Marshall JCM 900 fried! Need help :)

[sibo]

Interesting reading!
I've now taken the readings as described;

Cap1
1c (green-blue): Very jumpy. At first between 0-2VAC but after like 10 seconds it jumped between 0-1kVAC (note kV), after that it got pretty stable at 837VAC. After turning the amp off and on it stayed at 0V.
1e (green-yellow): 0V

Cap2
2c (red dot): 0V

Bias
4c (10 sec): 0V
4c (20 sec): 0V
4c (30 sec): 0V
4c (40 sec): 0V

So 1c acting like that feels weird, at least in my book?

[Audiotexan]

Interesting reading!
I've now taken the readings as described;

Cap1
1c (green-blue): Very jumpy. At first between 0-2VAC but after like 10 seconds it jumped between 0-1kVAC (note kV), after that it got pretty stable at 837VAC. After turning the amp off and on it stayed at 0V.

837v AC???!!? =o
Whoa.

That's crazy high. 0-2vac would have been expected/reasonable for ripple on the PSU caps...but 837? Just wow. I'm amazed you didn't mention 'large explosion' or smoking...or pieces of parts flying everywhere... :x
If that's accurate, you got LUCKY nothing went up in major fashion (inluding the meter). Incidentally, you might wanna check that meter on a 9v battery (or whatever you have handy) real quick since the rest of the readings that followed were 0v. *crosses fingers it didn't take the brunt* For future reference, when checking for AC ripple on a DC supply, any time it starts going above 20-30v (50v at the extreme outside), shut it down/pull the leads. There's major issues (or it means you're working with MAJOR voltages, -way- above 500v). As a general rule, AC ripple measurements should usually be a very small percentage of the DC operating voltage you're working with.

Check and make sure that the mains fuse (rated 2a if @ 230v) didn't go up in smoke as that AC drifted up to 837v (after the "10 seconds" in....), as the bias section (that measured 0v) is fed from a totally different secondary winding, so that should still have had voltage. So, either the mains fuse went up (hopefully), or the meter did (which hopefully has internal fuse protection as well).

Go back one last time, and measure DC voltage (chassis ground) at the OT, and PSU caps real quick, and turn it off as fast as you can once you've got the readings. If there's still voltage there, then that's a good sign.

It's definitely sounding like "Cap1" (the one with the yellow wire) needs to be replaced...but, given one's been stressed, I prefer to replace them in pairs. However, others feel only replace what goes down. I prefer the preventative approach though. Especially in the PSU.

[g1]

I was also thinking it might not have had a good contact with the probe and was arcing a little, giving a high reading.

[Audiotexan]

Possible as well!
The reason I hung on the latter was the 'steady 0-2vac for 10 seconds then it went nuts' part.

[g1]

It's quite possible there is a problem. Much better to be safe than sorry.
Re-reading the part about turning off and on again, he may have the probes clipped in place.

[Audiotexan]

It's quite possible there is a problem. Much better to be safe than sorry.

Indeed!

I feel like I've gotten a bit (ok, very) long winded on this thread, but, I'd rather cover too much than not enough. But if my impression of what sibo does/doesn't know is accurate, I'm trying to err on the side of slow and easy, but most importantly SAFE.

Re-reading the part about turning off and on again, he may have the probes clipped in place.

That's my hope as well.
I was also thinking though, that if he had it turned away from him, and the switches were on the back side, he might not have noticed them not lighting up when he got that 0v reading...
Time (and measurements) will tell though!

Cheers for your help and input at any rate! =)

[sibo]

I really don't mind taking this slow as I already learned a great deal by doing this. Also there is no rush getting this amp fixed as it's not my main amp.
Oh and I don't use clips when measuring as I don't have any. I am sitting with the amps front towards me so I see the switches
Fuses are still working, in fact I tested all fuses (again) just to make sure.

Now for the results!
This time I also used my old Fluke T5-600 which is less exact(no decimals) but at least I know it really works as intended.
Green text is measured using my Fluke device, blue digits are using the instrument I used before just for comparison.

