it's possible you would need more drive if this was a Fender Deluxe Reverb running with 420V on the anodes and fixed bias. Although a higher voltage on the PI may be enough.
- I don't have 1Ohm cathode resistors on the 6V6s so I can't specifically monitor the output of each valve individually, although I felt the way the speaker output started to clip was a good indicator.

Anyway, hope that was useful!

[catalin gramada]

You may well read those MEF topics to get well balanced PI outputs without mess with plate loads

https://music-electronics-forum.com/for ... b-question

https://music-electronics-forum.com/for ... ign/49744-

[2L man]

Good to see oscilloscope measures but if the voltage reading is right 5Vpp is only about 0.4W power to 8 ohms. 6V6 PP should produce almost 10W to resistor load before distorting much.

[Littlewyan]

Sorry for the late response, I had the probes set to 10x.

And that's quite interesting Catalin! Nice little trick worth remembering.