Marshall JCM2000 optocouplers and drive channel queries.

Psygoat

Techy curiousity question

I have been repairing a JCM2000 DSL.

I've got to say they are a level up in complexity from the 800's and 900's that I have worked on before.

 The channel switching is a bit of a maze and I can't see any clipping diodes in any of the circuits so am I right in thinking that the distortion is derived from applying different levels of gain to the tubes in the preamp section?

Does this make it an amp that overdrives the valves, or is something different happening here?

ICBM

Yes, it's all valve overdrive. However, you're right - the circuit is ridiculously complex.

There are also several things you need to do to reduce the risk of major catastrophic failures which these are prone to (not connected with the preamp).

1, if the main board is an earlier version the tips of the stand-off legs of some of the 1K screen resistors come perilously close to the wide traces for the filament wiring - within 1mm, and there's 470V across the gap, so they're prone to arcing there. (Next to the R of R74, but not only there.)



With a pair of fine-nose pliers, simply bend the points of the legs up so they're no longer as close to the traces.

2, on the back board there is a bit of unnecessary switching in the speaker output circuit - the current for the 8 and 4-ohm jacks goes through the contact in the ground connection of the 16-ohm jack. This means that if the amp is running at 8 or 4, and the jack switch doesn't close completely fully - which often happens if it gets stressed by the cable getting tugged - the amp will have no load and blow the output transformer. Jumper across the switch with a wire.

Although less serious - 3, also for 8 and 4 ohms, the selector switch only uses one half of the DPDT - this is stupid, it's far more reliable to use both halves in parallel. Jumper these as well.

This shows both of the above done - 



4, if it's a 100, near the last output valve socket there is a cap, C46. This is a band-aid solution to prevent arcing on this socket, but it isn't high enough rated (500V) and often blows. You need to replace it with the highest-rated cap you can get to fit, at least 1KV and preferably higher. Why this position is capable of generating such extreme voltage spikes I don't know, but I've seen a lot of these caps fail.

5, check the jumper connectors for the bias adjustment board - they're prone to making bad contacts, leaving one (or one pair in the 100s) of the EL34s with no bias. Then this happens...



If in any doubt, remove the connectors and hardwire the jumper cable to the boards at both ends.

There are other issues too, notably bias runaway and arcing through the thickness of the main PCB (!) which early versions of the board are especially prone to, but I don't think they ever managed to make them completely failure-proof even after more than one revision - they're just badly designed.

Did I say I don't like these amps?

Psygoat

Hi.

Thanks very much for the info. Very interesting. 

This JCM2000 had the fault with the 16 ohm speaker output switching jack, exactly as you describe so obviously a weak point. The impedance selector switch also had dry joints on it, so it was disconnected too.
Fortunately, the HT fuse blew, before the output transformer did, which was a lucky break.

There seemed to be a lot to go wrong with in these, and the build is nowhere near as good as JCM800 or 900 and I wasn't really a fan of the multiple relays, which just seemed a bit over engineered and just added another mechanical component which will fail eventually.

This particular unit had seen a lot of use, and I didn't see any evidence of arcing, but I can see from your pic, how close the screen resistors stand offs are to the traces.

Because the schematic is in sections I was curious to trace the signal path, well. I was just curious about the whole design, because there seemed to be so much stuff in what looked like a fairly straightforward head.

Cheers. M

Alterlifeson

They do seem to be overly complex compared to their predecessors. I stupidly did the C14 mod to mine about 10 years ago and it is now giving me all sorts of hassle. I am sure it is down to that. 

I would now want to reverse the mod anyway! It made the jump between channel 1&2 not so severe but the use of channel switching has gone down now.

Psygoat

It could be that the relays need replacing.
I did this in a hi fi amp recently and it made a big difference to the sound quality, much more than I expected, but it was intermittently cutting out on one channel or the other.

CHrisP86

These kind of comments on the DSL’s are mostly why I chickened out of doing the ‘Joey’ mods on mine. 

They look awful to work on.  

Psygoat

As far as I know this amp is still going.
Thanks for all your help.