Marshall JCM 2000, DSL of 2003

C13 100V 1uF at V1(A) cathode , of all things, is highly sensitive to hot air. Monitoring the hum over the output load. Once force warmed and hum level risen it is reluctant to go down again from natural cooling or from freezer spray - what process is going on with in it? Now to replace by fudge fitting to in situ component leads or take the whole thing apart to replace properly? Marshall saw fit to bodge 3 resistors by cutting off the originals and flying replacement ones to the cut wires - so whats good for the goose ....

Fudge it to verify the fault then do a proper number.

Rheilly P

hot

freezer

whole

....

Any ideas about the origin of the "heat ramping" effect? Allowing the whole amp to cool for 1/4 hour obviously resets the effect

This might be the infamous output PCB fault. Keep metering the bias voltage from the 3 pin connector provided, (Should be

90mV both sides). If it starts to climb/runaway in sync with the hum increasing, then that's what the problem is.

You will need to buy another PCB from Marshall if so.

Gareth.

If it is the output PCB, it can very quickly destroy the output tubes by putting HT where it shouldn't, so do not put any new ones in there til you have eliminated this possibility.

Gareth.

rising.

input

be

I will check that, I could understand with that old lino or whale hide they used to use on Fenders. Marshall fault , some chemical getting capilliary fashion down the glass fibres of the composite?

It is the board material itself that goes faulty. I do not know the exact mechanism, but clearly parts of it become conductive, causing absolute havoc with the bias. The output tubes often get red hot and are destroyed.

Unfortunately the first thing the owner tends to do is replace the valves. Which very soon get destroyed again.

(Expensive business when it eventually gets fixed properly - that would be

12 x EL34's, the output PCB, and a fair bit of labour).

Keep a careful eye on the bias as it warms up, adjusting all the time to keep it less than 90mV each side. The moment you run out of bias adjustment is the moment you have pretty much demonstrated the faulty PCB, and you want to turn it off immediately.

Gareth.

>

Couple of other tips on the DSL and TSL amps:

Resolder (and preferably reinforce with wire) the DIN footswitch socket. It is only held in place by the solder joints which often (always) fail in regular use.

The back panel PCB mounted mains and HT fuseholders also tend to suffer from dry joints. Whip off the PCB and resolder them. 3 minutes tops.

The 16 ohm speaker output is a switched socket that disables the 4 and 8 ohm sockets when used. These often get damaged, becoming a problem when the other sockets are used instead. Check, repair or replace.

The speaker ohmage selector switch also tends to suffer from dry joints.

(Either of the above 2 problems can destroy the output TX, output valves and other stuff, so best do it)

The output mute switch is stupidly relying on the switch contact to feed the signal to the power amp, rather than being a mute. Clean it.

Gareth.

It would be nice to find a schematic for this specific variant , but only one Google Ref for the number on the end face label JCM2-20-06

I'm wondering if there is any significance in the bodged 1/3 W R replacements R70 of 5K6 R7 of 5K6 R48 of 100K

unfortunately none of these numbers agree with any JCM2000/ DSL schemas that I have. Similarly I cannot find the 2K2 HT line dropper for (a) half of V1 mentioned on these schema

This site mentions people finding 220k grid resistors (R70, R7 etc) instead of the 5k6 resistors they expected, along with other details with regards to this board problem.

Looks like someone may have tried to sort this problem out before, but didn't replace the PCB?

Gareth.

New PCB's ARE available from Marshall, at about £60 to the customer if I buy and fit one. Marshall tell me they sell a lot of them. I understand costs are much higher outside the UK, however.

I totally agree re: the bodges and drilling malarky - the problem IS the PCB material.

Charging the customer a whole bunch of labour to bodge a way around this problem is, IMHO, unethical, as he STILL has a faulty PCB, and has been charged loads of money not to fix the real problem.

Replacing the PCB is relatively cheap (compared with the alternative labour costs) quick and easy, and is the only logical solution here.

