Marshall IBS 3520 200 watt problem

[woodyc]

...Could that BR have been at the end of it's life?...

I'm not following your diagnostic process but the bridges do fail. If you look at the 3520 schematic they spec 25A bridges which was smart. But what they installed was 6A bridges with no heatsink and no airflow on one side. To get them out without damaging the board, I cut them in pieces...

[DrGonz78]

Anyhow... Well the good news is that I have voltage readings measuring from the secondary CT ground. Voltage going into to each side (~) of the BR is +45 vac -45 vac. The B+ and B- voltage is sitting at 61 and -61 vdc. So this looks good. I will start by testing with the power board connected and running light bulb limiter. Will test for DC on the output again too. What else should I be checking?

I'm not following your diagnostic process but the bridges do fail. If you look at the 3520 schematic they spec 25A bridges which was smart. But what they installed was 6A bridges with no heatsink and no airflow on one side. To get them out without damaging the board, I cut them in pieces...

Yeah I was meaning if a bad solder joint on the B+ or B- part of the BR could drop voltage feeding the power board. Probably just one of my many over processed thoughts... This one had BR84-8432 8A bridge at least but yeah I love when I see those 25A type bridges. I see those 25A's mounted to the chassis and then point to point wiring.

[DrGonz78]

So I finished it all up. I then ran it through the light bulb limiter and tested voltage coming off the output. I can't remember but it was less than .6 vDC. Now I am not sure if that is okay to have any DC on the output but I figured lets plug this in to 120 mains. I was a bit nervous since the last time it flamed up!! Anyway I powered it up and no more 7 volts DC on the output! I am getting about .6 vdc on the output still, but that seems pretty insignificant compare to 7. So, I couldn't resist plugging in a speaker to test the amp... It sounded great!!!

I guess it is fixed but still wondering about the .6 vdc and if that is something to worry about? Also, the speaker does push out when turned on but does not stay in that position and when I turn it off the speaker does that push thing again. Does not make a popping noise and really not sure if that is normal but could be...??? Any comments are greatly appreciated before I do more from this point forward.

Just wanna thank all that helped contribute on this thread and especially Juan and Markus!! You guys Rock

[MarkusBass]

I guess it is fixed but still wondering about the .6 vdc and if that is something to worry about?

If you read the forum carefully, you will find a very simmilar thread started by Lowell in the Guitar Amp forum - Biasing Trace Elliot (which is very simmilar to the Marshall you are working on). He also has problems with DC on the output. In my opinion DC offset in the range 30-50 mV is acceptable but 600 mV is not. This may be a result of not matched transistors (especially differential pair on the input: TR1/TR2). I suggest that you desolder them and measure the Hfe coefficient of each of them. It also may be a result of not matched MOSFETs but you cannot do much about it. Just check the input pair. You can also measure DC voltage on base of TR1 and TR2 (before desoldering). As I mentioned in the other thread, I was fixing lately simmilar amp (actually it was Marshall 3520 ) and it had only 30mV on the output. I didn't investigate the problem since for me such offset is acceptable.

EDIT:

I had to change out one A06 since it fried up.

It seems that you made the same mistake as Lowell did: if you change transistors in differantial pair amp, you have to change them both and make sure they are matched. Otherwise, you get DC on the output.

Mark

[Enzo]

The whole point of the solid state power amp is that it is one big loop. If your output devices were unbalanced so as to cause an offset, the differential circuit will correct it. That is why offset adjust controls - in amps that have them - are on the left side rather than the output side of the circuit.

>6v offset is more than I like, but does it remain when speaker load is ppresent? Or does it go away?


You say it doesn;t pop on power up/down, but the speaker cone pulses in or out. Well, that is just a pop at too low a frequency to be audible. Same deal.

[Enzo]

The whole point of the solid state power amp is that it is one big loop. If your output devices were unbalanced so as to cause an offset, the differential circuit will correct it. That is why offset adjust controls - in amps that have them - are on the left side rather than the output side of the circuit.

>6v offset is more than I like, but does it remain when speaker load is ppresent? Or does it go away?


You say it doesn;t pop on power up/down, but the speaker cone pulses in or out. Well, that is just a pop at too low a frequency to be audible. Same deal.

[J M Fahey]

*Fully* agree with Markus' and Enzo's posts and add: as Enzo said, even quite mismatched power transistors will be corrected .................... if "the corrector" itself is balanced !!
In this case it's TR1 and TR2, MPSA06 ; which in my experience do not have that much Beta/Hfe/current gain nor are particularly consistent.
I much prefer BC550/560 family, etc. as input differentials, although I'm not telling you top replace them , at least not yet.
Start by measuring (carefully, your meter tip slips and you make a mess) voltage to ground on Tr2 base (measure across R5) and on TR1 base (from top of C10 or left of R10 to ground).
Let's see what you find, but basically if
1) you measure around 600mV on Tr2 base, the differential is doing its job, simply the reference voltage is bad.
Usually because Tr2 has very low Beta.
2) or said voltage is low (50/100 mV) but Tr1 base is far from it: very mismatched transistors or:
3) C10 is leaky, so the loop is no DC unity gain any more.
Check this by lifting one of its legs .
Is offset voltage much lower now?
In that case replace C10 by a known good new one.
No, no NOS babble here, use a fresh modern one.
And good luck.

