JCM800 50W power tube melt and fuse blow

[Neptune]

Hi guys

here's Marco with a technical question on power tubes, fuses and all that stuff.

Main character: Marshall JCM 800 50W combo (4104); I got it used in 1992, never replaced any component and, as far as I can tell after having checked it several times, it is completely stock.

Short version (for the absolute experts who need no details to grasp the nature of problems and the lazy lads who don't like to read):

- the stock T 500A fuse blew and one of the original Siemens EL34 power tubes had its glass enclosure molten and darkened in the areas above pin 2 and 3 and (to a lesser extent) in the area above pin 6 (which in EL34s is not present);
- new T 500A fuse and new pair of TAD JJs put in (plate current bias probe in series with the JJ tube placed in the socket where the previous Siemens tube failed): no reading from the probe, new fuse blown, new JJ tube with darkened glass enclosure in the area above pin 6, pin 7 and pin 8.
Need diagnosis and instructions to solve the cause of the problem.



Detailed version (for the meticulous, the curious, and those who want to spot if I made some stupid mistake and answer the post just to tell me that I made a mistake):

Weeks ago I switched the amp on and immediately noticed that it didn't sound as usual: lower volume, intermittent sound output...a bit like a car engine when half of the cylinders don't work, if the analogy sounds acceptable.
I went with the common tests to evaluate the problem (switch on/switch off, standby on/standby off, change input, roll on/roll off potentiometers...you name it) but the problem remained.
I don't think all of this took more than five minutes when I suddenly smelled a vague burning odour and I immediately switched off the amp. I let it cool down and checked it, here's the results:

• T 500A fuse blown;
• one of the power tubes (Siemens EL34) had the glass enclosure molten in two spots (the area above pin 2 caved in, had a dark tint and at the bottom there was a hole in the glass; the area above pin 6, which is not present in EL34s, had a milder version of the same appearance, that is to say it was slightly caved in with a dark tint)

My father, who's been in the hifi repair business for decades, suggested that the problem could be due to a failure of the tube itself (after all, they were there for 35 years) that led to the blowing of the fuse.
So I ordered a matched pair of TAD JJs (and, while I was at it, also 1 Gold Lion and 2 JJ 12AX7 premp tubes), a plate current bias probe and a cathode current bias probe (to be used with a multimeter).

Here's how it went:

• re-checked the original preamp tubes and all of them seemed fine so I left them in place to avoid introducing another variable to the problem;
• replaced the T 500A fuse
• replaced the power tubes with the new JJs (put the plate current bias probe in the socket in which the original Siemens tube that failed was connected)
• connected the speaker cable;
• set all pots to 0;
• connected the power cable, switched the amp on and let it warm up while in standby mode (both tubes lit up);
• switched the multimeter on, set it to mA and put the red plug in the <400mA input (as reported in the instructions);
• flipped the standby switch to get the party started but the multimeter doesn't move from 0.0/0.1, while the cones make little random, low volume popping noises.
  I guessed the multimeter did not give any reading because plate current may be above 400 mA so I put the amp in standby, placed the red plug of the bias probe in the >400 mA input of the multimeter and lit the amp up again: same as before, multimeter reads 0.0/0.1 and the same noises can be heard from the cones; after less than a minute, a feeble howling sound (like low-pitched feedback) could be heard which started to ramp up in volume so after a couple of seconds I switched the amp off to prevent it to reach considerable volumes.

I let the amp cool down and went to check it.

Results:

• new T 500A fuse blown;
• new JJ EL34 power tube (the one with the plate current bias probe in series, placed in the same socket of the Siemens tube that previously failed) has the glass enclosure darkened in the area above pin 6 (not present), pin 7 and pin 8, the JJ logo partially involved and faded as a result.
• inside the amp, everything seems to be "visually fine", that is to say there are no evident signs of burning, bloating, leaking etc.

It is now clear to me that it was not a failing tube that caused the initial problem (and the fuse to blow) but, instead, that there is a problem somewhere else causing the tubes connected to that particular socket to dramatically fail. I am no expert so I may be saying the dumbest thing in the world, but I like to try (and I also like the idea to give the true expert among you a laugh): I would assume that the socket receives excessive current, and that could be because of some resistor not working properly in the path leading to the socket; as a result of excessive current supply, the tube "redplates" and overheats rapidly, leading to the darkening of the glass enclosure (or its melting, if the amp is switched off too late) and to the failure of the tube itself, which in turn causes the fuse to blow. I would rule out an OT issue as a cause of the problem because the power tubes connected to the other socket (the old Siemens tube and the new JJ tube) don't show any sign of darkening/burning and seem perfectly fine; for the same reason I would also rule out the failure of one of the preamp tubes as a cause, as this would affect both power tubes (besides, I checked the preamp tubes several times and they look perfectly fine).

The question is simple (and, fortunately, short): any expert advice to precisely diagnose the problem and an explanation on how to eliminate the cause of the problem itself?

Thank you so much for your help (and, most of all, for your patience, if you read it all)!!!

