Marshall JTM30 output and bias question

[Jazz P Bass]

Not too sure why that bias voltage is -37Vdc.
That is way too low for a 6L6 tube.
(did Marshall screw the pooch on this when they swapped in 5881 tubes?)

[Jazz P Bass]

So Marshall calls this a "Class A" amplifier.

In that case the -39V bias is o/k.

If you are not comfortable with it that voltage, you could increase the zener string voltages.

Fellow MEF member here, ToneMeister, has installed a bias pot mod on these amps.

[Randall]

That's what I was asking earlier to change the 9.1v zener to 15v. Would that give me another 5 - 6 neg. volts? Or, if it is Class A, and this is how it was designed, do I just put some 5881WXTs in it and send it out? I'm inclined to do the latter at this point.

[g1]

only component that shows any failure out of circuit is the safety supply cap C127, it is supposed to be .047mF, but it reads .017mF.

C127 is perhaps the most critical cap in the amp. (bias feed) Unless you have reason to distrust your cap meter, I would highly recommend you do NOT re-use that cap.
Is that cap meter something you use often and are familiar with the accuracy?

I've never tested a class X cap with a meter, does anyone here know of any reason a class X might give a meter reading different from a regular cap?

[Randall]

I just replaced it. I measured it as .017 on a Fluke 77 cap setting. The new one measured pretty close to .047, that's the extent of the accuracy here. The other parts I changed all measured OK out of circuit.

[Dave H]

I just replaced it. I measured it as .017 on a Fluke 77 cap setting. The new one measured pretty close to .047, that's the extent of the accuracy here. The other parts I changed all measured OK out of circuit.

LTSpice says the raw bias voltage is 29V with a 0.017u cap which isn't enough to turn on the zeners. With a good 0.047u cap the raw bias is 59V which is clamped to 39V by the zeners so you could change the 9V zener to 15V to increase the bias voltage to about 44V.

[Dave H]

I've never tested a class X cap with a meter, does anyone here know of any reason a class X might give a meter reading different from a regular cap?

I can't think why it shouldn't measure OK. A class X cap can fail short, a class Y cap is supposed to fail open but short or open you've still lost your bias voltage. I know it isn't directly across the winding but it doesn't feel right to have a 250V cap connected to a 325V winding.

[Randall]

I swapped out the 9v zener for a 15v, and got interesting results. bias voltage went to -42v, and somehow that was enough to drop current from 53ma to 35ma @ 442v = 15.5 watts. How could this be? I don't want to put these 5881 in class A at that level, so if it is so I shall put the 9v zener back in.

edit: I had only measured one 5881 before, probably a different one each time. WHen I measure them both the other one is at 46mA at 442v = 20 watts. That I can live with. So my 5881s are not matched, and I don't know if one is week or one is pulling too much. I guess I'll go look for another pair to compare.

edit again. A NOS pair of JAN/Philips gives 50mA and 52mA at 431 volts. That's 21.5w and 22.5 watts. All three zeners are 15 volts. Does this seem right?

[Enzo]

Sure, why not? TUbes vary all over the place. I never expect them to match or be consistent.

[Chuck H]

A NOS pair of JAN/Philips gives 50mA and 52mA at 431 volts. That's 21.5w and 22.5 watts. All three zeners are 15 volts. Does this seem right?

That seems fine. And welcome to the world of modern tubes. New tubes don't act like old ones, except some do.?. So you just have to keep buying pairs until something works like it's supposed to Actually there's usually more leeway. Since that amp is working on a "fixed" fixed bias at a fairly critical idle current it's more of a problem. Even the non adjustable fixed bias in most amps is more adjustable that this amp. Not sure your customer will want to spring for a pair of NOS Phillips tubes though.

[Randall]

Well, more confusion. With the 3 zeners all at 15 volts, and a -42 bias voltage, 6L6s back in it, pulling 46 and 47mA ar 441 volts, I am getting very low output, about 2 watts on my Fluke and 15 p-p on the scope. This couldn't be from bias change, could it? I'm getting close to 10 X the signal on the grids as I am plates.

edit: I put the 9v zener back in, bias voltage -37v, not much change in output. What the heck?

[Chuck H]

I'm surprised you changed the zener thinking the bias condition, such as it was, would be the problem with the output. I believe it's a coincidental, secondary issue. Maybe a measurement error. Because 15V p-p on the scope is a lot more than 2 watts on the meter.

[g1]

Because 15V p-p on the scope is a lot more than 2 watts on the meter.

That would depend on the load impedance. I get 1.7W @ 16ohms.

[Chuck H]

That would depend on the load impedance. I get 1.7W @ 16ohms.

Yeah, ok. I admit I fact checked that and did the RMS math AFTER writing. Apparently forgetting that 16 ohms is a common standard on many Marshall models. Just the same...

It's not being caused by the bias condition.

[Randall]

I didn't think the zener was probably the cause of the low output, but I wanted to rule it out, before someone asked me if I had tried that. My power readings are just rough here, the point is why is it so low? I have plate voltage, lots of idle current, and gobs of signal on the grids, but a very low output. The OT measures 136 ohms across, with center tap at 75 and 61 ohms.

Any ideas?