Marshall JMP MkII 50 - acting funny

[dai h.]

2) The 100K ohm pull up resistor coming off the resistor bridge at the phase inverter input, and on the other side of the presence pot, is measuring only 4K ohms. Seems very low for a 100K resistor. The only reason it could be that low is I'm measuring it in circuit, and there could be a very low impedance circuit load in parallel with the 100k resistor. It's labeled 47x or 47k with (100K) in parenthesis just under the 47x. Its is in fact a 100K resistor that's installed there.

one end of the feedback resistor (which is what I think you are referring to) is seeing a very low impedance connection to ground since it's connected to a speaker output tap (output transformer secondary output tap). The other side sees ground ground through a 5k pot or 4.7k resistor, so this (100k in parallel with 4.7k or so = about 4.5k) is probably what you are reading when you see "4k ohms".

3) I can't read DC resistance through the two .022uf caps feeding the phase inverter output from V3 into the power amp stage input (V4 & V5). This could be fine & normal, because they are caps.

on a good cap you shouldn't be able to get a DCR reading

5) A side thing I noticed is the secondary side of the output xmfr, which connects to the speaker jacks via the impedance selector read nearly a short to each other and also to ground. Does this sound normal for the speaker side of the xfmr? Seems like awfully low resistance.

> Orange to Violet = .4 ohms
> Orange to Grey - .8 ohms
> Orange to Green = .6 ohms
> Orange to Yellow = .4 ohms
> Orange to Ground = 0.0 ohms

looks normal. Output transformer secondary from 0 ohms (orange) to yellow (4 ohms), green (8 ohms), then grey (16 ohms) and the DCR is going up incrementally, so seems ok.

Could it be the bias caps causing the current to keep increasing & the voltage to keep dropping from 410VDC all the way down to 320VDC?

the more the current draw (from the power transformer) the greater the voltage drops. Sort of like if you grabbed a branch on a tree and it bent downward. The more weight you put on the branch the more it bends downward. Put too much and it breaks. The more current drawn, the greater the voltage drop. Draw too much current, and something gives out (fuse, copper winding (insulated wire) inside a transformer that is only capable of passing current up to a certain point).

What else could be causing the current to keep increasing, besides the bias caps ?

a bad power tube, possibly other things.

How do I soecifically do this to assess the 50+50uf cans? ---> "it will probably be cheaper to use a series resistor, apply a voltage to a cap, and measure the voltage to figure out the leakage"

(If you google you should find a better explanation, but basically) take the loads off of the B+ path (high voltage supply). That is, the tubes (since they are sort of like resistors across the power supply drawing a few milliamps for the preamp tubes and tens of milliamps for the power tubes (but more when there is a signal). Install a resistor, say a 100k (make sure the fuse is still in the path), power up, monitor the DCV across the 100k resistor. The voltage seen here should drop over time (a good cap should slowly repair itself with a current limited applied voltage). Ohm's Law: V/R=I, (Voltage/Resistance = Current) will tell you the current total. If the caps are good, the current (leakage current) should be small. (If you look at aluminum capacitor datasheets, they will give a formula such as "0.02CV" (0.02 * capacitance value * voltage) for the amount of leakage a cap should have under specific conditions. So if you see something that looks abnormally high, then that should be a hint that one or more caps have too much leakage (and should be replaced).

On the 9VDC battery test, do I just jump (+) & (-) of the battery directly to the 50+50uf cap can (+)n & (-) cap inputs? How long do I apply the battery to the caps? If the caps are good, should I just see 9VDC (or closer to that) across the caps? How long should the caps hold that 9VDC charge before they dissipate back down to nothing? I drained all the can caps using a resistor on added to the jumper clip.

(Assuming the cap is discharged and not holding a high voltage first.) you need to "see" (monitor) the voltage, so the leads from the DVM need to be on the cap while you test (alligator clip-equipped leads or retractable hook type leads are useful for "hands free" connections). For the 9V battery, you can use a battery clip with alligator clip lead extensions or touch the battery contacts directly to the cap, but you MUST get the polarity correct (or you can damage the capacitor--IIRC polarized alu electros only withstand 1-2V of reverse voltage) . Hook up the battery to the cap. You should see 9V(or so if the battery is fresh). Disconnect one side (red or black--remember that as the term "electronic circuits" implies, circuits are made up of loops (the electricity goes somewhere and it has to come back through a path). (The loop will be battery plus -> cap plus -> cap minus back to battery minus.) If after interrupting the path (disconnecting), the voltage drops very quickly (immediately), the cap is likely bad. If the cap is good, the voltage should stay for a while. How long exactly? I don't know but probably at least tens of minutes or (possibly way) more (the less leaky the cap is the longer it should hold the charge).

