Another JCM2000 bias question

[Randall]

Sorry if I am repeating myself, I have been away from my shop with illness. I installed a Dr. Tube stable bias mod in a Dual Super lead 100. It was suscessful, but I couldn't get the tubes to bias less than 19 watts, which I thought was too high. I adjusted a resistor, and got it to adjust into range, but then found a Marshall adjustment bulletin in a couple of Marshal forums that call for clipping on the 3 pin connector at the back, and putting each side at 90mV. With a B+ of 475, that's 21 watts, if I am not mistaken. Thinking something wasn't quite right, I went to the bias probe method, this time measuring 37.5mA, which was better but still high in my estimation.

What is going on here, and where should I bias this thing, givien these amps tendency to burn up?

[DrGonz78]

The marshall bulletin stating 90mv is too high honestly, like you said. I usually set JCM2000 DSL or TSL 100's so that it is around 60% dissipation. What I always encountered testing bias on these amps is one tube out of the four would eventually be pulling a ton more current than the other three. Moving the tubes around configures the current a bit side to side but still only matters if all tubes are close. That is why I would only judge overall bias by looking at each individual tubes bias and not simply with that 3 pin connector in the back.

[Jazz P Bass]

Just a suggestion.
For a 100 watt amp, measure the current through the amp from the mains.
For a 100 watt amp set the bias that the current reads 120 watts.

[Helmholtz]

Just a suggestion.
For a 100 watt amp, measure the current through the amp from the mains.
For a 100 watt amp set the bias that the current reads 120 watts.

That relation may be valid for a specific amp model, but hardly for other 100W amps. Power consumption and especially PT efficiencies vary.

Also the method won't tell if one side runs too hot and the other side is consequently biased too cold.

[Randall]

I used the bias probe and put it at 60%, and am calling it good. I don't much like the rear pin method, it doesn't seem very accurate. Thanks again guys!

[g1]

I don't much like the rear pin method, it doesn't seem very accurate.

Both methods rely on 'sensing' resistors, and they should be precision types. The internal ones in the Marshall are 1 ohm, the ones inside your bias probe are probably the same. Unless your bias probe requires your meter to be in DC current measurement mode. In that case it will not have a sense resistor.
If the resistors are a bit off of their value, you will get differences in the readings like that.

[Helmholtz]

Both methods rely on 'sensing' resistors, and they should be precision types. The internal ones in the Marshall are 1 ohm, the ones inside your bias probe are probably the same. Unless your bias probe requires your meter to be in DC current measurement mode. In that case it will not have a sense resistor.
f the resistors are a bit off of their value, you will get differences in the readings like that.

In case of doubt there is a third method which gives real plate current: The OT primary voltage drop method.

[g1]

In case of doubt there is a third method which gives real plate current: The OT primary voltage drop method.

It does, but I'm not sure I trust the resistance reading to be highly accurate. It does fluctuate somewhat due to temperature, and it's not possible to check it at the same time as the voltage measurement. Certainly the reading it gives is close enough, just maybe not perfect.

[Helmholtz]

It does, but I'm not sure I trust the resistance reading to be highly accurate. It does fluctuate somewhat due to temperature, and it's not possible to check it at the same time as the voltage measurement. Certainly the reading it gives is close enough, just maybe not perfect.

I know.
Idle current heats up the primary winding and slightly increases DCR. That's why the DCR should always be measured shortly AFTER the voltage measurement.
But this method gives real plate current, while the cathode current contains the screen current which doesn't contribute to plate dissipation.

And finally there's the OT primary shunt method. This requires some care to avoid accidental shorts: Clip on mA-meter leads BEFORE turning on the amp, turn off, discharge filter caps and clip on to the other side.

[g1]

And finally there's the OT primary shunt method. This requires some care to avoid accidental shorts: Clip on mA-meter leads BEFORE turning on the amp, turn off, discharge filter caps and clip on to the other side.

It's critical with this one to know the meter's resistance in 'shunt' mode. It is sometimes higher in mA range than A range. And can be higher than you would expect when using less expensive meters. This is especially a problem with OT's that have low primary resistance.
For an example, one of my meters has 10ohm resistance in mA mode. Across a 40ohm OT resistance, you will not get an accurate reading at all.

[Helmholtz]

It's critical with this one to know the meter's resistance in 'shunt' mode. It is sometimes higher in mA range than A range. And can be higher than you would expect when using less expensive meters. This is especially a problem with OT's that have low primary resistance.
For an example, one of my meters has 10ohm resistance in mA mode. Across a 40ohm OT resistance, you will not get an accurate reading at all.

Thanks for reminding me. I actually mentioned that method just for completeness, never used it.

I remembered that I once verified that my meters have a shunt resistance of 0.2R in the 2000mA range, which seems good enough.
But as the default mode is autorange, it will switch its shunt to 2R in the the 200mA range. And that would already introduce a significant error with high power OTs.

So I agree to not recommend the method.

[Pedro Vecino]

Two or three months ago one DSL100 had one side set to 90mV and the other to 65. What was the problem? That the 65 mV side had a tube deactivated. It was just an EL34 set to 65 mA.
I have never paid attention to those systems for measuring tube pairs. I always measure them individually.