Variable cathode resistor(pot) with diff size bypass caps? Ampeg Mod

[Mr Johnny Birchwood]

So, I just got my Ampeg Gemini VI in good working order. Being as I am an incorrigible tinkerer and being that the parts for my next project wont arrive until next week, I've been thinking about mods for the gemini. The Gemini VI has a 100k cathode balance pot on the power tubes. This is useful if you don't have matched pairs of tubes. I have 2 good matched pairs, so I'm thinking of rewiring the pot with both cathodes going to the input, and different size bias caps on either side of the outputs. I believe I should run a resistor in parallel with each cap, as when the pot is open completely to one side, there will be no resistance. Does the size of a cathode bypass cap affect the amount of resistance to run in parallel with it? Let's say 25uf on one side, 10uf on the other side, just for the sake of discussion. Any suggestions for the amount of resistance I should use with each cap? The way it is set up now, with the pot centered, each cathode(with 50k of resistance) is pulling .020 amps(.556VDC.) Currently I have 436VDC on the plates of the 7591's with -22VDC grid bias. Schematic pictured below. It's pretty accurate.

Oh, and please don't ask me WHY I want to do this. I am just bored and itching to solder something. It's an easily reversible mod.

Thoughts? Suggestions?

[Enzo]

That is a 100 OHM cathode balance pot. If you run it all the way over to one side, yes, that side has zero ohms. Zero ohms is what MOST push pull output tubes see in guitar amps. The OTHER tube now has 100 ohms in series. At for example 40ma tube current, that results in 4 extra volts of bias. Cenetered, there will be 50 ohms per tube or about 2 volts. That is just 2 volts colder than the -50v already there from the fixed bias. The pot is truly just a balance, a small adjustment to trim things, it is not a real source of any bias. A bypass cap across that small amount won't do much, this is not a cathode bias circuit.

If you had 20ma through a tube, and 50k ohms for it to travel through, Ohm's Law tells us the voltage drop across the resistor would be: E = IxR so 0.02A x 50,000 ohms = 1000 volts.

If you have .556 volts across 50k ohms, then the current through the tube would be: I = E/R so 0.556/50,000 = 0.00001 amp, or 0.01ma

If you have 0.556 volts across 50 ohms, then 0.556/50 = 11ma.

If you want to eliminate the balance feature, just wire the cathodes to ground.

You only have 22 volts of bias?


AM I missing something here?

[Mr Johnny Birchwood]

That is a 100 OHM cathode balance pot. If you run it all the way over to one side, yes, that side has zero ohms. Zero ohms is what MOST push pull output tubes see in guitar amps. The OTHER tube now has 100 ohms in series. At for example 40ma tube current, that results in 4 extra volts of bias. Cenetered, there will be 50 ohms per tube or about 2 volts. That is just 2 volts colder than the -50v already there from the fixed bias. The pot is truly just a balance, a small adjustment to trim things, it is not a real source of any bias. A bypass cap across that small amount won't do much, this is not a cathode bias circuit.

If you had 20ma through a tube, and 50k ohms for it to travel through, Ohm's Law tells us the voltage drop across the resistor would be: E = IxR so 0.02A x 50,000 ohms = 1000 volts.

If you have .556 volts across 50k ohms, then the current through the tube would be: I = E/R so 0.556/50,000 = 0.00001 amp, or 0.01ma

If you have 0.556 volts across 50 ohms, then 0.556/50 = 11ma.

If you want to eliminate the balance feature, just wire the cathodes to ground.

You only have 22 volts of bias?


AM I missing something here?

Sorry about the inaccuracies in my question there..You are right it is 11ma, I meant to say 20ma, just typing fast and not thinking. I just did a mental calculation to get 20ma. The ampeg schematic calls for 80ma through the 7591 cathodes and a -50 bias. With a -50 bias the amp barely worked. Weak, grainy, very distorted and nearly no sound coming out. I tweaked the bias until it sounded good, which happened to be at -22. Didn't start to get much sound out of it until about -30. I posted a question about this in the maintenance, troubleshooting, and repair forum, but(and I believe it was you, sir, who said it--no offense) it was suggested that I stop fixing it if it sounds good. The low bias volts and low current draw through the cathodes gave me some hesitation(putting extra wear and tear on my tubes?) but I am happy with the way the amp sounds. The idea behind the bypass caps was to have another tone control(changing the capacitance of the cathode caps affects frequency response, I'm told.)

The other idea was that I could adjust between grid and cathode bias, just for the sake of research. Like I said, I'm bored. I don't watch TV.

[Enzo]

OK fair enough, my little slogan doesn't always apply to real life. No offense taken, if I say something wrong, I can't correct myself if no one calls me on it.


However, if you have to turn the bias down to 22 volts instead of 50 to get it to work, that ain't right. I suspect your tubes were barely conducting, and you turned the bias up to half what it should be just to get sound. You want to set the control for -50v at those grids, then make it work at that voltage. Then later, if you need to you can tweak the bias, but 22 instead of 50, no no.