OT
Chassi to OT red: 490VDC 512VDC
Chassi to OT white: 487VDC 512VDC

Cap1
1c (chassi-blue): 337VDC once stabilized, started at about 480VDC 365VDC
1e (chassi-yellow 481VDC 514VDC

[Audiotexan]

I really don't mind taking this slow as I already learned a great deal by doing this. Also there is no rush getting this amp fixed as it's not my main amp.
Oh and I don't use clips when measuring as I don't have any. I am sitting with the amps front towards me so I see the switches
Fuses are still working, in fact I tested all fuses (again) just to make sure.

Glad to hear you're picking up on what we're doing as we go! =)
(It's important to how (and why, and when to check for each), but a nice side effect for those helping you, it will shorten things up in the future for folks, not having to go 'by the numbers'! We can simply say 'test for ripple at point x' and you'll know that you test across the cap in AC mode, or if we're looking for "voltage", we can say 'check voltage here for x' and you know to be in DC mode on the meter!)

Now for the results!
This time I also used my old Fluke T5-600 which is less exact(no decimals) but at least I know it really works as intended.

You say less 'exact', however, just judging by the numbers, it's actually more precise. Given the differential between the two readings, it appears as though there's approximately 8-10% variance on the 'cheaper' meter you're using (Though only time and familiarity with the meter will let you know for sure, as I'm only making an educated guess here). The difference isn't large enough to be that one's measuring RMS and the other one's not, but it is large enough that the looser tolerances of the cheaper meter still show.

I'm trusting that the Fluke is in fact actually more accurate, even though it may not go out to the 'decimal count' that the other does, the reading appears to be 'truer'. Which helps shed some light on the parameters of what we were looking at before in the other meters you were looking at.

Green text is measured using my Fluke device, blue digits are using the instrument I used before just for comparison.

OT
Chassi to OT red: 490VDC 512VDC
Chassi to OT white: 487VDC 512VDC

Cap1
1c (chassi-blue): 337VDC once stabilized, started at about 480VDC 365VDC
1e (chassi-yellow 481VDC 514VDC

Ok, so again (this time on DC) 1c was jumping around and had to 'stabilize'. It's probably a safe bet to say that it needs to be replaced, but if you want to check it one last time, put your meter back on ac and check it for ripple.

If it starts out at 0-2vac again and stays there, that's good. If it rises outside of that in a hurry...pull the meter leads and shut-her down ASAP. And then order the cap. (Two of them if you want/can afford to do preventative maintenance.)

From the schem's, a single cap should be:
50uf + 50uf @ 500v (that's DC rated btw)

Probably an LCR branded cap IIRC..but that's not as important as making sure the ratings and physical size of the replacement match to make it easy to install where the old one came from!

[sibo]

Again, big thanks for world class support.
I bought that multimeter and did some more testing and indeed the cap has to be changed.
I spent an hour trying to find in at my only local dealer (Welcome to elfa - Everything can be organized). I also had a look at mouser.com as they ship to Sweden but can't find the damn thing x(


Text on it says
LCR
RED 50uF 500VDC
YELLOW 50uF
PLAIN NEG
FAC/RD37D
U.K.
90-47

It measure 73,5mm height and 35,37mm diameter.

So there are two cap-circuits each at 50uF in there. I googled and found that LCR is no longer in business. Good replacement should be F&T but still nothing with those specs at 500VDC..

Is that type of cap called something in particular, other than electrolytic capacitor so I can find it a little easier?

[g1]

You want a dual 50uF 500V can, radial type. 47uF is basically the same and may be easier to find.
But here are 2 examples:
https://www.tubesandmore.com/products/C-EC50-50-500

Capacitors (LCR-5050)

Another option is to disconnect it but leave it where it is for looks, and install 2 regular 47uF 500V axial type caps inside the chassis.