Anyway if you really want to service the amp yourself it would be a good idea to find some good sources (tons on the web) and try to understand some of the basics at least. Youtube looks quite good as well for basic theory on tubes and electronics. (Like these: )

Electronics At Work (1943) Westinghouse
Vacuum Tube Primer
Vacuum Tube Amplifier Theory

(Lots of books here: )

http://www.tubebooks.org/

some theory, amp calculators:

http://www.ampbooks.com/home/tutorials/lesson-001/

etc.

I don't know very much but I still find the bit that I have learned useful for making simple repairs, troubleshooting etc. so I would say it was worth the time I put into learning things.

[kdude12345]

Thanks again Dai. Great input and resource links!

1) Good, so the 100K resistor, (2) coupling caps, & OT are all good & measuring within norms.

2) "(me) Could it be the bias caps causing the current to keep increasing & the voltage to keep dropping from 410VDC all the way down to 320VDC?

(you) the more the current draw (from the power transformer) the greater the voltage drops. Sort of like if you grabbed a branch on a tree and it bent downward. The more weight you put on the branch the more it bends downward. Put too much and it breaks. The more current drawn, the greater the voltage drop. Draw too much current, and something gives out (fuse, copper winding (insulated wire) inside a transformer that is only capable of passing current up to a certain point).

(me) What else could be causing the current to keep increasing, besides the bias caps ?

(you) a bad power tube, possibly other things."

I have extra sets of used GE 6550 & 6550-A tubes I've been using for the tests. Now matter how many tubes I put in the socket closest to the PT, they are glow bright red, so its not a bad power tube. In fact, I can take that exact same tube & swap it with the one in the other socket that is not glowing bright red, and the problem does not follow the tube & remains with any tube I place in the suspect socket. And its always the same. So back to my original questions, if its not the tubes, and everything meters correctly in the signal chain between the phase inverter (V3) & the power amp section (inputs into V4, V5), except the bias caps, which read 15uf & 22uf, what could possibly be causing the extra current draw, making the plate voltage start at 410VDC, and then quickly & continuously drop to 320VDC, before I have to shut the amp down?

Q: What else could be causing the current to keep increasing, besides the bias caps ? One or more of the power filter cap cans? Some pins not making good contact on the one 6550 tube that keeps glowing red?

3) "[me] How do I specifically do this to assess the 50+50uf cans? ---> "it will probably be cheaper to use a series resistor, apply a voltage to a cap, and measure the voltage to figure out the leakage"

[you] (If you google you should find a better explanation, but basically) take the loads off of the B+ path (high voltage supply). That is, the tubes (since they are sort of like resistors across the power supply drawing a few milliamps for the preamp tubes and tens of milliamps for the power tubes (but more when there is a signal). Install a resistor, say a 100k (make sure the fuse is still in the path), power up, monitor the DCV across the 100k resistor. The voltage seen here should drop over time (a good cap should slowly repair itself with a current limited applied voltage). Ohm's Law: V/R=I, (Voltage/Resistance = Current) will tell you the current total. If the caps are good, the current (leakage current) should be small. (If you look at aluminum capacitor datasheets, they will give a formula such as "0.02CV" (0.02 * capacitance value * voltage) for the amount of leakage a cap should have under specific conditions. So if you see something that looks abnormally high, then that should be a hint that one or more caps have too much leakage (and should be replaced).

Q: Where exactly should the 100K resistor in installed for measuring V-Drop, to determine tube current, and whether the power filter caps cans are bad or good?

4) "[me] On the 9VDC battery test, do I just jump (+) & (-) of the battery directly to the 50+50uf cap can (+)n & (-) cap inputs? How long do I apply the battery to the caps? If the caps are good, should I just see 9VDC (or closer to that) across the caps? How long should the caps hold that 9VDC charge before they dissipate back down to nothing? I drained all the can caps using a resistor on added to the jumper clip.

