Marshall Mosfer Lead 100 diy repair help

[jjonas]

Hi,

I've had a Marshall Mosfet Lead 100 for some years, it's needed repair twice, and now it's the third time there's something wrong with the sound. I don't feel like handing out another 100€ for a repair, so I thought I'll try to do it myself (rather than give up the whole thing). I'm familiar with electronics in general, I have built a dozen guitar effects and some synthesizer stuff, but I'm not particularly familiar with amplifiers. I have a rudimentary ability to read schematics, know e.g. the basics about voltage divers etc., but I can't really read complex schematics like for amps without someone holding my hand.

So I do have intermediate skills in electronics, in a way that I believe allows me to understand most advice I will get here, but I won't manage without help.

Here's a schematic for the amp:
http://www.drtube.com/schematics/marshall/3210-iss7.gif

The problem: When I plug in, I can hear the guitar only faintly (the faint signal's kind of clean). When I strum harder, the volume jumps up a lot, and the sound is very distorted, almost in an "8-bit" way or something. When I switch the amp off, there's a loud swoosh, plus some guitar sound for about a second, if the strings are ringing at the time of switch-off.

When I plug into the return jack, it's basically the same, only without the faint clean sound, so I suspect it makes sense to start by checking the stuff between the return jack and output. Please correct me if this doesn't sound like a good assessment.

If I remember correctly (and I'm pretty sure I do), last time the amp was in a repair shop, IC5 was replaced. At any rate it's not the chip specified in the schematic: according to the schematic, IC4 and IC5 are MC1458, but IC5 is in fact BA4558.

So I could just desolder IC5, put a DIP8 socket in it and replace the chip with some suitable op-amp, and see if it starts working, but I'd like to complicate things a bit and ask for your help to improve my amp repair skills as well, if you dont mind :-)  Now I know that many op-amps are kind of interchangable, but I'm not sure which ones are compatible, and what are the crucial characteristics to look for in a datasheet.

What I have available in the local electronics shop are at least RC4558 (0,50€) and MC1458P. The pinout is the same, but do you think it matters which one I choose? Here are the datasheet for them, in case it's useful:

RC4558: http://pdf.datasheetcatalog.com/datasheet/texasinstruments/rc4558.pdf
MC1458P: http://pdf.datasheetcatalog.com/datasheet2/3/07j9c8li1155f0estliki76oc2cy.pdf

In case it's relevant, I've used the amp with a Boss GT-10 multifx so that the mfx connects to directly to the return jack. My impression is that this is ok, but if you think it's problematic, please explain. If it is problematic, it could explain why a chip behind the return jack gets broken repeatedly – if indeed that is the broken thing here. Anyway it's been 6-9 months since it was in repair, and the amp has been used every second week for two hours at a time.

If you are able and have the time to help, but need to know something further, please ask! I dont' have an audio probe, but I can build one if that's necessary for your troubleshooting tips.

[Roly]

Hi and welcome jjonas.


You are getting a bit ahead of yourself here.  Before we start replacing stuff we do a bit of diagnostics first, that way we hopefully identifiy exactly what is wrong and fix that, rather than the scattergun approach of randomly replacing stuff in the vain hope we will get lucky (which you almost never do).


1. The very first test is to unplug the speaker(s) and measure the DC voltage across the output.  This should be less than half a volt, but post what you find.  If it is more than half a volt do not reconnect the speakers.

2. The amp has an Fx Send and Return, so the next test is to plug a known good lead into both and see if that makes a difference.

3. Next we take a signal from Fx Send to another amp and see if that sounds okay or has the fault.

4. Similarly you can try plugging your guitar directly into Fx Return and see if that is clean or has the fault.

In this way we can tell of the problem is in the pre or main amp.

5. If you have it out of its case you can also check all the power supply rails are within about 10%, +/-15V, +/-34V and +/-49V.

Post your findings here and we will move to the next stage.

[jjonas]

You are getting a bit ahead of yourself here.  Before we start replacing stuff we do a bit of diagnostics first, that way we hopefully identifiy exactly what is wrong and fix that, rather than the scattergun approach of randomly replacing stuff in the vain hope we will get lucky (which you almost never do).

Hi, thanks for a quick reply!

(UPDATE: I think the problem is solved in its acute form, please read to the end! I found out the problem was solved while I was going through with the tests you recommended. Nevertheless, I decided to let the reply to reflect that.)

I must confess that I've already made an audio probe and checked where I can hear a signal. I could hear it at R38 (the send output resistor), but at R42 and R43 the signal was crackling. I checked the voltages in IC5 pins 4 and 8, and they checked out (-16,6 and +16,6VDC). When I found an MP1458 in one of my boxes I did replace the BA4558 (with an added DIP8 socket). Now I can hear the signal also at R42 and R43.

