How close of a match for PI tube?

[snakej200]

How close does the pi tube have to be matched? I’m auditioning some nos tubes for the pi and I have an rca that one triode is testing 118 on my Jackson 648a, the other is at 105. Clearly both strong. How matched must they be for pi?

 

[StingRay85]

Does your meter measure in gain or transconductance?

I don't think anybody ever proved that one or the other sounds better. I guess worst thing is that the inversed signal is a little stronger or weaker than the other side, but I highly doubt you'll hear that in a push-pull system.

Makes me wonder what would happen if you would play around with values of the PI coupling caps to the power section. What if one side is 10 nF, and the other is 47 nF, what would be the effect on the output to the speaker

 

[PelliX]

Marshall amps' phase splitter circuits are imbalanced by design.

 

[StingRay85]

Inbalanced when having a balanced tube. You could accidentally balance them with an unmatched tube

 

[NickKUK]

Inbalanced when having a balanced tube. You could accidentally balance them with an unmatched tube

True. All that would happen is it would sound sterile.

 

[snakej200]

Does your meter measure in gain or transconductance?

I don't think anybody ever proved that one or the other sounds better. I guess worst thing is that the inversed signal is a little stronger or weaker than the other side, but I highly doubt you'll hear that in a push-pull system.

Makes me wonder what would happen if you would play around with values of the PI coupling caps to the power section. What if one side is 10 nF, and the other is 47 nF, what would be the effect on the output to the speaker

Micromhos

 

[nortiks]

Marshall amps' phase splitter circuits are imbalanced by design.

Interestingly, when you model the Marshall PI on LTSpice it turns out to be "compensated" as opposed to "imbalanced". The plate resistor values compensate for anomalies created by the "tail" resistor values and the net result is identical amplitude, 180 out of phase waveforms at the plates.

I was messing with various PI tubes about 6 months ago and scoped it to see why I was hearing certain things with certain tubes, and seeing fairly different things on the scope with different tubes that logic would say should be about the same. That rabbit hole led me to model it on LTSpice to try to get a baseline of what is the expected ideal result...and the above is what the modeling revealed. And it also turns out to be why some different PI designs have different "imbalance" in the plate resistors. They too model to have identical amplitude. But change the tail values and that can go out the window pretty quickly.

 

[GAStan]

Personally I would audition each tube and go with the one that sounds best to my ear.

 

[printer]

Personally I would audition each tube and go with the one that sounds best to my ear.

Heretic.