[you] (Assuming the cap is discharged and not holding a high voltage first.) you need to "see" (monitor) the voltage, so the leads from the DVM need to be on the cap while you test (alligator clip-equipped leads or retractable hook type leads are useful for "hands free" connections). For the 9V battery, you can use a battery clip with alligator clip lead extensions or touch the battery contacts directly to the cap, but you MUST get the polarity correct (or you can damage the capacitor--IIRC polarized alu electros only withstand 1-2V of reverse voltage) . Hook up the battery to the cap. You should see 9V(or so if the battery is fresh). Disconnect one side (red or black--remember that as the term "electronic circuits" implies, circuits are made up of loops (the electricity goes somewhere and it has to come back through a path). (The loop will be battery plus -> cap plus -> cap minus back to battery minus.) If after interrupting the path (disconnecting), the voltage drops very quickly (immediately), the cap is likely bad. If the cap is good, the voltage should stay for a while. How long exactly? I don't know but probably at least tens of minutes or (possibly way) more (the less leaky the cap is the longer it should hold the charge).

I did a few tests on the cap cans. With quick disconnects on each, I apply amp power, then Standby is turned on, and ran them all up to around 400VDC. Turned off Standby, Removed amp power, then quickly disconnected the caps from the circuit to see how long & at what voltages they dissipated at after 10 minutes or so:

Preamp side can: ~380-400VDC w/amp on, then after quick disconnect was @ 140VDC on both sides
HT can: ~380-400VDC w/amp on, then after quick disconnect was @ 250VDC on both sides
mains can: ~380-400VDC w/amp on, then after quick disconnect was @ 80VDC on one side, 100VDC on the other

Same test, but with cap cans left in circuit and dissipation times measured at 1 minute instead of 10:

Preamp side can: One side ~370VDC w/amp on, ~300VDC after 1 minute; other side ~385VDC w/amp on, ~300VDC after 1 minute
HT can: One side ~385VDC w/amp on, ~300VDC after 1 minute; other side ~400VDC w/amp on, ~300VDC after 1 minute
mains can: ~420VDC w/amp on, ~310VDC after 1 minute (both legs jumper-ed together)

With the above results, is it safe to assume the mains can is bad, but the HT can is good, and the preamp can is questionable in the middle. Or would you deduce a different conclusion from these tests?

Here are some additional measurements I took during this process that might provide additional clues:

Measured at 6550 tube socket pins. Measurement 1: Voltage to chassis ground w/power & standby ON. Measurement 2: DC resistance to chassis ground, with amp turned off:

Pin 1: 0 ohms; 0VDC/VAC [ground jumper to Pin 8]
Pin 2: 0 ohms; 3.3VAC (6.9VAC to other side of heater) [Heater]
Pin 3: 5M ohms; 380-400VDC [Plate]
Pin 4: 1.6M ohms; 385VDC [Grid 2 - Screen, 1K resistor to Pin 6]
Pin 5: 209K ohms; -37.5VDC (input from V3/bias) [Grid 1]
Pin 6: 1.6M ohms; 385VDC [no tube connection, used for connection from other side of 1K resistor to Pin 4]
Pin 7: 0 ohms; 3.3VAC (6.9VAC to other side of heater) [Heater]
Pin 8: 0 ohms; 0VDC/VAC [ground - Cathode & Beam Plates]

NOTE: All tests done with tubes not installed & bias caps (10uf) removed. No excessive current or falling Plate voltage observed in this condition.

Does this mean the bias caps were the cause of the excessive current & falling plate voltage? Or does it mean the one cap can (and/or the others) that looks bad was causing it? Or could it be a combination of the both?


5) Anyway if you really want to service the amp yourself it would be a good idea to find some good sources (tons on the web) and try to understand some of the basics at least. Youtube looks quite good as well for basic theory on tubes and electronics.

Thank you for the reference material. Some good stuff in there. I should tell you though, I do have a background in electronics & electrical engineering, and am a techie gear head type guitar player, work/mod/repair my own gear (unless its really serious), and though I am not an analog, tube, or amp design engineer, I am fairly familiar with Ohms Law, basic circuit design, general concepts & purposes of circuits, high voltage caps, soldering, component types & purposes, etc. So while I'm not so familiar with these amp designs & some of the nuances, you can assume I know enough of the basics to be proficient at testing, metering, scoping, soldering, and fixing things. Once I have a decent idea of what the problems are, what they're being caused by, and what needs to be done to fix them, I am pretty capable of executing the solution. I just wanted you to know what level of expertise & knowledge you can communicate with me at. It should save you some time & effort when reviewing & responding to content on this thread, because there are some things you can assume, that you don't have to explain or provide too much detail on.

I appreciate very much your input & feedback on this thread.

