Gemini 2 bias changes while playing

[tdlunsfo]

This one is new to me but I'm sure you guys have experience it at some point. I noticed when I cranked the amp the tubes started glowing brighter when ever I played harder. Threw the meter on it and saw my DC bias voltage changing at the grid. Amp is fixed biased. I have them biased cold at 18 mA or approx -28v while I troubleshoot. I would see it jump down to -60v at times while playing.. strange. Running a pair of used old stock 7591 sylvania's. I plan to buy some new TungSol 7591A tomorrow. Any ideas? Think it has something to do with the difference between 7591 and 7591A's?

[J M Fahey]

All tubes do this.

Hidden in plain sight are two diodes: each grid/cathode combination is one ... the grid being the anode .

Normally they are reverse biased and not passing any current but when power tubes get driven hard, as soon as grid voltage gets positive (even if amp is nominally class AB1 and not exactly designed as AB2) diodes become forward biased and pass current.

Now that current is "sucked" from coupling capacitors, and since they do not pass DC by themselves, they become more negative than before.
In your case, shifting bias more negative ... big time ..... what you measured.
Those grids will recover regular bias voltage when you lower volume or stop playing, but that takes a fraction of a second, is not instant, you will have a recovery time constant defined by coupling caps and bias resistor values.
As in .022 caps and 220k resistors will give you a time constant of 5 milliseconds.
Looks too short to be noticed but in fact changes note attack which gets compressed.
Thatīs why some guitar players prefer SS amps because they are "faster", the opposite being Blues players who prefer "creamier/smokier/greasier" sound, which in part is explained by slower attack.

This causes a bunch of things:
1) bias shifts more negative, colder, which is visible in waveform kink near zero crossing.
Some mistakenly try to kill that by biasing amps way too hot, or limiting excessive negative bias by using Zeners, "Ruby mods", etc.
A non cure for a non problem.
2) colder/more negative bias means lower gain, because of lowered tranconductance.
Net result is amp compresses applied signal, becomes smoother, shows liquid sustain, etc.

Killing this is killing part of tube sound.

[tdlunsfo]

Juan thanks for the reply and the education. I had a feeling this may be normal and I had just never seen a power tube driven hard like this. The amp sounded great but watching the power tube brightness change like that made me nervous, especially old stock 7591's.

[Enzo]

Remember, bias is only a static measure at idle. When playing through the amp, all you are measuring is tube current.

[Helmholtz]

I noticed when I cranked the amp the tubes started glowing brighter when ever I played harder.

I would see it jump down to -60v at times while playing.

There is no direct relation between the 2 observed effects. If it was only a shift of the operating point to more negative grid voltage, the tubes would run cooler.
But as Juan explained, the reason for the -60V is grid current in the other half-cycle. And grid current means heavy overdrive which strongly increases screen dissipation and makes screens glow.

[The Dude]

Remember, bias is only a static measure at idle. When playing through the amp, all you are measuring is tube current.

^^^^^^ That! Stated another way: Bias isn't changing. Tube current is, which is normal operation.

[Helmholtz]

Stated another way: Bias isn't changing.

Right, but average DC grid voltage is changing (getting more negative) when the power tube is driven into grid current.

[The Dude]

Understood completely, but bias is, by definition, an idle measurement.

[tdlunsfo]

I see the fallacy in my thinking now. Bias is at idle and measuring at the grid while playing likely doesn't tell me much. Those tubes sure are getting a workout though. I plan to do some experimenting tomorrow. I am curious how the tubes react to hotter bias and a larger screen resistor as Juan alluded. Current screen resistor is at 470ohm but did see where ampeg swapped to a 1k when they replaced the 7591s with 6L6s. Not trying to accomplish anything, more for my own education than anything.

[tdlunsfo]

I took a video of these tubes. Just wanted you all to see it to ensure this is normal. Sorry for the bad playing its below freezing in my shop.
https://youtu.be/l8t5DLtK-uY

[Justin Thomas]

I've seen the blue glow get more intense, but WOW, look at those screen grids light up!

As far as vintage 60s Ampegs having less-than-stellar distortotron sounds, I call that normal. They were designed to be as hi-fi as economically possible. The founder/owner HATED rock&roll & was pretty up front about it.

Justin

[Helmholtz]

I took a video of these tubes. Just wanted you all to see it to ensure this is normal. Sorry for the bad playing its below freezing in my shop.
https://youtu.be/l8t5DLtK-uY

The video clearly shows over-dissipation of the screens. This is not safe operation, power pentodes/tetrodes have screen dissipation limits and are not designed for bright glowing screens. Expect reduced tube life.
Screen dissipation can be lowered by reducing screen supply voltage and using higher value screen resistors (and by decreasing load resistance/impedance).

[nickb]

The video clearly shows over-dissipation of the screens. This is not safe operation, power pentodes/tetrodes have screen dissipation limits and are not designed for bright glowing screens. Expect reduced tube life.
Screen dissipation can be lowered by reducing screen supply voltage and using higher value screen resistors (and by decreasing load resistance/impedance).

That and add, check that you are not using too high a load impedance

[Helmholtz]

That and add, check that you are not using too high a load impedance

That's what I implicated when saying "and by decreasing load resistance/impedance".

[tdlunsfo]

I'll check to be sure but believe it is an 8ohm tap and an 8ohm speaker. I assume you would lower screen voltage by increasing the screen resistance? I am at 470ohm currently. My gut told me this wasn't normal.. sounds like there are a few that agree.