Marshall JCM 900 fried! Need help :)

[J M Fahey]

Yeah, I was wondering about that as well...

I wish all manufacturers would put the damned voltages down.
/minor annoyance rant off

Well, in that particular case actual voltage is much higher than apparently rated .

Literally, something rated "250VAC" should be rated for 1.4142 times that in DC voltage, or somewhat over 300VDC ... which no sensible company would do, if you read the datasheet small print, you´ll find the "400V DC" rated caps are "100% tested with 600V DC " and so on.

BUT this is not a standard "in circuit" cap like all we use, but a safety product rated to meet or exceed official Safety rules covering 250VAC lines or apparatus (as they call them) connected to them.

And although 250VAC has the peak value I mentioned, fact is that they can often and easily have peaks much higher (rays hitting a power line anywhere along the grid) so AFAIK 120V rated stuff must at least stand "1500V DC for 1 second" and 240V one "3000VDC for at least a second" .

Quoting from memory but basically that.

So I guess an X or Y rated "250V" cap must probably be a 2000VDC rated one, "tested at 3000VDC" .

So, inside a tube amp connected to some 350VAC it must be well within ratings.

[sibo]

Would this be a better choice then? USB UT61D Multimeter - UNI-T | Clas Ohlson Says Cat4 600V, Cat3 1000V so it should work up to 1kV then.

[sibo]

Update:

I now changed R27 and R30 (R30 measured faulty even with one and two ends disconnected).
Fired her up and started playing. After like 30 seconds there's a vague but increasing sound like the one when you tap a tele jack that is connected to your amp with your finger. The overall hum starts to increase and something smells burnt so I turned it off again. I tried opening it fast to locate the source of the heat but couldn't really manage. There was most heat coming from the two first tubes or perhaps the blue "cups" beside them, whatever is located inside those..

This all happened at the same time as I was fiddling with eq, presence and reverb but that might just be a coincidence.
Also I might recall it being louder at lower volume settings than it was now that I tested but I may be wrong as it was a while ago I used this amp..

[Audiotexan]

Update:

I now changed R27 and R30 (R30 measured faulty even with one and two ends disconnected).

Ok, that would have been a good place for us all to resume with you, however...

Fired her up and started playing.

I know it's hard to resist, but in order to be able to have consistency in troubleshooting, *we all have to be on the same page at the same time*. That way neither you nor us (anyone following in hopes of helping you out) get lost in what's going on, as we have to know the current status of the amp at each stage in relation to the progressive steps of troubleshooting.

If that status is constantly changing in between being able to chat with you here, there's no way for any of us to know 'where it's at now'.

So if you would, help us out on this, and please try and resist the urge to fire it up and "see if things are fixed" at each replacement of parts, unless requested to do so...
If we verify through proper steps, and see that a voltage is too high, or too low, before you plug the output tubes back in, hopefully we can save you from blowing that new set of tubes (if they weren't already damaged from before, which again, hopefully they were not!)

After like 30 seconds there's a vague but increasing sound like the one when you tap a tele jack that is connected to your amp with your finger.

While this is also good for us to know, we know it still has problems, but there are some more basic things we can do, and ideally, in a manner that doesn't put additional (or costlier) parts at risk. Which is always a side-goal as we go (to minimize the expense of the repair).

The overall hum starts to increase and something smells burnt so I turned it off again.

Ok, so we'll pick up here.

- Make sure power's off, and unplugged
- pull the output tubes (and leave them out)
- drain the caps
- crack her open and see if you can see anything that's obviously burnt.

Report back with your findings!

I tried opening it fast to locate the source of the heat but couldn't really manage. There was most heat coming from the two first tubes or perhaps the blue "cups" beside them, whatever is located inside those..

Also good to know! Since you say 'blue cups', I'm assuming that you mean the two nearest the mains power supply (PSU) caps. So those two tubes were /noticeably/ hotter than the other pair???

[sibo]

I know it's hard to resist, but in order to be able to have consistency in troubleshooting, *we all have to be on the same page at the same time*.

Totally with you on that. Will not start until we reach the point if it being safe.

- Make sure power's off, and unplugged
- pull the output tubes (and leave them out)
- drain the caps
- crack her open and see if you can see anything that's obviously burnt.

Nothing's obviously burnt.

Also good to know! Since you say 'blue cups', I'm assuming that you mean the two nearest the mains power supply (PSU) caps. So those two tubes were /noticeably/ hotter than the other pair???