 

[Pete Farrington]

What’s the conditions of the output valve sockets, ie is the insulating body material discoloured, especially in the area where the valve bases were discoloured?
Check from both sides of the chassis, inside and out.
Replace the T500mA HT fuse and power up without any valves fitted; what’s the VDC at pins 3,4&5 of each output valve socket?

 

[Gunner64]

Until the real experts chime in, and ONLY if you have the skillset to do it safely, I would pull the power tubes, put in a new fuse and power it up and check voltages at the sockets..and that problem socket itself may have been arcing, or otherwise compromised...loose pins ect. With no bias voltage it will redplate and blow the ht fuse. Pin 5 you should have -40ish volts measuring + on pin 5 of the socket and - to chassis ground. Again ONLY if you can work safely in a hot chassis.

Very reluctant to instruct someone with unknown skills to work inside an amplifier. The last guy shocked the shit out of himself..

 

[Gunner64]

What’s the conditions of the output valve sockets, ie is the insulating body material discoloured, especially in the area where the valve bases were discoloured?
Check from both sides of the chassis, inside and out.
Replace the T500mA HT fuse and power up without any valves fitted; what’s the VDC at pins 3,4&5 of each output valve socket?

Ha ha..I was typing at the same time you were Pete...listen to Pete, Mr. op..he's one of the experts.

 

[Ken Underwood]

I dont think the question would have been asked if they were technical like Pete.

I get worried, maybe you have noticed, asking just anyone to do technical things on tube amps.

It seems as now many know how to bias output stages then they are tempted to go further, thinking thats easy i am now my own Tech, BANG OMG exit stage left the smoke is getting to me.

 

[nickfox]

 

[Chris-in-LA]

@Neptune I’m hoping that your father is available to assist with some of this. As you probably know, there’s high voltage that can kill you inside lurking around those tube sockets.

 

[neikeel]

From the brief description you may have had flash over between pin 2 (heater) and pin 3 (anode) which has damaged the output valve socket, but without inspecting directly it is hard to say, once it is scorched it is irretrievable and will need replacing.
Until you have verified that all of the other aspects of the output section are intact there is no point putting any more valves in.
Things that need checking are the integrity of the 1k screen resistor and, before installing any output valves that each output pin has the correct voltage (ie heaters 3.15 vac each, pin 3 = 450vdc, pin 4 slightly lower = 445vdc and similarly pin 6. Pin 5 should be somewhere in the -25v to -35vdc depending on the bias pot setting.

Check the above
Inspect the socket for soot, scorching.
If the screen resistor is 1k and the voltages without load or valve in are correct then it is likely the socket is compromised and you will need someone expert/trained/competent to look at it and replace.

 

[Neptune]

Hi guys, and thanks to all of you for the quick replies and for your concerns about my physical integrity

I uploaded some pictures of the tubes and socket in my original post, but it seems that for some reasons they were not included when it was made available on the forum: I'll try to repost them below this comment and see how it goes. Anyways, to answer Pete's question on the subject, both sockets seem fine, with no discoloration or burn/shade marks whatsoever, neither on the upper nor on the lower side; the fit of the tubes in the "problem socket" is tight just like in the unaffected socket. Moreover, before installing the new tubes I had sprayed some contact spray on the pins of the old tube and plugged it in and out of both sockets a bunch of times to have cleaner contacts.

Anyways, this evening I'll follow Pete's suggestions on measuring the VDC at pin 3,4 and 5: I'll measure the unaffected socket first, and then the other one and see what I get: I'll write down the readings and let you know all the values

 

[Pete Farrington]

Yikes, thanks for the photos!

I think the HT fuse is Marshall’s great technical contribution to valve guitar amp design, but that poor Siemens EL34 demonstrates why I’m doubtful a HT fuse with a time delayed action is particularly beneficial in preventing collateral damage when something bad happens.
eg a T500mA is expected to pass over 1A for 2 minutes before blowing, see https://dalmura.com.au/static/Valve amp fusing.pdf

The strange thing is, the position of the HT fuse in a 4104 should allow a fast acting fuse type to be used, no change required other than amend the fuse to F500mA.

https://www.drtube.nl/schematics/marshall/jcm800pw.gif

 

[Spanngitter]

That poor Siemens EL34….I feel sorry to see it molten like this.
At all it seems like this valve did run without Bias and run into core meltdown, also it’s successor. I have to echo Ken here, knowing how to Bias doesn’t mean you have all the knowledge to understand and resolve issues within a tube amp and the outcome are wasted valves…
For sure voltages need to be checked on both sockets without valves fitted and the resistors be checked if still in spec. If these checks do come up good then next thing is to measure Output Transformer primary winding resistance, both sides should come up with fairly identical value. Also, due to the vintage of the amp, the power supply capacitors and bias circuit capacitor might be EOL and require replacement, even when this unit seems to be still fitted with DALE and not with LCRs what actually puzzles me a bit…

 

[Neptune]

Hi guys, here I am with the measurements!

