I have a Marshall JTM60 guitar amplifier that I am having problems with and am looking for some suggestions. The schematic for this amp can be found at http://www.marshallschematics.com (there are 3 pages to the schematic).
The problem is this: there is a knob on the back of the amplifier that is called presence(VR201 on mains schematic). When this knob is turned up to between 8-10 I get a low frequency oscillation(haven't measured this frequency yet). When this knob is turned down below 5 I get an 8khz oscillation(measured with scope). The oscillations aren't constant, but they eventually happen when the right note is played I guess. Once the oscillation starts I can make it go away by turning the presence knob to around 7 or so, or by hitting the standby switch off(kills the power to the output which is approx 500Vdc) and then on.
This amp has a footswitch to switch between normal and boost modes and the same oscillations occur in both modes which points me to V104(12AX7WA) or V105,106(EL34). I have tried to measure the components(Rs and Cs) around V104,V105 and V106. The Rs measure close to the values indicated by the stripes on the component body. Some of the Cs measure close to the proper values some don't. I'm not sure they are bad though because the measurements were done with the components still soldered in the board.
I am looking for some advice on what to do next. Should I take the Cs out and test them, or should I start somewhere else? Any advice is appreciated!
Thanks
The problem is this: there is a knob on the back of the amplifier that is called presence(VR201 on mains schematic). When this knob is turned up to between 8-10 I get a low frequency oscillation(haven't measured this frequency yet). When this knob is turned down below 5 I get an 8khz oscillation(measured with scope). The oscillations aren't constant, but they eventually happen when the right note is played I guess. Once the oscillation starts I can make it go away by turning the presence knob to around 7 or so, or by hitting the standby switch off(kills the power to the output which is approx 500Vdc) and then on.
This amp has a footswitch to switch between normal and boost modes and the same oscillations occur in both modes which points me to V104(12AX7WA) or V105,106(EL34). I have tried to measure the components(Rs and Cs) around V104,V105 and V106. The Rs measure close to the values indicated by the stripes on the component body. Some of the Cs measure close to the proper values some don't. I'm not sure they are bad though because the measurements were done with the components still soldered in the board.
I am looking for some advice on what to do next. Should I take the Cs out and test them, or should I start somewhere else? Any advice is appreciated!
Thanks
I'd be looking at coupling caps and (even more) the electrolytic decoupling caps on the power supply rails. This sounds a lot like motorboating, which is usually induced via the power supply rails.
Hi,
I've seen this before. It is motorboating due to supply decoupling. Look at the filter caps for sure. Also make sure of the ground paths and that no one has increased the gain.
-Chris
I've seen this before. It is motorboating due to supply decoupling. Look at the filter caps for sure. Also make sure of the ground paths and that no one has increased the gain.
-Chris
What am I looking for when scoping the power supply raiils? I take it that I would be AC coupled on the scope @ 10V/div. Will the rail be showing something abnormal when the amplifier not oscillating? or do would I only see the abnormality when the amplifier is malfunctioning? How much ripple is too much? If the electrolytic caps are bad where is the best place to get high voltage caps?
thanks for the help!
thanks for the help!
It may or may not be visible on a digital scope, but it will be a challenge to spot on an analog scope because of the long period. Given the symptoms, it's safe to assume it's happening and should be treated accordingly. "If you hear hoofbeats, expect to see horses, not zebras" as they used to teach interns.
With low volume, look to see how much audio you see on the rails. You will need to turn the sensitivity up on the 'scope. Turn the audio level up and watch the waveform. There shouldn't be much audio at all. If there is a problem here you will see it easily.
-Chris
-Chris
Thanks for the replies. I finally got a scope on the rails. I was ac coupled at 10V div. There was a large 120hz component, it was approx 10V pk-pk. This seemed pretty large to me, is this to be expected?
On top of that I had a signal generator as the input and could see this signal superimposed on the 120hz, especially at higher input frequencies( 5k-10khz region, I didn't go any higher than 10khz since I saw this).
Would you change all the alum electrolytics(pre amp and output stage) or just the ones at the output stage?