What voltage is on the screens? That is pins 4,8 on the power tubes? It ought to be within a few volts of whatever is on your plates. Missing or very low screens will make the tubes not conduct. And that balance control, power off, do we have 100 ohms more or less between the two cathodes? And centered, do we have roughtly 50 ohms to ground from each? Measure from the tube pins to chassis, not just at the control legs.

[Chuck H]

I think the schematic is incorrect. There is another Ampeg schem for a different Gemini that shows -100V on the grids! That -50 is probably supposed to be the base voltage for the bias supply behind the divider. I say this because I know 7591's to be sort of like big el84's in that they require lower bias voltage than many other tubes. Look up the data and see the typical operating conditions and you'll see.

[Enzo]

Well, I seem to be doing real well in this thread...

I sure wish we got to see the original schematics instead of these darn Piazza re-draws.

[Mr Johnny Birchwood]

Well, I seem to be doing real well in this thread...

I sure wish we got to see the original schematics instead of these darn Piazza re-draws.

I have an Ampeg schematic I got from the company, with voltages on it, but they made me swear on pain of death and loss of future tech info privileges not to share it with anyone.

The screen voltage is 426VDC. Plates 436VDC. Power off, with balance pot centered, I read 50 ohms to ground from each cathode pin. The output of the pot is to ground so I don't read 100 ohms between the cathodes. 52 Bill also previously suggested that the schematic(s) may be in error. The Ampeg sourced schematic does, however, say that the amps are adjusted at the factory to pull 80ma at the cathodes. If I was pulling 80ma through the cathodes I would need more grid bias -volts, correct? Also, plate voltage would be lower as the tubes would be conducting more.

Also of note. I took readings at the cathodes with new matched EH 7591A's and a pair of old 7591s(they are labeled 3M Revere Wollensak--RCA made I believe.) The cathode readings were virtually identical with both sets of tubes, so it's not the tubes. The bias voltage on the Ampeg schematic is unreadable, but ends in a 2. could be a 4 maybe? -42?

[Chuck H]

For the tubes to draw more current you need LESS negative voltage on the grids.

[Mr Johnny Birchwood]

For the tubes to draw more current you need LESS negative voltage on the grids.

Sorry, I shoulda been more clear. I was trying to say that if the cathodes were conducting 80ma instead of the 11ma they are currently(pun intended) conducting, THEN the tubes would be conducting more, drawing down the plate voltage. If this were the case I would then need LESS or MORE bias volts(-) on the grids? More?

[Enzo]

The output of the pot is to ground so I don't read 100 ohms between the cathodes.

That doesn;t sound right. Ohm meter probes - one on pin 8 of tube A, the other on pin 8 of tube B. Should get 100 ohms if the pot is 100 ohms. That the wiper is grounded shouldn;t matter. Imagine you are holding a yardstick. The ends are 36 inches apart. You can hold the stick at the center, or 12" from one end, or anywhere you want, but the ends will always be 36 inches apart.

OK< wait a minute. Two things: WHICH schematic do you now have? What number is the drawing, lower right corner? I am an authorized Ampeg service center, I can look up and download anything they have. You don't need to violate your promise, just tell me which one you have. The other thing is this, just WHAT model number is your amp? In post #1 you showed a GS15R drawing. Is that what your amp says? I have G15 drawings that do say -22v and 80ma. And I have some that say -50v. There were various versions of this amp.

[Mr Johnny Birchwood]

That doesn;t sound right. Ohm meter probes - one on pin 8 of tube A, the other on pin 8 of tube B. Should get 100 ohms if the pot is 100 ohms. That the wiper is grounded shouldn;t matter. Imagine you are holding a yardstick. The ends are 36 inches apart. You can hold the stick at the center, or 12" from one end, or anywhere you want, but the ends will always be 36 inches apart.

I can take the reading again, though probably not today as I won't home til after midnight.

OK< wait a minute. Two things: WHICH schematic do you now have? What number is the drawing, lower right corner? I am an authorized Ampeg service center, I can look up and download anything they have. You don't need to violate your promise, just tell me which one you have. The other thing is this, just WHAT model number is your amp? In post #1 you showed a GS15R drawing. Is that what your amp says? I have G15 drawings that do say -22v and 80ma. And I have some that say -50v. There were various versions of this amp.

The ampeg schematic says GS-15-R Gemini VI. I don't see a drawing number in the lower right corner, but there is a -43- to the left of the drawing. I don't know the exact model number of the amp, as some previous owner sawed off the speaker enclosure, just leaving the head and the reverb tank. There are no identifying markings on the amp chassis giving model number or serial number. I completely rebuilt the amp though, using the piazza schematic, and I found said piazza schematic to be 99% accurate compared to what I found in the amp when I was tearing it apart. Thanx for taking the time to help me hash this out.

[davohilts]

how about taking 2-1 ohm resistors and running them from pin 5 of the powertubes to ground? then you can just take a voltage measurement across the resistor and you have your bias in mV which in this case should equal mA. you can dial in the bias voltage with the bias pot.