JCM900 50w dual reverb low on power

[Alex R]

Well this forum was so helpful with that ol Hartke I thought I'd run another by you. This is one of those little 1x12 JCM900 50w 2xEL34 dual reverb combos. It puts out 14watts only and sounds tinny; it got this way at a gig all of o sudden. Here are my observations:

• Preamp puts out a good signal
• Output valves test fine and new subs make no difference
• Screen resistors fine
• Power reduction ('triode' switch) works as it should
• Substituted new filter caps too - no difference
• Bias is ok - current draws are v unbalanced though, despite the fact that the valves match ok on the tester, but it is possible to get 30-40mA across the legs of the OT at 440v idling.
• OK Voltage shows up on the plates and screens at idle, and bias voltage on the grids
• Cathodes are grounded and only develop about 1v when it's fully driven, presumably through D6,D7, R36 etc (what are that little lot for, by the way?)
• However, the waveform on the plates has about the same amplitude as that on the grids, with what looks like a bit of crossover distortion added.

Sounds like the OT? Tried hooking up a good stand-up Marshall 50w OT. The amp put out just under 30w through that. Better but still not right - this may be just the difference between the smaller JCM900 OT and the fatter JMP type I subbed in?

Resistance on the legs of the JCM900 OT are 33 ohms and 42 ohms. So maybe it has some fused windings, or maybe not anyhow, that's why the currents don't match.

The mains transformer? Well maybe, I'd like your opinion. It does run hot, but then I've seen that before in these amps. Fully driven, the amp draws 0.75w from 240v AC mains. The plates idle at 440v, but sag to 389v when driven hard.

Like I said, it's not the filter caps. Looks like I need both new power and output txrs. What do you reckon? Am I missing something? I've had to replace quite a few of each on these amps, but I won't badmouth them because Enzo will accuse me of observer bias again.

The guy says it goes through quite a few power valves, one a year or so. I think he drives the poor little box quite hard. I was thinking of putting raw bias voltage on the suppressor grids and lifting them from earth a la Traynor, what do you think?

Enough questions already.

[Alex R]

"Fully driven, the amp draws 0.75w from 240v AC mains."

0.75 A-M-P-S. Jeez.

[Alex R]

Schematic attached, or at least that was my intention.

[Amp Kat]

When you said 32 and 42 ohms. Is that from center tap ? What does it read across the winding ? Check your plate load resistors and make sure they are all good. Preamp and power.

[Alex R]

Yep, from centre tap to the two plate connections. The problem is definitely after the PI stage, big healthy signals to the power tube grids, then it all goes wrong. I thought those winding resistance weren't too far out for an EL34 OT... not totally sure...

[Alex R]

...and across the whole winding, the sum of the two halves right enough.

[Amp Kat]

Those numbers seem right for that OT. I think you have another problem that's not staring you down. Maybe a tube socket connection or even speaker connection ? However most Marshall 900 trannies are very close from ct to each tap ect.. 32ohms/34ohms/ 66 ohms and I still haven't seen one yet that equaled up and was bad. You would also assume that if you did RG's test and fed a very small voltage from secondary and measured the primary they would be dead on as a ratio of the taps and winding. Have you tried connecting a cord from the send jack to the return. Could be your loop jacks.

[Enzo]

Yeah, I don't get power amp vibes from this one either. Of course I could be wrong.

WHAT JCM900 amp do you have - model numbers please!!! DO out have the 4500 with its partly solid state front end? Or the 2500 with its more or less tube preamp.

What happens if you pull the OT secondary wires OFF the board and clip a speaker directly to them? Eliminates certain possibilities from the output jack board.

When you clipped in aanother transformer, was the NFB then connected to it? If not, that would explain the little louderness.

If you clip in another transformer and connect its secondary directly to a speaker, even if it is louder, does it still have that thin sound?

What does it sound like if you apply the signal to the loop return? Or patch across the loop as KB suggests?

The first thing to my mind from your description was crosstalk. The curcuit through the amp is stage to stage to stage, like the cars of a passenger train. If any one stage cannot pass the signal along to the next, the signal path is broken. But there is enough gain in the amp that later stages can still pick up the signal radiated from earlier stages, and it sends THAT along. That "signal" will inevitably be weaker and less robust than the original. Imagine a message passed along from the rear of the train towards the engine. In the middle of the train, the door is stuck shut, but the next car up can still hear the message from the car further back through the door.

Two stages coupling signals together is called crosstalk - imagine a hot signal on channel 2 of a mixer. And on this mixer, if you brought up channel 3, you could hear that signal from channel 2 a little. That is crosstalk, and not desired ina mixer. In a guitar amp it is sequential stages and not an issue, unless it is all you get.