Do you know where to get a 550Vdc 47uF Alum Elect cap that is snap in mounted?
Thanks
On top of that I had a signal generator as the input and could see this signal superimposed on the 120hz, especially at higher input frequencies( 5k-10khz region, I didn't go any higher than 10khz since I saw this).
Would you change all the alum electrolytics(pre amp and output stage) or just the ones at the output stage?
Do you know where to get a 550Vdc 47uF Alum Elect cap that is snap in mounted?
Thanks
If you look at the first node of the B+ supply, the one that connects to the center tap of the output transformer, it is normal for a lot of ripple to be present. It cancels out in the push pull output stage so they get away with it. But when you move ofer to the screens node and further into the preamp supplies, that ripple should be gone. Can you still see the oscillation on the low voltage end of R108. That is the resistor between the screens and the preamp tubes.
The supply is not regulated, so it is normal to see the output signal on the first B+ node to some degree.
But if the decoupling between preamp stages is faulty then they can couple together and if phase relation is right will oscillate.
Before you dig into the board, please tap on all the small tubes to see if any are microphonic. They will scream if they are from time to time. Use a pencil or your fingernail.
Also turn the reverb off. Does that have an effect? If the reverb squeals, try changing R20 from 4k7 to 1k.
Vr201 is the presence control, and it is in the feedback loop of the power amp. REgardless of the genesis of the symptom, turning it up will exacerbate the problem.
One other thing to try is to disconnect the speaker in the combo cab and play through a n external cab. This gets any vibration away from the tubes and chassis. That will also tell you if there is a mechanical connection involved.
Are you solid reading this thing? It is an inconveniet format.
The supply is not regulated, so it is normal to see the output signal on the first B+ node to some degree.
But if the decoupling between preamp stages is faulty then they can couple together and if phase relation is right will oscillate.
Before you dig into the board, please tap on all the small tubes to see if any are microphonic. They will scream if they are from time to time. Use a pencil or your fingernail.
Also turn the reverb off. Does that have an effect? If the reverb squeals, try changing R20 from 4k7 to 1k.
Vr201 is the presence control, and it is in the feedback loop of the power amp. REgardless of the genesis of the symptom, turning it up will exacerbate the problem.
One other thing to try is to disconnect the speaker in the combo cab and play through a n external cab. This gets any vibration away from the tubes and chassis. That will also tell you if there is a mechanical connection involved.
Are you solid reading this thing? It is an inconveniet format.
Hi sparkaholic,
10V is too high for ripple. The ESR is high on these. Take a small value cap (say 0.22 ~ 1.0 uF) and place it from B+ to the cap negative. This may solve the oscillation and prove this point. Bypass all the supply caps.
-Chris
10V is too high for ripple. The ESR is high on these. Take a small value cap (say 0.22 ~ 1.0 uF) and place it from B+ to the cap negative. This may solve the oscillation and prove this point. Bypass all the supply caps.
-Chris
WHile bypassing the filters may be good design engineering, this amp works well without them. His amp being an exception. Self-oscillation is not a problem with these amps, so I suspect his trouble is not the lack of bypass caps or it would have been this way out of the box.
I would also add that 2-3% ripple is pretty good on an amp like this. Plenty of amps with more. AS long as it is gone by the second node, we won't hear it. His trouble is oscillation, not hum.
I would also add that 2-3% ripple is pretty good on an amp like this. Plenty of amps with more. AS long as it is gone by the second node, we won't hear it. His trouble is oscillation, not hum.
Hi Enzo,
My suggestion was to determine if the HF impedance of the existing caps has gone too high with age. These caps are much cheaper to buy than just replacing the existing caps. That is why I suggested it, not to modify or "improve" the design of the amp in question.
-Chris
My suggestion was to determine if the HF impedance of the existing caps has gone too high with age. These caps are much cheaper to buy than just replacing the existing caps. That is why I suggested it, not to modify or "improve" the design of the amp in question.
-Chris
Fair enough.
