Lowering plate current in Marshall 1974X Reissue Combo Amp

[nevetslab]

A client of mine just picked up a Marshall 1974X Reissue Combo Amp and great condition. He brought it to me to check it over, while picking up his Hiwatt SS20/40 112 Combo Amp that had some warranty work done. When I looked at the Amp assembly of the Marshall,I noticed the EZ81 Rectifier tube and the two different EL84 power tubes that were installed were noticeably blackened on their plate structures inside the tube glass. I ordered new J/J EL84 & EZ81 tubes, which came in today.

Marshall-1974X-18W-Schematic.pdf

When I measured the plate current of the new EL84 power tubes, they were 56mA, and plate dissipation was 17.6W with plate voltage of 321VDC. Cathode Bias resistor was the stock 91 ohm/10W value.

This seems considerably higher than I'd expect. On Cathode bias amps, can you get lower plate current by selecting lower Transconductance in what's available from the distributor? Tubes and More/Antique Electronic Supply was the vendor. Of course, increasing the Cathode Bias resistor would also lower that.

[Helmholtz]

Cathode bias works by DC current feedback which tends to equalize variations in transconductance.

How did you determine plate currents?
What is the cathode voltage? Does it change when you lift the bypass cap?
Does it change when you lift coupling caps C10/C11?

The original 1974 had a 100R cathode resistor. Later 2xEL84 Marshalls used 125R.

[The Dude]

I'd have no problem just raising the cathode resistor value. Even if you could find tubes that bias a bit lower, I'd be concerned about the guy who throws new/different tubes in there and doesn't even bother to check bias. IMO, it's better to build in some safety/tolerance.

[Chuck H]

Well the others here that are actually dealing with vendors more frequently may have a different answer, but I'd say no. You probably can't request a set of tubes with a lower transconductance. I still remember when vendors offered "selected" tubes a couple of different criteria (though transconductance wasn't one of them). That disappeared something like fourteen years ago and IMHE it's a crap shoot if even the premium graded tubes even perform as advertised (which would indicate that they aren't tested or at least not properly). And now, with availability changes such as they are it seems even less likely vendors are "selecting" any tubes out of batches.

I guess Groove Tubes sort of did this when they offered tubes matched and graded for early or late "breakup". They didn't specifically say transconductance, but they didn't have to. A higher number meant the tube would stay cleaner and get louder with a given drive signal than a lower number. So if you wanted lot of breakup you bought a low number. If you wanted maximum headroom and power you bought a high number.

That said, there's no harm in asking. You don't get anything if you don't ask. Sometimes I ask about something or other in the course of my daily grind and I'm sometimes surprised and heartened by the response. Once in a while people will restore your faith in humanity and demonstrate pride in their product and the craft they have pursued. Mesa and Peavey are two companies that have done that for me. Though that was long ago I picked those two because they are in the amp genre.

So ask. Then take your lumps and increase the cathode resistor value

[christarak]

I'm with the Dude and Chuck. If you confirmed that both of the EL84s are healthy with a reasonable match, and the 91R resistor is within tolerance, I’d be increasing the cathode resistor to get the present 147% dissipation to somewhere between 95 and 105%, and damn the torpedoes.

[nevetslab]

Cathode bias works by DC current feedback which tends to equalize variations in transconductance.

How did you determine plate currents?
What is the cathode voltage? Does it change when you lift the bypass cap?
Does it change when you lift coupling caps C10/C11?

The original 1974 had a 100R cathode resistor. Later 2xEL84 Marshalls used 125R.

I used my Compu-bias Probes to measure the cathode current, seeing 56mA each tube, and seeing 10.3VDC thru the cathode resistor. Bias probe meter shows cathode current, plate voltage and dissipation for a pair of tubes.

So, I think increasing the cathode resistor to 120 ohms is probably the safest bet.

[nevetslab]

The 120 ohm 7W cathode resistors arrived yesterday, so I removed the stock 90 ohm 10W Cathode resistor and installed the 120 ohm resistor in it's place. Powered back up. Having seen the new tubes were matched, I just looked at the voltage across this new cathode resistor, that being 11.9VDC. Plate Voltage was 321VDC, so subtracting 11.9VDC from that, we have 309.1VDC, cathode current per tube being 49.6mA, power dissipation is 15.3W. Lower than the 17.6W by a little, though still well above the max limit published at 12W. Everybody does let these run hot, so its a little better. I'll settle for it.

[Helmholtz]

As you're measuring cathode current, actual plate current will be lower by around 5mA (screen current), resulting in a plate dissipation of 13.8W.
Essentially same operating conditions as with an original AC30.

[pdf64]

The EL84 12W anode limit is design centre. Similar to 6V6, its limit under the design max system might be about 14W.

[nevetslab]

As you're measuring cathode current, actual plate current will be lower by around 5mA (screen current), resulting in a plate dissipation of 13.8W.
Essentially same operating conditions as with an original AC30.

AH....THAT's good to know! Thanks for the info. I forgot to measure the screen current thru their common 100 ohm resistor.

[Chuck H]

AH....THAT's good to know! Thanks for the info. I forgot to measure the screen current thru their common 100 ohm resistor.

And... You can also subtract the cathode voltage from the plate voltage to use that figure as the "working" voltage. Not very significant with el84's since they don't require much bias voltage but between subtracting screen current and using the actual working voltage for the plate current calculation the actual working conditions for the tube start to look more reasonable.

[Helmholtz]

And... You can also subtract the cathode voltage from the plate voltage.

See post #7 and check dissipation numbers.

[Chuck H]

See post #7 and check dissipation numbers.

Ah. Thank you. I actually HAD assumed it would be obvious to someone as experienced as nevetslab but I didn't back read.

[nevetslab]

I was going to open the amp back up and get the actual screen current drop across it's 100 ohm resistors and then recalculate it all.

[mozz]

Is it odd that the plate voltage did not rise from 321v when you increased the cathode resistor?