[sibo]

You want a dual 50uF 500V can, radial type. 47uF is basically the same and may be easier to find.
But here are 2 examples:
https://www.tubesandmore.com/products/C-EC50-50-500

Capacitors (LCR-5050)

Another option is to disconnect it but leave it where it is for looks, and install 2 regular 47uF 500V axial type caps inside the chassis.

Thanks! Ordered from tubesandmore so now I'll have to wait for some shipping over the atlantic ocean

Get back when they're here!

[km6xz]

I am new to this thread and I can see that it is going around in circles with too many redundant measurements that is confusing the situation. The heating and increased hum suggests a bias problem as everyone noted about, the test would be simpler to just measure at the grid of each tube to assure that all the bias circuit is working instead of taking lots of individual measurements and trying to figure out if the individual components. If the voltage on the girds are right and stable,move on, everything in the divider, rectifier caps is OK. If it does drift down, pull all the power tubes and trace what is happening in the bias supply.

Resist the beginner trait of wanting to replace everything it COULD be, it does no good, only causes the cause of the problem to become uncertain, it lowers the value of the unit unless you solder better than the wave soldering machines. If you are not skilled with soldering, you ARE going to damage the board and traces pulling parts out. A lot of incompetent techs repair by shotgunning which greatly increases the cost to the owner and decreases the value of their unit.
If a part is proven bad, replace it but note why it went bad. Resistors seldom fail on their own. Replacing parts require verification that they are the correct value. The original was the correct value, as evidenced by the fact that ALL the parts were the right value or it would not have worked as it did for 20 years. If after verifying the diagnosis and the repair is complete, it is a judgement call as to whether you want to restore the unit by replacing parts just to increase its potential reliability into the future. Replacing anything, even with the correct value, introduces uncertainty. Is the trace or pad damaged, well soldered, soldered with no bridges or movement of the joint while cooling, air cavities are eliminated, and many other uncertainties are introduced. A pro tech hates working on gear that people have messed with and given up, and often refuse them because tech induced defects seldom fit a logical or normal part or circuit decay. A wire or part put back wrong will confuse the diagnosis and relevance of readings.
What is your line voltage at your home? It might be high to get those values so check the mains voltage selector to see if it is set to the closed to actual value.
You found one resistor that was open, I think, from reading the thread. Why, is there anything in the circuit in series with it that could be trying to pull more current than normal? Focus on the symptoms and not parts, one part will be the cause, but there can be a ripple effect where it caused enough current in other good circuits to become damaged. In any case, the first cause is the key to a successful repair. Any diagnosis much fit the evidence and be the expected result from abnormal readings found. The readings are symptoms and considering all the symptoms together will point to a single cause. If the diagnosis does not explain all the symptoms your diagnosis is wrong.
When measuring resistance in-circuit, the readings are usually different than the value on the component because other components are connected to the part also. So the resistance measured can be lower than expected but higher only when there is still voltage stored in the circuit you are testing.
This repair is one that would require very little time for a tech to figure out because they have fewer guesses in measurements because more is revealed from having the right test gear and being able to determine the actual values quickly, seconds in fact. With just a volt and ohm meter, and no known load, ac mains current metering, test signal level control, and waveform display are the basics that would show in 2 seconds what is not right, although does not directly pinpoint the cause, just the current operating parameters. Knowing precisely what the amp is doing to the signal and how much current it is drawing doing it would have eliminated most of the measurements you were asked to do over the last couple weeks.
Can you list what you know for sure has been found. One resistor was open? Anything else?

[MarkusBass]

Thanks! Ordered from tubesandmore so now I'll have to wait for some shipping over the atlantic ocean

I'm also new to this thread and I haven't read it (too many measurements).
I just wanted to mention that the capacitors that you are looking for are available either from TAD (Germany), JJ Electronics (Slovakia), or from many shops in Poland (for example here: Gold Cap 50uF+50uF 500V - TAD - Elektrolityczne - Kondensatory). The price is the same but the delivery fee and delivery time is lower.
If you have that many problems with fixing the amp, you could consider visiting an amp service. It would be fixed in 2-3 days.