I have so many encounters with customers who think upgrading to the newest operating system ROM in something will cure the sudden loss of one channel or something. SO I have a built up sensitivity to modifications in the face of trouble symptoms.
I have so many encounters with customers who think upgrading to the newest operating system ROM in something will cure the sudden loss of one channel or something. SO I have a built up sensitivity to modifications in the face of trouble symptoms.
amplifier oscillation
I took a look at the ripple in various parts of the amp. There is ripple on C111 and C112. Should there be ripple on C111? I could not see any ripple on C110.
Turning reverb up or down didn't affect the oscillation.
I tapped on the tubes with a pen and this didn't cause any problems, so the tubes are not microphonic.
Turning the presence pot down to 0 or turning treble up near 10 causes the oscillations to start.
I took a look at the ripple in various parts of the amp. There is ripple on C111 and C112. Should there be ripple on C111? I could not see any ripple on C110.
Turning reverb up or down didn't affect the oscillation.
I tapped on the tubes with a pen and this didn't cause any problems, so the tubes are not microphonic.
Turning the presence pot down to 0 or turning treble up near 10 causes the oscillations to start.
amplifier oscillation
I also soldered some .47uf 1000V film caps in parallel with C111 and C112. It did not solve the problem.
I also soldered some .47uf 1000V film caps in parallel with C111 and C112. It did not solve the problem.
You still haven't done anything to properly decouple the various stages. Adding small caps can help high frequency problems, but motorboating can only be fixed by separating the power rail feeds to the various stages with individual RC filters and/or tight, active regulation.
He may have a decoupling problem, but JTM60s are not prone to motorboating, so I can't think of it as a design flaw.
Have you played with lead dress at all?
I wouldn't expect very much ripple at C111. A little. C110 is on the other end of a 10k. Ought to be clean.
My guts, for what it is worth, are telling me power amp problem. If it will do it with the master all the way down, VR13, that breaks the path from the preamp for decoupling problems created there.
I might pull CN104 and verify the resistors in the tail of the phase inverter, and also check C129 there. And is C127 OK?
Just a wild stab - if R107 were open, then the presence feedback resistor through the speaker winding would be the ground path for the phase inverter. Short of testing that resistor - which we already did above - pull CN104 and see what happens. DO DC voltages in the phase inverter get screwy? That is the feedback path, so maybe the feedback is all that is keeping the amp stable.
It is not in my schematic, but if Marshall has added a small disc cap between screen and plate of one of the output tubes, check it for shorted. It would be like 27pf or something and near the socket. This cap is on some larger models.
Have you played with lead dress at all?
I wouldn't expect very much ripple at C111. A little. C110 is on the other end of a 10k. Ought to be clean.
My guts, for what it is worth, are telling me power amp problem. If it will do it with the master all the way down, VR13, that breaks the path from the preamp for decoupling problems created there.
I might pull CN104 and verify the resistors in the tail of the phase inverter, and also check C129 there. And is C127 OK?
Just a wild stab - if R107 were open, then the presence feedback resistor through the speaker winding would be the ground path for the phase inverter. Short of testing that resistor - which we already did above - pull CN104 and see what happens. DO DC voltages in the phase inverter get screwy? That is the feedback path, so maybe the feedback is all that is keeping the amp stable.
It is not in my schematic, but if Marshall has added a small disc cap between screen and plate of one of the output tubes, check it for shorted. It would be like 27pf or something and near the socket. This cap is on some larger models.
I am sorry. What I had in mind was that maybe C111 is not doing its job. I should have said clip another similar filter cap in parallel with C111 to take the place of any missing capacitance. I did not mean add HF bypass cap. I should have been more clear.
Did you try the phase inverter stage stuff?
I have a TSL100 here on the bench that I can't figure out, so I am hoping that I might have some insight into yours. Then at least something will be going right today. TSL problem is not oscillation though.
Did you try the phase inverter stage stuff?
I have a TSL100 here on the bench that I can't figure out, so I am hoping that I might have some insight into yours. Then at least something will be going right today. TSL problem is not oscillation though.
- Status
- Not open for further replies.