[kdude12345]

A side note: i just checked 2 or the 3 cap cans after an hour from when they were disconnected from the live circuit:

Mains can: 40VDC; 60VDC; after an hour or so, after disconnected from live circuit
HT can: 190VDC; 160VDC; after an hour or so, after disconnected from live circuit
Preamp can*: 10VDC, 9VDC; after an hour or so, but still left in circuit, w/amp turned off

*NOTE: So the preamp can is not a good test to see how long it holds a charge, because to has the entire circuit hanging off it draining it fast then if it were isolated and left open, so its true internal dissipation rate could be observed.

All measurements & tests were done w/o tubes & bias caps installed.

[neikeel]

I appreciate that this is a learning exercise (hopefully we are all learning someting al the time

If you want to get the amp working I think you will find that new bias caps are the place to start and unless you have a supply of tubes (and possibly OTs) to throw at it in proving the point I would do that first, whether the original filter caps need replacing will be evident if you get hum or poor performance (low output lack of crispness in tone) will be evident later. A pair of BC 160v 10uF caps from Valvestorm is $5 so not a bank breaker.

Defective sockets usually cause arcing and leakage at high voltages ie when the amp is pushed (the highest voltages are on pins 3 , 4 and 6) sometimes you can see carbon trails on the socket sometimes you have to play the amp open in a dark room!

Please let us know how you get on

[kdude12345]

Thanks for the input Neikeel. Much appreciated.

I actually order replacement bias caps & one for the open cap on V1 a few days ago. Parts should arrive today or tomorrow. I'll be able to replace the caps & check out the amp shorty. At that point I can assess the amp & sound like you mentioned.

I have 6-8 original GE 6550 & 6550A tubes to use for this amp, a few that can be used to troubleshoot the amp, and if it comes down to it, be sacrificed in the name of progress. Though you should know that all the tests with the actual tubes installed were limited to 1-2 minutes tops with standby ON. I've been monitoring the plate voltage each time I power up the amp to run tests. When standby in turned ON & the current starts to run away on the one output tube near the PT, causing plate voltage to fall down at the same, from around 400VDC down to 300-320VDC, and it doesn't matter what tube I put in there, the result is the same, I power down the amp immediately. This always happens within 1-2 min of turning the standby ON.

The one answer that nobody has seemed to be able to answer, that I've asked repeatedly, is what could be causing the run away current/dropping plate voltage scenario, assuming its not the tube itself?

1) Could this be solely attributed to the bias caps being bad?
2) Could this be solely attributed to the power filter caps cans being bad?
3) Could it be a combo of the two?
4) Is there another possible root cause different from the above that could be the culprit?

[neikeel]

I would suspect the coupling cap on the PI output feed for that socket being leaky, that is if it follows the socket rather than follows the tube...........

[kdude12345]

Thanks Niekeel.

The symptom of the tube 'glowing bright red' (a lot more than the other tube) does follow the socket, not the tube, and it doesn't matter how many or which tubes I put in, it was always the same. But keep in mind, that was with the tubes & old bias caps installed.

However at the moment, and with the tubes & bias caps not installed, both .022uf coupling caps on the PI output, that feed the 6550s input (pin 5), read nearly perfect on my DVM w/capacitance meter, as far as I can tell:

Cap coupling PI output to pin 5 on V4 input (6550 tube); DVM measurements:

.02uf across cap legs
1.5M ohms across cap legs
209K ohms to chassis ground on pin 5 side
5M ohms to chassis ground on PI output side (w/power filter caps drained)

Cap coupling PI output to pin 5 on V5 input (6550 tube); DVM measurements:

.02uf across cap legs
1.5M ohms across cap legs
209K ohms to chassis ground on pin 5 side
5M ohms to chassis ground on PI output side (w/power filter caps drained)

So based on these results, it looks like the coupling caps between the PI & power amp stage are fine.

Do you see anything I could have missed, that would indicate there is an issue with coupling caps?

Or is there are a different part of the circuit that could make the excessive current come from the PI side of things, other than the bias caps?

[neikeel]

If you have a scope you could examine for DC leakage......................

ideally if you have a scope and signal generator you can test for DC contamination on the output grids without tubes in and the amp powered up.

Static testing will only test for a short or open circuit but leakage of dc across the cap will be a possibilty. Of course Dai's point about tracking across from pin 4 or 6 (high voltage dc of 400v plus) to your grid on pin 5 will do it too

I see you get -37v at idle on pin 5.

Sometimes you can get oscillation issues and a 5k6 or 4k7 grid stopper on the pin of the socket between the wire and the pin can help with this but it is often a bit of a sticking plaster over other issues such as poor lead dress.

[dai h.]

