JCM900 Low Output

[rf7]

I'm working on this JCM900 2502 combo and I'm only getting 36 Watts RMS out of it. The tubes are new and biased properly, but it's only putting out 72% of what it should. I've fed it thru the receive jack and it's still low. I'm running EL34's in it. It used to have 6L6's, but I changed a bias resistor to get the bias voltage correct.

I've done the following:

1. Tried other tubes.
2. Checked the OT out of circuit.
3. Swapped in another OT.
4. Screens are ok. 2.2Kohm as standard. Grids are 1.5Kohm, with a 2.2Kohm added in series on one side?
5. Tube voltages look reasonable.
6. Triode/Pentode switch cuts power as expected.
7. Tried it with 6L6's too.

The plates are sitting at 440V at idle and drop to 398V at clipping. I think that's reasonable.

Filter caps are original, but the output is not exhibiting any 120Hz or 60Hz hum.

What am I missing? What else could cause this?

Thanks in advance!

[jamesmafyew]

Sounds like maybe not enough signal voltage on the power tube grids? The FX return does not feed the phase inverter directly, there are still a couple of opamp stages to go through. The master volume pots are variable resistances sitting in the switching dual opamp's feedback loops. Maybe check in there to see if there are resistors not measuring as they should, check the opamps IC1 and IC2. Plus the PI area too.

[Chuck H]

+1
Check with a scope to see that when trying for a "clean" sound the power tubes are the first to clip. If an earlier stage is clipping ahead of the power tubes (due to low signal at some other point in the circuit) then the amp can never get to full output.

Check the output at each stage to be sure there is enough drive to clip the power tubes first before any preamp distortion. If not then you need to trace the signal and isolate the early clipping stage.

[rf7]

It does not seem to be the preamp at all. I pulled the power tubes and set the amp up so that the PI just starts clipping with the Volume at 10. The input to the PI is not clipping at this point, just the output. With the power tubes back in the amp, the output starts clipping around 8 on the volume knob.

On a side note, if I run the amp set for 8 ohms into a 16 ohm load, I get 39Wrms output. And if I run the amp set for 4 ohms into an 8 ohm load, I also get 39Wrms. That's weird, since I'm only getting 36Wrms with the load matching the output selection. The 16 ohm tap into 16 ohms gives me 38Wrms.

Also, the lower half of the output waveform starts to clip before the top of the wave.

It's all very puzzling.

[Enzo]

You didn't report it, but may I assume that the screen pins on both tubes are getting B+?

James suggested the grid signals. Scope the power tube grids. Are BOTH getting strong signal at the grid pins? And following back then, is ther strong signal at both plates on the PI? No scope? Use an AC voltmeter. Are the plate voltages about even on the PI?

If one side of the push pull is not being driven somewher along the line, you will have one side of the waveform clipped off part way after a certain level, the part the other side would have handled.

[rf7]

You didn't report it, but may I assume that the screen pins on both tubes are getting B+?

Yes.

James suggested the grid signals. Scope the power tube grids. Are BOTH getting strong signal at the grid pins? And following back then, is ther strong signal at both plates on the PI? No scope? Use an AC voltmeter. Are the plate voltages about even on the PI?

If one side of the push pull is not being driven somewher along the line, you will have one side of the waveform clipped off part way after a certain level, the part the other side would have handled.

I had a chance to compare this with a JCM2000 head, and the JCM900 PI is definitely clipping early. On the JCM2000, the output tubes clip before the PI plate signals clip.

The PI plate voltages look OK.

I pulled BR1, the diode bridge, because it was causing clipping before V2a. Now the output of V2 is fine but the input to V3b is clipping early, it looks like.

Gotta check into that next.

[rf7]

Scope the power tube grids. Are BOTH getting strong signal at the grid pins?

Yes.

And following back then, is there strong signal at both plates on the PI?

Yes.

Are the plate voltages about even on the PI?

241V and 250V.


OK, I just learned, courtesy Mr. Frondelli, that the second half of the long tail pair is run common grid, so the input comes off of the cathode. I've been chasing my tail on that, looking at the grid. Oh well, now I know.

I'm seeing 34Vrms on both outputs of the PI before clipping. But I'm only getting 17Vrms into 8 ohms out of the OT. Which is 36Wrms at clipping. Are the outputs of the PI strong enough?

[J M Fahey]

OK, so you have 240/250 V on the PI plates:
1) what +B voltage do you have on the rail feeding them (i.e. "the other end" of plate resistors)
2) What voltage do you have on both PI cathodes?
3) when the PI plates signal clips: is the input signal driving it still clean?
Does one of the plates clip significantly earlier than the other one?
Thanks.

[km6xz]

I am a little confused. You are looking at the PI and preamp, but also say "The plates are sitting at 440V at idle and drop to 398V at clipping. I think that's reasonable."
Are the output tubes saturating when driven hard or not? If they are and screens are OK, look at the output matching, the selector switch for example, wiring etc.

[Chuck H]

I might test the OT in this case just to be sure. Other than that I've been guessing at a bias condition. Bias voltage is not bias current. No more than frequency equates to volume. I would test the OT for a proper impedance ratio and check the bias CURRENT next.

[g1]

Agree with Chuck, the alarm bells started going off when you mentioned higher power with impedance mismatch. Is it the proper stock OT?
Otherwise, if it were me , I'd be taking a step back and verifying my test gear/setup, load box etc.

[rf7]

OK, so you have 240/250 V on the PI plates:
1) what +B voltage do you have on the rail feeding them (i.e. "the other end" of plate resistors)

Those plate voltages were taken with the EL34's out, so the PI plates are sitting at 237V and 224V. The B+ is 371V.

2) What voltage do you have on both PI cathodes?

It's 39.4V.

3) when the PI plates signal clips: is the input signal driving it still clean?

Yes.

Does one of the plates clip significantly earlier than the other one?

No, they seem to clip at the same point. I checked this with the EL34's removed.

I might test the OT in this case just to be sure. Other than that I've been guessing at a bias condition. Bias voltage is not bias current. No more than frequency equates to volume. I would test the OT for a proper impedance ratio and check the bias CURRENT next.

OT Impedance Ratio: 392:1

So, 6.27K:16 ohms, 3.1K:8 ohms

Bias current (measured at tube cathodes): 35mV and 30.5mV (not the greatest matching). Plates are 440V.

[J M Fahey]

Thanks for measuring.
I was somewhat suspecting that the PI might be either improperly biased or fed, but I must discard that possibility now.
Oh well.
Everything looks normal.

[jrfrond]

I've been watching this thread, and from the looks of it, all is normal for a typical 50W amp. So, I must ask this question: exactly HOW are you calculating your power output? How are you deriving your RMS value? What's the peak power reading? I'm not saying that you are doing it wrong, but everything else LOOKS right. I just have the sneaky suspicion that you MAY be chasing a ghost that isn't there.

[rf7]

I'm running the amp into an 8 ohm 300W resistor and measuring the AC voltage across it with a DVM. It starts clipping at 17Vrms.

Then V squared/R = (17*17)/8 = 36.125W for the power.

I can get it to put out 50W but that's pretty much fully clipped.

I'm used to getting closer to 20Vrms out of 50W amps.

I would calculate peak power to be 48W using:

(Vpeak)Squared/2R = (Vrms)squared/R