Marshall JVM 410H bias ramping up

[christarak]

Greetings. This is a Marshall JVM 410H with the bias set nicely to about 30-32mA per tube with the Bias voltage at around -40v. Within 5 minutes the bias starts to heat up on its own. The PCB is JVM410-60-00 Issue 3Marshall jvm410 power amp.pdf . The amp appears to be from 2007.

Eg after 5 mins on V4 Ic=39.1mA (Vb=-37.1 V)
After 8 mins Ic=43.4mA (Vb=-35.7v)

I ended up with both bias pots wound fully clockwise (Lowering bias current direction) but the current had ramped up to around 56 mA per tube with a corresponding increase in the bias voltage (ie less negative).

I removed the power valves and tried the bias voltage measurements again. No drift after 30 mins, constant at -40V. Would this clear C12 and C14 as being leaky?


I have read about this bias ramping upwards being a problem in the DSL and TSL amps with drastic holes being cut in the PCB around Grid (pin 5), but not with the JVM 410. Is there another fix that you know of? With a steady bias voltage with EL34's removed, does that test clear the DSL/TSL condition of the conductive PCB?

[Jazz P Bass]

Are both bias supplies acting up.
The post is not clear on that point?

When you are monitoring the negative bias voltage, put your meter on Volts ac to measure the ripple.
There is an initial filter capacitor in the circuit (C35) and two additional capacitors on each bias 'leg'. (C22/23)

If the ripple increases when the bias voltage goes less negative on both legs, then the initial bias filter capacitor is suspect.(C35)
At the worst, all three could be bad.
Using the attached schematic crop you can test each node.
JVM410-bias supply.pdf

If you have spare caps you could try clipping in a parallel cap (observe the correct polarity!) to see if Any change occurs for the better.
If you do see any improvement then that cap is bad.

[g1]

You didn't mention trying other tubes so I have to ask that first.

[christarak]

Are both bias supplies acting up.
The post is not clear on that point?

When you are monitoring the negative bias voltage, put your meter on Volts ac to measure the ripple.
There is an initial filer capacitor in the circuit (C35) and two additional capacitors on each bias 'leg'. (C22/23)

If the ripple increases when the bias voltage goes less negative on both legs, then the initial bias filter capacitor is suspect.(C35)
At the worst, all three could be bad.
Using the attached schematic crop you can test each node.
[ATTACH]47284[/ATTACH]

If you have spare caps you could try clipping in a parallel cap (observe the correct polarity!) to see if Any change occurs for the better.
If you do see any improvement then that cap is bad.

Thank you Jazz P. Yes, both bias supplies are acting up.

I'll get on with those tests and come back to you.

Cheers
Chris

[christarak]

Sorry G1. I forgot to mention that. One of the EL34's had died (I assume from the bias going too hot). The other 3 all tested OK so I have used a known good used EL34 in its place. Judging from the cathode current they are all pretty well matched except for the new one running a little hotter (8mA hotter). I have got a new matched quad of TungSol's but I didn't want to risk frying those.

( I am trying out the Orange VT-1000 that I've borrowed from a customer. SO far I'm very impressed. Its way more compact and I don't need to use a 110-240VAC transformer as I do with the B&K. The Orange read 9,9,9 on the exisitng EL34's and a 10 on my second hand valve. I'm not sure what the units are on the Orange, but they relate well against the B&K)

[christarak]

Thank you Jazz P. Yes, both bias supplies are acting up.

I'll get on with those tests and come back to you.

Cheers
Chris

HI Jazz P and G1

I put in a brand new set of JJ EL34's to eliminate that possibility of shot valves. Sadly the ramping had me reaching for the power switch after 8 minutes. For expediency I just took readings of each bias supply (V4 and V2). Here's what I saw:

After 1 minute:
V4: Ic=53mA VbDC= -35.2 VbAC = 8.4mV (Cathode current readings are per tube, not per pair in this 100w amp).
V2: Ic=36mA VbDC= -35.3 VbAC = 8.2mV (so much for matched quads!)

