Idle current measurement weirdness

[g1]

Sunn Spectrum II.
This amp has a bias pot and someone has added a balance pot. I don't think the balance pot is done well as it has replaced the bias feed/grid leak 100K's so now you can put AC ground on the grids. So that needs to be corrected, suggestions for better methods welcome.

The weirdness is that as I turn the balance pot, I can see the grid voltage change with my meter, but the voltage across the cathode resistor (22R) does not change. If I adjust the real bias pot (bottom of drawing) for the same variation of grid voltage, I see the corresponding change at the cathode resistor.
What is going on?

[TomCarlos]

Greg..... And why does this amp have cathode resistors? Is that part of the mod? I am looking at a schematic of the Spectrum. The amp normally has 2 100K resistors where the pot has been inserted and the cathodes connect to ground, correct? So the mod is to help dial in the gain for symmetry? Just curious and asking.

Are you measuring the voltage across each 22 ohm resistor?

[g1]

Yes, the 22 ohmers have been added. I'm measuring the voltage across them to calculate idle current.
Perhaps the pot was added as a symmetry adjustment, I had assumed bias balance but could be wrong.
Anyway, it does change the measured voltage at the grids, but does not change the voltage across the cathode resistors.
I'm wondering if it is some kind of meter induced anomaly.
The regular bias pot also changes the grid voltage, which causes the expected changes in the cathode voltage.

[MarkusBass]

The weirdness is that as I turn the balance pot, I can see the grid voltage change with my meter, but the voltage across the cathode resistor (22R) does not change. If I adjust the real bias pot (bottom of drawing) for the same variation of grid voltage, I see the corresponding change at the cathode resistor.
What is going on?

There is noting weird about it. This is exactly what is expected. The input impedance of the grid is very high and it doesn't much matter whether you connect it to a DC voltage with 100k, 220k or 1k resistor. The voltage on the grid will be still the same. The only thing is that by turning the balance pot, you decrease/increase AC signal that goes to each of the tubes. I suggest that you remove the pot and install two 100k or 220k resistors (as it was initially). Instead of 22 Ohms resistors, use 1 or 10 Ohms resistors. This will make measurements and calculations much easier. And the amp works exactly as expected

Mark

[Enzo]

Ever tryo to measure the voltage at the grids of a LTP phase inverter? You can't. You have to measure voltage grid to cathode.

Look at that balance pot. I don't see how it would work, since there is no voltage divider formed. Until you connect your meter. Never forget your meter has internal resistance. let us say your meter offers a 1 neg resistance. Connect that grid to ground, and at one extreme of the balance pot, you have a 220k/1meg voltage divider, so the reading would be about 0.82 times bias supply. But rotate the pot the other extreme, and we are just having your 1 meg meter to ground from the bias supply. ie. 1.0 time bias. But the only place we see the change is the meter.

I have no beef with the 22 ohm guys, though 1 or 10 ohms is a lot easier to calculate with.

So tell me if my theory seems correct or not. You have two meters? put one across the 22 ohm cathode resistor for a tube, and then put the other meter on the grid. Does touching the second meter to the tube grid alter the cathode current at all? This will only work when the pot is dialed away from that side.

[Dave H]

As Mark said it looks like it's for AC balance not bias balance.

If you are using the same meter to measure both grid and cathode voltage, when the meter is connected to the grid its loading causes the grid voltage to vary as the pot is turned but when the meter is moved to the cathode its loading is removed from the grid and the grid voltage and therefore cathode voltage doesn't change.

[g1]

Thanks guys. I figured it was probably a meter thing but didn't think to try 2 meters at once. It was one of those rush jobs that should have been simple but got frustrating due to the deadline.
I'll probably be working on it again.
Any suggestions for best method to change that symmetry pot to a bias balance? Or examples from other model amps?
I would look to the SF Fender balance circuits but they use that funny 4 legged tapped pot.

[Enzo]

Rush job? Well Rush is a Canadian band after all.

[g1]

And I was using the term 'job' loosely. Rush favour is more like it.

[Chuck H]

Just thinking out loud now, and since this is a freebee it may be a project for another time, but... I'd lose the 'doesn't balance pot' and, since the owner likes the 22R cathode resistors (?) use a 50R rheostat in place of the 22R cathode resistors as a bias balance. Right and left lugs to the cathodes and the CT to ground.