Yeah it felt that way but the heat may be coming from elsewhere but at least that area seemed warmer to me. Not like-crazy-warmer, just warmer.

[Audiotexan]

Would this be a better choice then? USB UT61D Multimeter - UNI-T | Clas Ohlson Says Cat4 600V, Cat3 1000V so it should work up to 1kV then.

Ok, the important part of those specs are that it's DC rated for 1000v.
Looks like it should be fine! And as a bonus (compared to the first one) it's capacitance is rated to 4000uf, which will be much more useful than the 100uf limit! So IMO yes, good pick on that one for a starter!

[Audiotexan]

Totally with you on that. Will not start until we reach the point if it being safe.

Thank you!

Nothing's obviously burnt.

Ok.

Yeah it felt that way but the heat may be coming from elsewhere but at least that area seemed warmer to me. Not like-crazy-warmer, just warmer.

I know you mentioned before that you were near having to return the meter that was borrowed....do you still have it ATM? If not, let us know when you have recieved your new meter, and then we'll look at getting some voltage readings from there!

[sibo]

I know you mentioned before that you were near having to return the meter that was borrowed....do you still have it ATM? If not, let us know when you have recieved your new meter, and then we'll look at getting some voltage readings from there!

I still got it. How and what do you want me to read?

[Audiotexan]

Ok, given that you've mentioned that it develops a hum (and then a burning smell) the longer it's left on, I'm going to strongly recommend clear eye protection at this point. As given the amount of time that it will take getting basic readings of the mains/secondary voltages...it's going to have been longer than your 30 seconds reference point. If a drifting cap explodes while you're checking other things, it can potentially send 'shrapnel' everywhere. Not to mention scare the crap out of you, and risk jerking your leads/hands into somewhere you didn't intend them to be....so please, OBSERVE SAFETY BEFORE ALL ELSE! Think: one hand behind your back rule while taking readings in a live amp! If that means you have to alligator clip your ground lead in place so you hold the red lead and only use one hand, DO IT! We want you around to be able to enjoy life, regardless of if an amp gets fixed!!

I still got it. How and what do you want me to read?

Cool! Just wanted to take some basic measurements to start with...
Make sure that you keep the output tubes out of the circuit while you measure. Turn all control knobs to '0'. We do expect that some readings will be slightly high, as those tubes are not there to 'weigh things down' so to speak. This is normal.

Checking the power supply:
_______________________
(looks like it should be under (but near) 500VDC, as the mains caps are rated 50uf+50uf @ 500vdc.)
You should be able to take readings at the caps (the blue 'cups' you referred to earlier).

Mains voltage:
*Set meter to DC (highest voltage if there are selectable options)*

1 - locate the blue PSU cap that has blue/yellow/green wires.
1a. Put your black lead on the lug of the cap that has the green wire soldered to it, as that should be the grounded lug.
[ref:http://i.imgur.com/ISUg51Xh.jpg]
1b. Now put your red lead on the lug that has the blue wire soldered to it.
1c. Take reading
1d. Now put your red lead on the lug that has the yellow wire soldered to it.
1e. Take reading


2a. Repeat 1a. (Leave your meter 'ground' where it is)
2b. Now, on the other adjacent blue cap, put your red lead on the lug that has the /red dot/ inked on it (looks like it should be a brown wire soldered to it that feeds the center tap of the OT -but I can't confirm this from the pic. I can only assume it from the schematic.)
*note* both parts of this cap are tied together with a jumper, effectively making it a 100uf/500v cap.
2c. Take reading



Heater voltages:
Set meter to AC voltage 50-100v setting should be fine.

[ref: the long twisted pair of red and black wires that connect all of the tubes (both preamps and outputs) http://i.imgur.com/KNWJJjVh.jpg]
*note* The filaments (heaters) are going to be pins 2+7 on the EL34's, and on the 12ax7's they're pins 4+9, and pins 5+9. Pin 9 is common to both halves of the ax7, and that's why you see the preamp tubes tied together at pins 4+5 by the heater wire.*

Hopefully you've got a rock solid 6.3vac here.



3. There's also a 20vac line that feeds the preamp board. (There should be a black and white wire from the primary transformer, and you need to follow those leads. Check again that your meter is on AC mode, and then put your black lead on the Black wire, and your red lead on the White wire)

Again, this 20vac should not fluctuate.