I'd go through the whole amp stage by stage from the front input and see if the signal breaks somewhere in there. Be mindful of teh various voltage dividers.

D6,7, etc are circuit references to chassis. The chassis is NOT the circuit ground. CLose, but not ground WHen you are measuring several hundred volts of B+ or even -50v of bias, the little difference won't matter. But when yuo try to measure the cathode voltage of something which is to be some millivolts or a volt, then the difference matters. FInd real ground for such measurements

[Alex R]

Thanks Enzo. This is a 4501, sorry not to say that. I took another look in the light of your comments and observed the following:

With a speaker hooked up directly to the disconnected secondaries (grounded properly) it's just the same.

With the other transformer it's low output and thin sound the same. I understand about the absence of NFB accounting for the wattage difference; that would indeed seem to discount the likelihood of a bad OT.

power tube voltages today with the OT connected direct to a speaker:
grid -44.5v
screen 455.8v
plate 459.1v
cathode to circuit ground resistance 0 ohms

This is the funny thing: with the trace on the grids nice and clean and several volts amplitude - normal-looking - the trace on the plates is the same amplitude but with what resembles heavy crossover distortion - notches half way up and down the waves, no clipping. Yet the valves seem to be biased ok, drawing at the last count 31 and 40 mA respectively, pretty much in line with the unbalanced primary resistances.

Do you think that 60v or so of sag on the power tube plates between idle and fully open is normal, when the output maxed is only about 10 watts? Power tfr problem? (NB it's not the filter caps.) I'm stumped - the only idea I have to account for all this is that either both tfrs are faulty ot the power supply has issues. Maybe I should look if the bias voltage sags when you turn it up? Or somehow rises... I will have a look next time. But in truth I'm out of ideas and strongly suspect something is staring me in the face...

Many thanks for the help so far...
Alex R

[Mars Amp Repair]

Just a thought,
At full power (just before clipping) the drive voltage on the grids of your power tubes peak, should just about equal the DC bias voltage there...IOW, if the bias voltage is -44 volts, then the drive voltage should be about 44v Peak.

That helps in knowing if the drive voltage has enough amplitude at that point in the signal path.

If you really only have 2vp-p on the control grid of the power tubes, then I'd have to say the problem is not in the power stage. glen

[Enzo]

I made my suggestion, I don't think it is the power amp, I think you have a break in the signal path in the preamp.

Did you inject a signal at the return loop jack? Did you send the signal from the loop send to another amp to see what it sounded like? Classic divide and conquer. WHich half of the amp has the problem?

Rock the amp to crash the reverb pan. Is that noise loud and strong or is it also thin and weak? If strong, then everything after the reverb, including the power amp, is OK. If weak, then problem is after the reverb.

With the 450x you have all those op amps in the preamp, as opposed to the 250x which is just tubes. SInce both channels are affected, we look for somehting they have in common. First thing I see is IC1, but also the two preamp tubes - got signals on those where there ought to be?

[Alex R]

Sorry, I should have been clearer about this.

It's not possible to get more watts out of it by injecting signal into the effects return or indeed the reverb return - the distorted tinny low-wattage output remains.

Respect to the suggestions that it is a preamp fault; I ought to have made my reasons for not thinking so because of the signal on the power valve grids clearer than 'several volts'. I'm used to what a good signal for power valve plates looks like on the scope and didn't measure it. The preamp can put a clean sine wave of 31VAC peak to peak onto the power valve plates - above this it begins to distort as Marshall preamps do but it can go up to 44VAC with distorted waveforms. With that 31VAC clean sine wave on the power valve grids there is 32VAC on the plates - with what looks like heavy crossover distortion. Just as if the thing was struggling with way too much bias voltage. This is the bit I don't get as the bias seems to be ok to judge by how it idles. Never saw one like this before. I was wondering if it has a huge lot of sag due to a power supply fault. The plates drop by 60v under full signal; I don't know whether this is normal. Still puzzled and open to suggestions...

[Enzo]

Fair enough.

SOunds like you are scoping it, so I assume we can rule out HF oscillation?

I can't visualize the innards. All pcb or hand wired sockets on this one?

When you subbed OTs, were the plates still connected to the pc board? Any possibility something is shunting the two phases together?

Trying not to make too many assumptions, did you actually test the existing OT as KB suggested?

[Alex R]

Yep, no sign of oscillation on the scope.

This is the small box 1x12 combo. Sockets are handwired and sit half under the narrow pcb on the back panel on which are the output jacks etc.

Yes the plates were still connected to the board. I will look at this.

I didn't do the backwards small-AC-voltage test on the OT and will do so.

...more later...

[Joel DC]

New to this forum and I ain't a tech and I know this may sound stupid but have you tried a set of 5881/6L6 tubes in it? I remember reading about a similar problem coupla years back and the 5881's made the difference.