First of all, to answer Spanngitter concerns, I will specify that the melting of the glorious Siemens EL34 was not due to some moron who was incapable of biasing the power section and set the bias control to 11 thinking he was an expert in tube amps: as I wrote, I got the amp in 1992 and never changed a single component, and I also believe the Siemens tubes were the stock ones that came with the amp in 1986 when the first owner bought it. I just switched on the amp as always and after a couple of minutes of it sounding odd I switched it off and the tube was molten...that's it.

As far as the screen grid resistor (which is soldered on pin 3 and pin 5) is concerned, values are 0.990 KOhm for the one on the unaffected socket and 0.988 KOhm for the one on the affected socket (as you can see in the attached pictures).

The VDC measured on pin 3,4 and 5 (as per Pete's suggestion) are 514 (pin 3), - 42.5 (pin 4) and 513 (pin 5) for the unaffected socket; they are 514 (pin 3), - 42.6 (pin 4) and 514 (pin 5) for the affected socket (see attached table).

Again, I'm no expert but the values I measured seem quite good...so, what do we do next?

 

[Gunner64]

I think you're counting pins wrong. Pin 5 should be the -42 voltage, screen resistor should be between 4 and 6

 

[Neptune]

Hi Gunner, you're right, I stand corrected. I was misled by the placement of the pin numbers on the bottom side of the sockets (they are offset and I had a wrong impression)...I knew that screen resistors are placed between pin 4 and 6 but, nonetheless, I managed to get it wrong

Here are the right measurements: unaffected socket reads 517 (pin 3), 516 (pin 4) and -43 (pin 5), while the affected socket reads 518 (pin 3), 516 (pin 4) and -43 (pin 5).

 

[South Park]

Don’t forget the coupler caps if any DC gets by them tubes over heat

 

[Matthews Guitars]

A tube can develop a space charge short which means the tube is only in a shorted condition while at operating voltage. It can happen with absolutely nothing wrong with the amplifier's circuits or components.,

The older Chinese 6L6 tubes were notorious for developing this condition. I won't put those in any amp for love nor money because of it.

The shorted tube may not read a short when checked on a tube tester. In fact, it won't, if the fault is a true space charge short.

 

[Chris-in-LA]

Just so that we are clear, it looks like the voltages are reading at a normal level. But @Neptune has had an old tube and a new tube fail in the same socket.

I would suggest that he let the amp run for a while without power tubes and monitor p5 voltages in order to rule out a leaky PI coupling cap. i would also do another close up visual inspection of the suspect socket. Any other checks after that are above my pay grade.

 

[Spanngitter]

Thank you for the measurement values, however you only did half of the job as I am still pending the resistance values of the OT primary windings.
Ensure that your amp had been turned off, unplugged from the wall outlet and cooled down before taking any measurement, only this way it is ensured that the HV has been discharged as you need to have one test point at the CT of the OT which hooks up to B+....both sides should measure fairly identical in the range of 40 Ohm. If this is not the case then your OT might be shot.
Also I recommend to check as Chris-in-LA has mentioned if a coupling cap is leaking. This is not a very common problem on JCM800 but it can happen...

 

[Gunner64]

His probe shows no bias topside, but normal at the pin itself.

Have you cleaned the sockets? You have bias supply to the socket, but apparently, with none showing at your probe, it isn't getting any further. I would be certain it isn't the socket itself.

Unless your probe is shot, and the zero reading in the op is unrelated to your problem, I would highly suspect that socket. Of course I'm a moron who beats on stone all day, so what do I know..lol.

Experts??..Pete..what say thee?..

 

[Pete Farrington]

switched the multimeter on, set it to mA and put the red plug in the <400mA input (as reported in the instructions);

Please could you share a link to the probe’s webpage / instructions?
It’s just that most probes require the meter to be set for mV, rather than mA. ie a current sensing 1ohm resistor is fitted in the probe, and the meter is intended to read the mV across that resistor. 1mV corresponds to 1mA.
I’m only aware of the Eurotubes probe that reads current directly, anode current in that case.

Regarding a diagnosis, my guess is that your dad’s right, the old EL34 became gassy, resulting in excessive grid current counteracting the bias voltage. That process then feeds back on itself and causes the valve to spiral into redplating self destruction.

Have you or a friend got a light bulb limiter?
They’re extremely beneficial in these circumstances, as they’ll help to protect the amp from collateral damage (eg to transformers) until this is sorted.

new T 500A fuse blown

You have bias supply to the socket, but apparently, with none showing at your probe, it isn't getting any further. I would be certain it isn't the socket itself.

Unless your probe is shot, and the zero reading in the op is unrelated to your problem, I would highly suspect that socket

It’s certainly good practice routine maintenance to retension the socket contacts every few decades
But there may have been no reading from the probe meter because the HT fuse had already blown / because the meter was set incorrectly.
Bear in mind we’re reading bias voltage (eg-40V) at the grid pin5, and idle current (eg 35mA) at the cathode pin8 (typically mV across a 1ohm cathode current sensing resistor).