Incidentally, 2 old faulty boards I have here and have kept, both have 220k grid stopper resistors instead of the 5k6 the schematic shows.

Cheers,

Gareth.

*guffaw*

Gotta remember that one!

"Gareth Magennis" "Phil Allison"

** It's not just the cost in dollars or whatever - availability is crucial too.

Even common Marshall power and output transformers ( the shitty Indian made one ones that sit the lams right on the chassis) are regularly out of stock here in Australia for months on end.

New output stage PCBs are not even on the radar !!!!!!!!!

** See above.

The logical solution is the use a bloody FAN !!!!!!!!

Takes less than 1 hour to fit one to any combo model and the job is DONE !!

An IEC inlet mounted on the lower back panel supplies AC power to the fan and it is up to the owner to use an extra IEC lead to run it.

.... Phil

OK, I see that in your situation that makes a lot more sense than in mine.

Any thoughts on the 220k resistor anomaly? Someone stuffed the component insert machine with the wrong resistors? Bit of a disaster, this PCB.

Gareth.

I should be able to get back to it in a couple of hours. Only just got around to building a Dexion support frame so I can toss these sorts of amps around with gay abandon, and not risk the bottles.

The valve set that are in here are all 2009 , he took out the new ones and replaced these used ones. They look brand new except for simple dragon's teeth marks, ie not repeatedly moved around. The Russian markings are quite visible and there is no sign of overheating of the bases , still light brown , and the pcb around looks as new. But the first thing I will do is check for o/p bias drift, then heating the pcb with hot air and a 2Gohm megger and then remove V1 and see if hum returns then replace and monitor DCs around V1 on heating ,

No mention of that Marshall Forum page about hum but it could be these pcb problems around V1 the immediate problem

The "Hum" I have associated with the PCB problem is ripple due to the huge currents the output valves are taking from the PSU. One pair of valves inevitably draws way more current than the other pair, which again manifests as a hum. The more current the valves draw, the hummier it gets, until things start glowing red, then things start to break and the hum is gone.

Yours may not actually have this PCB problem, but the bodged in grid stoppers make me suspicious.

Gareth.

and

overheating

pcb

break

main bias results monitoring every 5 minutes for the first 15 minutes rate of increase is falling passing 10 to 15 minutes in one side rising 1.0mV per 5 minutes and the other side 1.5mV per 5 minutes starting from that 15 min reading of 76.9mV and 68.2mV should be about 90mV over an hour . locally heating with hot air (low setting) for 20 seconds the voltages shot up

5 or so mV but soon dropped back to where they were , unlike the problem around V1 it would seem.

now for the other preliminary tests

No E number of board supplier found but I would say, by trying to press a needle into the board, it is epoxy rather than polyester composite. Megger showed nothing untoward With no V1 in place , the grid socket pins measure about 10 or 20mV DC, wave hot air over the valve base and the readings shoot up to 100 to 200mV, just like applying a magic wand. The cathode lines have 2/3 orders of magnitude lower resistance to ground so any such effect not so obvious there

These are the small bottles not the hotter big bottles, something to do with that metal shield plate for them? I'm assuming the effect is at the valve base through board , hot pins rather than where the anode dropper leads pass through the pcb, and passing through rather than along the surface . I will explore this, as hard wiring the HTs to isolated valve bases ,only, is quite different to hard-wiring all HT traces. Now what is the physics /chemistry of all this ? I assume something hygroscopic is grabbing moisture then forming a conductive salt that stays within the micropores of the surface of the glass fibres, but why temperature increasing the conductivity , what salt has highly temp dependent conductivity, we're only talking 50 degree C or so

For those thru-board valve bases it looks as though you would have to replace with standoff types to retain structural integrity but with electrical isolation , as well as all that hard wiring and intermediary isolation/ mounting points. Looks like too much of a work up in comparison to paying Mr Marshall for a replacement board. I'm just wondering if this particular board , only 1 google ref, that they will not have a replacement