bias procedure for deluxe AB763

[mikepukmel]

Hi All!
For those of you who celebrate, Happy Thanksgiving!

Going to take a crack at setting the bias on my deluxe AB763 build later today, would be nice to do this correctly.

I have 1 ohm resistors mounted from the cathodes (pin 8) to ground. Measured around 1.1 or 1.2ohms with my crappy meter.

I have the idea to measure the voltage drop across that resistor and convert that to a current. so far so good.

Not sure about which plate voltage to measure. Is it plate to cathode (pin 3 to pin 8), or plate to ground (pin 3 and the other clip on the other side of the 1 ohm resistor), or does it matter since there's only a 1 ohm resistor in there?

More info:
These are Electro Harmonix 6v6's, web says:

Cathode current not more than 80ma (that sounds awfully high based on other articles here and there, they talk about current max in the 30 - 40ma range)

Plate dissipation not more than 14W

Some people say bias to about 60%?

Thanks!
Mike

[nevetslab]

Assuming the 6V6GT power tubes are aged (been in the amp for some time), you'll probably find the voltage you measure across your newly installed cathode resistors are not quite the same. And, I assume you removed the ground connection from pin 8 on both of your power tube sockets, so you do indeed now have the 1 ohm cathode resistors to ground.

That being the case, measure the voltage across the 1 ohm resistors at idle, no signal applied or no tremolo running. The voltage in mV will directly translate to current in mA, so if you're seeing 23mV on one and 28mV on the other, it translates to 23mA on one tube and 28mA on the other. To be more accurate, use the actual resistance measured of the two resistors you installed to take out the error. If you're looking to see what the plate dissipation is, then you'd want the plate voltage at pin 3 so you could calculate power dissipation. There, you'd multiply the plate voltage times the cathode current. There is some current flowing thru the screen resistors, but for simplistics sake, consider the cathode current to be the plate current. That will then give you the plate dissipation.

Adjusting the bias pot will vary the plate current that you're measuring across those cathode resistors. I typically set the Deluxe Reverbs up at 25mA per tube. If, you're seeing greater than 20% difference in the two tubes, you might want to replace them with a matched pair.

[mikepukmel]

Assuming the 6V6GT power tubes are aged (been in the amp for some time), you'll probably find the voltage you measure across your newly installed cathode resistors are not quite the same. And, I assume you removed the ground connection from pin 8 on both of your power tube sockets, so you do indeed now have the 1 ohm cathode resistors to ground.

That being the case, measure the voltage across the 1 ohm resistors at idle, no signal applied or no tremolo running. The voltage in mV will directly translate to current in mA, so if you're seeing 23mV on one and 28mV on the other, it translates to 23mA on one tube and 28mA on the other. To be more accurate, use the actual resistance measured of the two resistors you installed to take out the error. If you're looking to see what the plate dissipation is, then you'd want the plate voltage at pin 3 so you could calculate power dissipation. There, you'd multiply the plate voltage times the cathode current. There is some current flowing thru the screen resistors, but for simplistics sake, consider the cathode current to be the plate current. That will then give you the plate dissipation.

Adjusting the bias pot will vary the plate current that you're measuring across those cathode resistors. I typically set the Deluxe Reverbs up at 25mA per tube. If, you're seeing greater than 20% difference in the two tubes, you might want to replace them with a matched pair.

Thanks Nevetslab! These are brand new EH tubes. Probably less than 15 minutes on them. Thanks for the info! I will print this one.

One of the sites mentioned a rough estimate of 5% current flowing through the screen, is that a good rule of thumb?

[Jazz P Bass]

Aside: ideally you want to use 1 ohm resistors with a 1% tolerance.

[mikepukmel]

Aside: ideally you want to use 1 ohm resistors with a 1% tolerance.

Thanks I think they are, will check.

[mozz]

Tubes should be close, they don't have to be matched. Some like the sound of mismatched tubes better.
Tube bias calculator.
https://robrobinette.com/Tube_Bias_Calculator.htm

[mikepukmel]

Tubes should be close, they don't have to be matched. Some like the sound of mismatched tubes better.
Tube bias calculator.
https://robrobinette.com/Tube_Bias_Calculator.htm

AES said they're a matched pair, but I haven't measured anything yet. Thanks for the link, reading now ...

[mikepukmel]

Read through the various methods. seems like people who do this like the method where you measure dc resistance across each half of the OT (from plate to center tap), then measure voltage drop with the amp running, same places, figure out the current. Then, measure plate to ground on each tube. Finally use the 2 current values and 2 plate-to-ground voltages to compute power for each tube.

Does anyone know why this method (ot coil dc resistance) would be better than using the computed current across a small resistor installed cathode-to-ground? The first method would compute charge flowing out of the tube, and the second would be charge flowing into the tube. Don't you really want impedance, but that is frequency dependent? I mean, its power *at* a given frequency, right?

[g1]

I think it's simply because most amps don't have the cathode resistors, and with the OT method you don't need to install them.
Impedance doesn't factor in, we are just measuring idle current and using the DC resistance of the winding.

[mikepukmel]

I think it's simply because most amps don't have the cathode resistors, and with the OT method you don't need to install them.

Awesome, thanks. Maybe I'll go through both methods, first time through. Not so worried about a particular percentage, just don't want to melt anything expensive.