Mark

[Audiotexan]

It might pay to read the whole thread, but, I'll recap to see if I can at least provide my perspective...as I think your initial assessment is a bit harsh. But we all see things differently so..

Sibo's kindly let it be known pretty much up front that he doesn't know much about electronics, but wanted to learn -using this amp- as his guinea pig basically (ref: post 55 below).

I am new to this thread and I can see that it is going around in circles with too many redundant measurements that is confusing the situation.

Only recently, and due to changes in readings between tests.
The first set of readings (post 69) was to try and get baseline DC voltages for where the amp was condition-wise. (Is the supply stable at critical points?) (Sibo replied at post 70)
And trying to gather whether or not, the 'hum' was coming from the bias -through the signal chain -or through the PSU.

The second set of readings (post 71) was to confirm outlying readings of reply 70. But it appeared line voltage changed with post 73.

So I did a recap in 74, to get it all down in one spot, and reassess. And also took the opportunity to explain some more about why and how these readings were being interpreted so that he could hopefully continue to learn as he went.

Post 75 was supposed to be to get AC ripple, but that was delayed due to a problem being isolated with a PSU cap. Which currently, appears to be the source of the hum (being injected after a timed duration of operation as that cap starts to go south).

And then once a part was suspected, it needed to be confirmed.
Which is where we are currently. Both AC ripple readings, and DC voltages are fluctuating on a main PSU cap.

The heating and increased hum suggests a bias problem as everyone noted about, the test would be simpler...

.. If the voltage on the girds are right and stable,move on, everything in the divider, rectifier caps is OK.

He doesn't know what grids are, a divider is, nor rectifier caps.

Post 5 (OP) "What I'd like to know is rather where I'm supposed to measure and what results I should be expecting?"

Post 7 (OP)
"I noticed there's voltage on that big badass capacitor next to the text "JCM52C". Can I just use a shielded wire to unload it? Should I unload anything else to work safely? Anything else that's common sense or good-to-know?"

Post 12 (OP) "How do I adjust BIAS?"
(We haven't gotten there yet...)

Post 46 (OP) "What's a voltage dropper?"
He learned. Thanks to g-one.

Post 48 (OP) "What does that mean and what exactly am I measuring?"
(He learned how to measure a diode, thanks to 52 Bill)

Post 63 (OP) "There was most heat coming from the two first tubes or perhaps the blue "cups" beside them" ("cups" referring to capacitors, which he learned.)

Resist the beginner trait of wanting to replace everything it COULD be, it does no good

Post 51 and 53 already noted: "Shotgunning parts at it will only cost you money and time." And another important point from g-one: "For starters, you won't learn anything that way"
and Post 52 went into more basics (solely for educational purposes).

A lot of incompetent techs repair by shotgunning which greatly increases the cost to the owner and decreases the value of their unit.

Not sure if this is directed (based on your impression of me due to lack of reading the whole thread)...or just simple statement of known truths. If however, it's based on my stance of replacing the caps in the preamp secondary supply (with the shrunk casings due to overheating), I'll simply content myself with knowing we hold two different outlooks on minor/cheap details.

A pro tech hates working on gear that people have messed with and given up, and often refuse them because tech induced defects seldom fit a logical or normal part or circuit decay.

True enough, and there's also a lot of techs out there that wouldn't take the time to help someone in need (by doing a cheap $25 repair for someone when you actually spent hours in a unit simply because they NEEDED the break), or couldn't be bothered with going into this much detail with someone ON A PAID repair, much less for free.

A wire or part put back wrong will confuse the diagnosis and relevance of readings.

Which is also part of why I've been taking it slow, and emphasizing caution, and double-checking every step.

You found one resistor that was open, I think, from reading the thread.

Read again.