4. Bias supply. We're hoping for -38 to -40v (only as a reference)
I'm almost expecting that this line will have some drift in it, so take a bit longer to measure it (again, as you say, it takes up to 30 seconds for it to hum).
Watch closely and see what it does when taking this reading...if it drifts or bounces around (doesn't stay a steady voltage within a volt or two) then that's a sign of the bias circuit still having problems. This is where I'm thinking your comment of "After like 30 seconds there's a vague but increasing sound...." and "The overall hum starts to increase and something smells burnt..." comes in. However, by the time you've taken your other readings, it may have already well drifted up into enough to give us a bad reading flat out. Either that, or something that has been trying to smoke WILL while you're trying to get your readings. At least the output tubes won't suffer the wrath since they're not in there!

set meter to DC voltage (if selectable: around 100v should be fine, we're looking for a NEGative voltage here anyways.)

4a. black meter lead to chassis ground
4b. red meter lead to *the common tie point between* R26 and R27. From the looks of one of the pics, that common point should be the side of those two resistors that are *closest to* the 'D' in the small sticker that says "Dupe 4102". To verify that those resistors are tied together on that end, you should get the same reading on either of those resistors AT THAT END.
4c. take reading.


I could be wrong, but, that's my hunch offhand.
It could also quite easily be a signal/blocking cap that's gotten weak, and is allowing dc into where there should only be ac (signal) passing or vice-versa. Hence the hum. But since you don't have a scope (or at least that I recall offhand) it's going to be a bit harder to tell/see that on a DMM then it would be on a scope.

Again, If you're not comfortable for ANY reason, or have ANY doubts about your ability to safely do any of the above, please DON'T. Just take it to a pro.
I'd rather risk hurting your feelings, or have you mad at me, but this is a whole different level of safety precautions than simply changing components in circuit that's been properly drained. These will be LIVE voltages that can kill you if not given the proper respect they're due. I can't emphasize that strong enough. Know your limits, and respect them. I have FAR more respect for those who admit when something's too much than I do for those that barrel in without sufficient safety/knowledge.

Forgive the 'lecture' if you already know all of this. I'd rather say it than regret not having said it though. :x
-AT

[sibo]

Wow! Thanks for taking your time writing all that down for me <3
A lecture is always good when it comes to dangerous environments, I worked a couple of years as an electrician so I've seen what happens to those who disrespect directions - Einstein hairdos.

I did all those readings and they are as follows:
Mains voltage
1c: 368VDC
1e: 538VDC
2c: 538VDC
Heater voltage: 7.2-7.3VAC (measured every tube socket).
Preamp feed: 23.7VAC
Bias supply: -58.7VDC

[Audiotexan]

Thanks for recieving the warnings in the spirit they were intended! One can never be too cautious with things such as this -especially over the internet!

Wow! Thanks for taking your time writing all that down for me <3
A lecture is always good when it comes to dangerous environments, I worked a couple of years as an electrician so I've seen what happens to those who disrespect directions - Einstein hairdos.

And that's if they're LUCKY!

I did all those readings and they are as follows:
Mains voltage
1c: 368VDC
1e: 538VDC
2c: 538VDC
Heater voltage: 7.2-7.3VAC (measured every tube socket).
Preamp feed: 23.7VAC
Bias supply: -58.7VDC

Ok, round 1 is out of the way! I'm a bit surprised they're that high.. but, all in all, things appear ok (generally speaking) so far. Bias is a bit 'cold' but that's good. Although we're not done with that (bias section) just yet, I want to get a few more base readings just to be safe...

If you would, grab quick DC readings on all of the 2k2's (on both sides of R31/32/33/34) with the black lead on chassis ground. Expecting all of them to read 538v on the one side, and slightly less on the other.

Also, if you would, while keeping the black meter lead on chassis ground, take DC readings on each side of R24 and R25 (the 1k5's).

Next, take a DC reading of R8 and R9. You could also just take a reading at the fuses (FS1/FS2) too, but given the type of holder they use, you'd have to pull the board, so R8/R9 are easier to get to!
[ref:http://i.imgur.com/yq5wpjlh.jpg]

Lastly, if we can get a DC reading of the Red and White wires that come directly out of the Output Transformer (OT), again, keeping the black meter lead on chassis ground.

[Audiotexan]

Am gonna have to call it a 'night' here...as it's almost 8AM. lol
I'll pick back up after some sleep! (Unless someone else beats me to it! And then I'll watch with continued interest as usual!) *grins*

[sibo]

Yeah I get you, here it's 3 mid day

Here are the readings:
I will call the backplate of the amp (where you connect speaker) north just to locate which end of the resistors give what result.
All these DC readings are done with black lead on chassis ground..

