Marshall JCM800 50w Head 2205 - it's lacking power and so am I

[mikeydee77]

Hi guys,

I would really appreciate a second pair of eyes over my thinking here. This is the Marshall 50w twin channel, reverb model as per attached diagram.

It came to me not barely working and what I have found is a few little faults that have gradually brought it back to life - but it's still lacking. This is what I have tackled so far:

0) output transformer has been replaced at an earlier time and one of the connections to the power tubes was particularly dry - cleaned up and resoldered
1) filter caps were ruptured - replaced with new FT CAP
2) A few pre amp plate resistors were a little high e.g. 120K instead of 100K (replaced with new CC 0.5W)
3) The first 10K power supply resistor had drifted high - replaced
4) Numerous dry joints.


After all that it works and sounds nice enough. But it isn't all that loud; there are two channels and both have the same problem. With both pre and master volume on full it's loud, but it only feels like 20% normal volume to me. It has a send and return loop (thoroughly checked those pesky sockets) and if I connect an input straight into the return then the output is much stronger than through the normal pre amp input, and because both channels are weak I have been concentrating my investigations around V3.

I have attached some voltage readings and the circuit diagram. It would be great if someone from the forum could point some obvious ommission in my process.




A couple of other observations is that although my meter has an input impedance of >10Mohm I noticed two strange things. 1) when I measured the voltage on either grid of the PI the volume of the audio being played at that time dipped a little and 2) All the high voltage readings appeared as one value on my meter and then settled on a lower one e.g. V1 Anode measure 290v just for a second and then 190v as though the load of the meter was just too much to bear. Can't be right can it?

Any thoughts . . . .

[nickb]

V2 , triode 1, cathode voltage 0.9mV is not consistent with the plate voltage. Did you mean 0.9v?

V5 PI: grid to cathode voltage is such that the tube would be in cutoff. You meter is loading the grids giving an incorrect reading, putting the tube into cutoff and killing the audio, Looks like you meter is about 3M ohms.


Sorry for repost. Forum went to sleep and I thought my post didn't go. Edited read V2 ( was V1)

[Enzo]

Yes^^^

The only way to measure the voltages on the grids of the PIs is to read from grid to CATHODE, not grid to ground.

[mikeydee77]

V2 , triode 1, cathode voltage 0.9mV is not consistent with the plate voltage. Did you mean 0.9v?

V5 PI: grid to cathode voltage is such that the tube would be in cutoff. You meter is loading the grids giving an incorrect reading, putting the tube into cutoff and killing the audio, Looks like you meter is about 3M ohms.


Sorry for repost. Forum went to sleep and I thought my post didn't go. Edited read V2 ( was V1)

Thanks nickb - it is of course 0.9v that was a typo.
On the other point perhaps my meter is playing up. It's a low to mid budget meter made by Tenma although I've never had any doubts about it before. I do have a couple of other meters (one motor mechanic digital (very budget) and an AVO8 analogue which I know has a much lower impedance and also a scope - still interesting to compare against.

[mikeydee77]

That's interesting I didn't realise that about the PI. I will recheck and post the results.

[nickb]

Having a scope will help you diagnose your original problem. Suggest you start by measuring the output power into a dummy load to see if the problem is real.

[drewl]

Yeah, these amps will seem to not be putting out power unless the master is full up.

Also check the fx loop jacks, a common failure.
Run a signal generator with a couple volts ac into the return jack- is the power amp putting out full power?

[eschertron]

output transformer has been replaced at an earlier time

could there be an impedance issue with the OT that is causing the problem? I can't answer that, but I thought to throw that 'into the ring' since you didn't mention any other non-original part.

[mikeydee77]

Thanks for alll your thoughts and suggestions guys. A small update having measured the output power of the amp and tried a few other things.

So, I fed in a 150mV p-p signal at 1KHz sine wave measured the output as 65v p-p until it started clipping (clean channel). So if I am right the RMS is calulated as:
( 32.5 x 0.707 ) squared / 8 ohms; which is
22.9 x 22.9 / 8; which is
524.41 / 8 = 65 watts rms
Lead channel was similar although clipping started at about 60v p-p

Well if my working is sound then thats a pretty decent output so what am I talking about saying that it is weak sounding. Well my reference point could explain that - it is another Marshall; a 2204 50w amp (single channel, no reverb) but with 6CA7 tubes biassed at ~60mA and that one really barks it out. So does that sound ok?

I think there are still some dry joints to sort out but that seems to be a feature of these amps.

[mikeydee77]

That's a good point.
I checked over the wiring colours when I resoldered the poor OT connections as it was a bit of a mess - and they DO match with the original diagram and in fact the previous repairer had left tails dangling to help verify correct orientation of the wires so I was fairly confident that it had been done correctly and cleaned out the selector switch but worth another look. Is it the right part might be another question though?

Other than that amps is original apart from a couple of replaced pots and accompanying PCB track bodge job.

[mikeydee77]

DC Voltage cathode to grid is -1.1 volts on both grids of PI

[acorkos]

make sure to reflow solder on all the front panel pots

[mikeydee77]

Good point. Although I reflowed them initially as standard practice I later found a doubtful join. Only upon a second reflow I realised that this approah is not good enough for this PCB. Only by fully de-soldering the pots, thorough cleaning of the pads and pot legs could I achieve a reasonable join. Unfortunately, since only the pot legs support the PCB, and nothing else, I guess that it's an inevitable problem that will return. I was wondering if anyone had ever thought about re-manufacturing such boards with pots flying and panel mounted and the board mounted on pillars. Would seem quite viable if the PCB is fairly widely used.

[Alex R]

Good point. Although I reflowed them initially as standard practice I later found a doubtful join. Only upon a second reflow I realised that this approah is not good enough for this PCB. Only by fully de-soldering the pots, thorough cleaning of the pads and pot legs could I achieve a reasonable join. Unfortunately, since only the pot legs support the PCB, and nothing else, I guess that it's an inevitable problem that will return. I was wondering if anyone had ever thought about re-manufacturing such boards with pots flying and panel mounted and the board mounted on pillars. Would seem quite viable if the PCB is fairly widely used.

This problem is talked about more often than it actually occurs. I've seen it once in looking at perhaps 30 jcm800s over the last ten years. Reflowing the joints is enough to cure it.