Post 11 audiopete sees "Also R8 and R9 look scorched." (visually confirmed)

Post 36 R30 is burnt as well. (confirmed in post 63)

Post 44 R27 is found bad. (confirmed in post 63)

In any case, the first cause is the key to a successful repair.

Under normal circumstances, and with a single problem or fault, I agree.

When measuring resistance in-circuit, the readings are usually different than the value on the component because other components are connected to the part also.

Which is again, why when that was relevant, g-one had him remove one end of the lead to test it.
You missed a lot when you didn't take the time to read the whole thing.

This repair is one that would require very little time for a tech to figure out because....

Because they have the knowledge, training, and equipment to do it in short order.

Can you list what you know for sure has been found. One resistor was open? Anything else?

Post 55 Sibo: "...I do this because it's interesting. Otherwise I could've just turned it in at a tech shop and got it fixed!"

I was actually thankful when g-one poked his head back in. As I know myself well enough to know that I occasionally miss something (that's obvious to us) but there's no way Sibo would know to check me on it. And I don't mind a bit, hell I expect it. This is a NO BS zone. And again, why I dig this place. But I know that my mom having hypoglycemia, and father's side having a history of diabetes...is a large part of 'why I forget what I forget along the way' all too often. So I expect to be checked when I faff up. But this?

Sorry to everyone else if I got out of line anywhere in there, but I couldn't help but get the impression that Stan thinks I'm either leading him on a wild goose chase, or wasting his (Sibo)/your (member's reading) time.

[sibo]

Thanks for fresh input.
Winner don't quit, quitters don't win - Now that I started doing this I will at least try to complete it myself. Trust me, the thought of turning it in was appealing at first but since there's no rush in getting it fixed I'd rather learn while doing it, even though I'm very well aware that it probably won't be cheaper than turning it to a tech.
My socket voltage is 233VAC and as I live in a city there shouldn't be much ripple. No other electronics have had any problems. However this doesn't mean that this wasn't the problem as the amp was sold and the buyer lives outside the city so you might be on to something here..

To sum this thread up:

Problem(s):
The amp was started after about two years spent resting in the garage, sadly it was sold so I didn't see the problem myself but the buyer claims that it started with a buzz and after a while the first channel started distorting and at the same time he discovered a scent of something being burnt.

After changing resistors R27, R30, R31-R34, C13, C14 (power amp) and output tubes I fired here up (I now know I shouldn't have) and discovered a tapping sound that reminds of when you touch the tip on a instrument cable jacked in. I also changed C29, C30, C21 & C22 in pre amp before this sound.

Error search:
To start off easy there was a burn mark on R31 and R27 power section. Also R30 measured 38.5k when it should be 22k. Those were all changed.
Other than random resistor measurements that don't need mentioning I have so far measured the following:
PSU Caps

#1
Earth -> Red lug: 368VDC after some jumpy results, leading to the change of both these caps.
Earth -> Yellow lug: 538VDC
#2
Earth -> Red lug: 538VDC

I also did some AC measuring for ripple and found that #1 to red lug is jumping up to 1kV briefly. It ended up displaying 837VAC but after I drained caps and restarted amp that didn't happen again..

Heater voltages
Measured on every tube socket (#2+#7 on output sockets): 7.2-7.3VAC

Preamp feed
Measured between black and white wires: 23.7VAC

Bias supply
Measured between ground -> common point between R26 & R27: -58.7VDC

Other
Ground -> R8 & R9 both ends: 0V
Ground -> R24 & R26 both ends: -55VDC

The following results were from ground to following destination, north is backplate - south is front of amp.
R31 ”north”: 550VDC
R31 ”south”: 542VDC
R32 ”north”: 550VDC
R32 ”south”: 542VDC
R33 ”north”: 548VDC
R33 ”south”: 542VDC
R34 ”north”: 552VDC
R34 ”south”: 542VDC

Ground -> Output transformer red: 558VDC
Ground -> Output transformer white: 559VDC

That's where I'm at. Don't think I missed any part now.

I would once again very much like to thank all people involved in this and especially Audiotexan!