[DrGonz78]

So I just checked the voltage on the bases of both TR1 and TR2. What I got was TR1 very low voltage like .05mv and it would go down to 0mv as I held the probe there for readings. TR2 was reading about 500 mv or little more up to 530mv and it too would drop as I held meter probe on it. So yes this TR2 is new and TR1 is the old transistor. I decided early on to replace both C10 and C11. They are brand new parts ordered straight from mouser and are good replacements. So I will do some HFE checking later and try to match up two of these A06 transistors to try them out. The only parts I have replaced are the 4 resistors that blew, TR2, TR4, and C10/C11. Those old electrolytic caps needed to be pulled out was my thinking and would save me if one was leaky. That was my thinking when I replaced those, and sorry I forgot to mention those before.

Also, yes Enzo the speaker does this push outwards and then when turned off it is the same thing again. Like a hiccup. I did not test to see if the 600mv DC goes away while the load is connected. Will test that, but first working out the deal with TR1 and TR2.

[J M Fahey]

Cool.
Post results.

[DrGonz78]

Well the voltage is down I tested the old TR1 with hfe at 57 and tested the new TR2 with hfe at 170. So, I then went through the few A06 transistor to test each one to get a matchable pair. I went w/ the original one at 170 and the new one was at 173, so that worked good. I tested the output with light bulb limiter and it shot up to 45mv then moved towards 50mv then back down to hang around 42mv. Then I ditched the bulb limiter and tested straight mains power. It Jumped up 45 then almost 50mv for a few seconds then it went way down and hung out around 32mv. I then tried it w/ a speaker connected... It still did the thump thing Anyhow, with the speaker connected it shot up to 10mv then slowly up to 20mv then it mellow down to 10mv for a while. The confusing thing here is that readings fell to about 1mv then even started to dip into the -1 then -2mv. At that point I thought that might not be good or something, just not sure so I turned it off. Then I plugged in a guitar to hear it and it sounded excellent. So I think the bias is good now, but is this oscillation or something as far as the readings are concerned. One minor thing to note is that I had looked at output w/ the positive (red) probe on the ground (-) side on the clip from the speaker plug vice versa for the other one. That way it would show me positive mv readings. That seems a bit confusing so I am not sure why it is doing that at all. I know I had just worked on a Crate BX440H and output was steady on 0v the whole time testing.

Solved one issue and maybe found another?

Also, touched all the transistor afterward and none felt really hot? TR4 & TR3 felt a tad bit warm compared to any others. Which makes me wonder about C11. I take back what I said earlier or I am correcting my statement: I meant to say that I replaced C9 & C10. I really did not replace C11 so maybe that is leaky like Juan had discussed.

[Enzo]

SO you have a few mv of offset, and it wanders around. Try monitoring your main power rails, see if they ocasionally move around, I bet they do. Then monitor your mains and see if it moves around. I bet it does. What happens if your mains moves around? Your power supplies move around, and if they move around, your circuit will react to it.

What does offset mean? It means some DC voltage other than zero on the output. 10mv 20mv 42mv, whatever are positive voltages. -1mv or -2mv simply means it went to the negative side a bit. If it starts at 10mv, and drops 12mv negative from there, then you wind up with -2mv.

I myself sure won;t worry about a few tens of millivolts.

[DrGonz78]

That makes feel better about the amp and to understand DC offset better too. I always remember DC offset from my days of recording and mixing audio. Looks like I am finally really getting to the source of what is DC offset. Always could remove DC offset from a .wav file before mastering, but only knew about audio. This is way more interactive learning, thanks.

Well another thing to point out here is that earlier today while testing voltages on the bases of the transistors the amp's power switch got stuck in the on position. I then unplugged it and it was still not really clicking... you know? Then after about 30 seconds or so the switch started to click and I plugged it back in and it's been fine ever since. Just some info to throw out there...

I will test voltage on the power rails and ac mains from the transformer soon.

[J M Fahey]

Cool.
As Enzo says, a few mV don't matter, and they may be positive or negative.
While within a few mV, nothing to worry.
As of the Thump, usually it's unavoidable.
Why? Because the input differential *will* compensate the output voltage (it is a servo corrector, after all) *but* to do so, it needs to receive accurate voltage info from the output ... which it will *not* do forva second or two because electyrolytic capacitors in the path need to charge first and reach some equilibrium.
So the corrector is "stupid" for a short time (imagine you wake up because of a loud noise in the middle of the night).
As of the slow drift, besides wall voltage variations (which happen all the time), the transistors take some time in reaching operating temperature.
Remember Monitors have to be turned on a few minutes until colors stabilize, and so on.

PS: modern amps are designed in a more symmetrical and balanced way, and often avoid thump because of that.