R8 ”north”: 0V
R8 ”south”: 0V
R9 ”north”: 0V
R9 ”south”: 0V

R24 ”north”: -55VDC
R24 ”south”: -55VDC
R25 ”north”: -55VDC
R25 ”south”: -55VDC

R31 ”north”: 550VDC
R31 ”south”: 542VDC
R32 ”north”: 550VDC
R32 ”south”: 542VDC
R33 ”north”: 548VDC
R33 ”south”: 542VDC
R34 ”north”: 552VDC
R34 ”south”: 542VDC

OT red: 558VDC
OT white: 559VDC

[Audiotexan]

Just condensing readings to a single post real quick (for a single easy reference point)

Ok, that should yield the following:
---------------------------------
Heater supply [expected 6.3vac]: 7.2vac
Preamp mains feed [expected 20vac]: 23.7vac
Bias supply [expected -38 to -40vdc]: -58.7vdc

TP 1c: "HT to preamp" R30 (PSU cap1 side blue wire): 368vdc
TP 1e: "screen resistor feed" R35 (PSU cap1 side yellow wire): 538vdc
TP 2c: "center tap of OT" (PSU cap2 side brown wire): 538vdc

Unless I'm forgetting something, and assuming that you're on 240v mains in Sweden, is that there was a healthy 21v jump (as evidenced by OT readings from 538v to 559v) on the mains between the first test session and the second!!
This is also a factor that will make consistent readings a bit of a bear assuming you don't have access to a variac, so that you can set the incoming voltage to maintain a steady (240v supply) setting each time you take readings.

Cathode protection at R8 and R9 should both always be 0v, so that's good. If there's ever a voltage there, that's what blows the fuse to ground, and then whatever amount of voltage that might be, passes through R8/R9 and subsequently lights up the LED letting you know there's an output tube problem. (one LED for each pair)
Grid resistors R24 and R25 appear good as well.

Now, the next oddity comes in when you take readings on R31-R34. But, for the moment, I'm going to attribute that to line voltage (240v) fluctuating between you taking readings. Because if you measured in the order you posted the readings, it looks like there was a slight dip to 548 (on R33 -down from 550) and then it boosted to 552 (on R34) all the way up to 559 by the time you got to measuring the OT.

Screen resistors R31-R34 are a little odd, but should be good. If they all measure the same (resistance-wise) they should all drop the same voltage, but we had:

R31-R32 both drop 8vdc (while line volts stayed steady 550)
while
R33 drops 6v? (line drop to 548, dropped 2)
and
R34 drops 10v? (line boost to 552, gained 2)

So again, IMO that points to the source line voltage jumping around during the course of taking measurements. But it could also point to instability in the one half of "cap1" {-the yellow wire that feeds TP 1e: "screen resistor feed"}. Possibly more on that later...just making note of it for the time being.

Basically, what I'm trying to impart here is, it may well be important to get a reference on base readings when measuring voltages each time from here on, to note the "actual" changes, rather than making 'assumptions' that whatever the primary reading (regardless of what it may be) we're trying to measure should be under steady conditions (if that makes sense to you -it may read strange, but that's just me more than likely). Meaning: just for the basis of comparisons.

EG: If you know that the OT is reading 550v, and you measure -55dc at R24/R25, then the next time you measure 538v at the OT, and it reads -53.8 it should be safe to assume that nothing changed in the bias circuit, it's simply that the line voltage changed.

Plate voltage 1 (OT red): 558VDC
Plate voltage 2 (OT white): 559VDC

[Audiotexan]

Ok...so now that that's out of the way, I'd say time to check for ripple on a few points.

Set your meter to AC Volts. Start each check at the meter's highest voltage setting (just to be safe), and scale down as appropriate. There shouldn't be much ripple, but you never know until you get in there.

Refer to post 69 and measure:

Cap1
1c:
1e:

Cap2
2c:

Once complete with the above, turn the amp off, and then unplug, and drain all high voltage (HV) caps. Once that's done, prep to resume testing and move to checking the bias for ripple. Again, meter still on ACV. This is a good point to check for noise coming in. If there is, then we've got a couple of directions to check from there...but let's see what you find first!

I'd recommend taking continual readings ASAP after flipping standby on, so that you can watch and see if the hum (ripple) gets progressively worse at this location.

Bias section
4c (10 seconds in):
4c (20 seconds in):
4c (30 seconds in):
4c (40 seconds in):

or thereabouts.... Again, just trying to give an idea of what we're looking for here.