Of course Dai's point about tracking across from pin 4 or 6 (high voltage dc of 400v plus) to your grid on pin 5 will do it too

I think that is someone else but regardless still applies as one possibility (i.e. something is making the bias go positive, that is in the direction such as -30V from -40V (less negative bias voltage means greater current draw--if you think of the power tube as a resistor across the power supply, less negative bias voltage makes it a lesser value resistor so draws more current from the B+ (high voltage) winding. Could look at the socket for carbonization (will make for an unwanted leakage current path acting like a resistor).

re: the bias caps, I would think the red plating would happen to both(?--at least can't think of a reason at the moment)

[kdude12345]

UPDATE: Received new caps. Replaced bias caps & V1 Electrolytic cap.

Caps (amp powered off)

V1 = 335uf
Bias = 11uf
Bias = 12uf

6550 socket (amp powered on, no tubes)

heater pins = 6.9VAC
Pin 3 = 425VDC
Pin 5 = -42.5VDC

Both PI coupling caps (amp powered on, no tubes)

PI side = 400VDC
Pin 5 side = -42.5VDC

6550 socket (amp powered on, with tubes)

heater pins = 6.7VAC
Pin 3 = 395VDC
Pin 5 = -43VDC

Both PI coupling caps (amp powered on, no tubes)

PI side = 385VDC
Pin 5 side = -43VDC

No run away current draw or falling plate voltage. So some forward progress! Though I did not leave the amp on for very long. Only a few minutes. Everything seemed stable though. Have not plugged in a guitar yet, or ran it through speakers.

Something that caught my attention though. The plate voltage on V3 PI seemed too high. It was nearly the same as the output tubes, around the 400VDC. I had thought plate voltage for V3 should be almost half of that in the 200VDC-275VDC range, not 400VDC. And all the other pins on V3 were 0VDC, except for the two pins with high voltage on them. Something doesn't seem right with the PI...?

[neikeel]

UPDATE: Received new caps. Replaced bias caps & V1 Electrolytic cap.

No run away current draw or falling plate voltage. So some forward progress! Though I did not leave the amp on for very long. Only a few minutes. Everything seemed stable though. Have not plugged in a guitar yet, or ran it through speakers.

Good

Now how does it sound?

I would expect the plate voltages of V3 to be about half that on the output plates too, presume the output tubes were installed. As you have load and droppers on the system the voltages do drop as you go across the board, is something not connected properly, like the grid wires on V3??

[kdude12345]

Thanks Dai & Neil for all your input & insight. I was able to get the amp up & running sounding great. Basically, it appears the bias caps & power filter caps were the main culprit. Just to make sure, I did an entire re-cap job & replaced all electrolytic, bias, & power filter cap cans and that seems to solves the issues and replacing the power filter cap cans really made the amp come alive, and the old 1980s GE USA 6550A made for Groove Tubes i have been using really woke up & have razor sharp attack transients, and showed no signs of sagging or compressing. They are real strong & make the amp loud & a lot like a 100w Marshall.

For sure the power filter caps had gone south. I think one of the bias caps was bad also. Along the way for an intermittent connection in the PI section, that caused the high plate voltage in the PI section. I never did find what was causing one of the output tubes plate voltage to run-away. After replacing all the electrolytic caps & power filters, and troubleshooting the amp, somewhere in the process, the issue went away. The output tubes & sockets are still the same, nothing changed, yet the problem is gone.

Anyway, the amp works & sounds great. Thanks for your insights & input as I worked thru the issues.

[dai h.]

Good to hear. Glad you got it sorted.

re: the GE 6550 (I doubt they were especially made for GT. I think they just selected them to their standards), my understanding is that they are a really well made American tube (seem to be robust mechanically compared to say, the Mullard EL34, be long lasting, and have a high power output). That is not to say they are invincible and will sound good or have the highest output no matter what (the circuit they are in matters), but overall as a tube they seem to be excellently manufactured (esp. compared to the ones now).

[kdude12345]

Yes. They are simply USA made GE 6550A tubes selected, matched, & private labelled for Groove Tubes in the 80s. I have several sets. Though my main point was they were the tubes I was using the whole time with this amp, so it proved the problem of glowing red, run away plate voltage & current, and squishy crappy sounding, was not the tubes and was something with the amp itself, which is what I suspected all along, because I knew the tubes were good and were high quality made. In this case, it turned out to be electrolytic capacitors & the power filter cans. There still is some mystery in my mind of what was causing the glowing red in one specific tube location, and why that location had run away plate voltage & current. It's seems like the cap issues weren't serious enough to cause that drastic of a reaction. But the problem is solved now, so it doesn't matter any more.