After 5 minutes
V4: Ic=65mA VbDC= -33.2 VbAC = 8.4mV
V2: Ic=53mA VbDC= -33 VbAC = 8.9mV

After 8 minutes:
V4: Ic=78mA VbDC= -29.9 (quickly shut down at this point)

What do you think? Its not lookng like the Bias circuit caps are the problem. I scoped the bias on pin 5 and it was pretty much what you'd expect from a single diode rectifier: a 50Hz sawtooth (Australia) with a peak-peak of around 19mV.

[g1]

Well that should at least rule out the tubes. Still can't get the fault to show with the power tubes removed? Try a hair dryer or heat gun around the power tube area with them removed while monitoring the pin5 voltages. (V4 position seems consistent with the fault?)

[christarak]

Well that should at least rule out the tubes. Still can't get the fault to show with the power tubes removed? Try a hair dryer or heat gun around the power tube area with them removed while monitoring the pin5 voltages. (V4 position seems consistent with the fault?)

Thanks G1 Good idea about the heat gun with tubes removed. I'll try that next. I like you're reasoning in that it seems to be heat (or current) related. The bias is solid as a rock with no tubes installed. I assume this means we've cleared any leakage from C12 and C14. If C7 and C8 were leaking to ground and making the bias voltage less negative (hotter bias), do you think that would happen with no power tubes installed, too?

Yes, V4 is typical of the fault on all tubes, so I just used that as a reference, along with V2 as the "informant" from the other side of the bias supply.

[christarak]

Well that should at least rule out the tubes. Still can't get the fault to show with the power tubes removed? Try a hair dryer or heat gun around the power tube area with them removed while monitoring the pin5 voltages. (V4 position seems consistent with the fault?)

With the power tubes removed, I tried applying heat to various parts of the PCB, starting at the components relating to the bias, then spreading outwards. There were minor shifts (+/- 0.3v) but nothing of concern. I pulled out the PI and reinstalled the 4 EL34's. I set the bias to a very cool 20mA across the 4 tubes and monitored V1 (to check the other side of the bias supply). Within 3 minutes the ramping started again:

After 1 minute:
V1: Ic=20mA Vb= -42.3VDC

After 3 minutes:
V1: Ic=25mA Vb= -40.9VDC

After 7 minutes:
V1: Ic=37.7mA Vb= -37.9VDC


Can I ask a question of my professional repairers? I have spent about 8 hours on this amp so far and its still not fixed. Assuming I eventually can repair it, how would you bill such a repair? What if I can't repair this amp? When do you throw in the towel? How do you not lose face with one of your main customers (a store)?

[Jazz P Bass]

I would 'lift' the PI coupling caps & try again.

[TimmyP1955]

Sorry G1. I forgot to mention that. One of the EL34's had died (I assume from the bias going too hot). The other 3 all tested OK so I have used a known good used EL34 in its place. Judging from the cathode current they are all pretty well matched except for the new one running a little hotter (8mA hotter). I have got a new matched quad of TungSol's but I didn't want to risk frying those.

( I am trying out the Orange VT-1000 that I've borrowed from a customer. SO far I'm very impressed. Its way more compact and I don't need to use a 110-240VAC transformer as I do with the B&K. The Orange read 9,9,9 on the exisitng EL34's and a 10 on my second hand valve. I'm not sure what the units are on the Orange, but they relate well against the B&K)

Those Orange numbers tell you very little. I've had matched numbers - and even a number apart - that were within a few ma or less. But I've also had matched numbers that were 10ma apart. 95% of the time it's a good go/no go tester, but that's it.

[christarak]

Thanks for that feedback Timmy. What do you use as a valve tester? My old B&K still only gives me a % figure with no actual Gm value. Buts always been a reliable indicator of valve health. I calibrate it annually. The only downside is its size and that I have to use a step-down trannie.

[christarak]

I would 'lift' the PI coupling caps & try again.

Between me taking way too many hours to not solve the ramping bias problem, and the customer keen to get the amp back into his rental stock, we agreed to buy a replacement board from Marshall. Sadly, this decision based on pure economics, has limited my contribution to the Forum's pool of knowledge. As usual my sincere thanks to you Jazz P, G1 for your invaluable advice and suggestions.