As to the signal from the speaker, it is not breaking up anymore, but it's not amplified to a level it should.

1. The very first test is to unplug the speaker(s) and measure the DC voltage across the output.  This should be less than half a volt, but post what you find.  If it is more than half a volt do not reconnect the speakers.

It's -300mA. I'm not sure whether it matters that it's negative, but I'm sure I have the multimeter heads the right way.

2. The amp has an Fx Send and Return, so the next test is to plug a known good lead into both and see if that makes a difference.

Assuming you mean that the Send/Return should be connected together using a single lead: done. The reading stays the same.

3. Next we take a signal from Fx Send to another amp and see if that sounds okay or has the fault.

That seems to work, i.e. with the other amp I get a good clear signal, even when I turn the Marshall's volume up.

4. Similarly you can try plugging your guitar directly into Fx Return and see if that is clean or has the fault. In this way we can tell of the problem is in the pre or main amp.

I'm using a looper pedal for the signal, I connected it to Return only now, and the amp seems to be working now..! This is actually the first time I notice that it's working properly after replacing the op-amp, I must have had something wrong with my test setup.

5. If you have it out of its case you can also check all the power supply rails are within about 10%, +/-15V, +/-34V and +/-49V.

Measured values (at R44-47):
-49,6 / +49,7
-37,0 / +36,8
-16,6 / +16,5

Anyway, the amp seems to be working now. The follow-up question is, though, what could cause the same chip to fry repeatedly? The Vcc and Vee voltages were within tolerance, though it must be said that when the absolute maximum ratings in the datasheets for both BA4558 and MC1458 are +-18VDC, the measured +-16,5 or so sounds pretty close. On the other hand, there's another MC1458 in there with the same voltages, and it's doing fine.

Any ideas?

[g1]

  Stick with 1458's and see if it improves reliability.
Normally 4558 is ok to sub for 1458.  However, there are odd cases where they don't work well, and if I'm remembering correctly, the only place they seem to have issues is in some Marshall amps.  I can't remember which models or why it is a problem.  I think I have seen Enzo mention the same issue and he may remember more details.

[J M Fahey]

The 4558 is "better" than 1458 , as in less distortion and noise, higher slew rate and highfrequency response, simply because it´s more modern and made foe Audio, while 1458 is an older general purpose Op Amp.
But I think that in practical use it can supply more current into the output, so it was retained in the Reverb drive stage, even if 4558 or TL072 (or M5201) were used somewhere else in the same amp.
In the loop job, both are equivalent.

As to what fries it: loop out/in and pream in are often fried by strong external signals, usually some speaker out there and blows it to pieces.

It´s advisable to add a series resistor and a couple protective diodes to send dangerous voltages away.

In your case Rs may be 4K7 and diodes plain 1N4002 .

[jjonas]

Thanks for insights!

As to what fries it: loop out/in and pream in are often fried by strong external signals, usually some speaker out there and blows it to pieces.

Can you elaborate what you mean with "external signals, usually some speaker out there"..?

It´s advisable to add a series resistor and a couple protective diodes to send dangerous voltages away. In your case Rs may be 4K7 and diodes plain 1N4002 .

Do you advise that this be done to both IC4 and IC5 (with series resistors for both sides of the opamps), or only IC5..? So far IC4 hasn't caused any problems, but in principle is it in the danger zone in the same way as IC5..?

One way to implement this that occurs to me is making a small daughterboard with the added components and the chip, which is then pressed into a DIP8 socket on the main board.

[DrGonz78]

1. The very first test is to unplug the speaker(s) and measure the DC voltage across the output.  This should be less than half a volt, but post what you find.  If it is more than half a volt do not reconnect the speakers.

It's -300mA. I'm not sure whether it matters that it's negative, but I'm sure I have the multimeter heads the right way.

2. The amp has an Fx Send and Return, so the next test is to plug a known good lead into both and see if that makes a difference.

Assuming you mean that the Send/Return should be connected together using a single lead: done. The reading stays the same.

When you say the reading stays the same are you referring to the DC offset on the output, the 300mv? If that is what you meant then the problem might have been a dirty return contact. I may just be misunderstanding the comment but it might not have been the opamp at all. What you wanted to know is if the audio signal was restored by connecting the lead between the send/return connection. If that be the case then by inserting the jack you might have just woke up the amp and not realized it, if you were not monitoring the signal.

Edit: Sorry I get it now that the opamp really was bad. But still I would give a good cleaning to the send/return jacks.

[J M Fahey]

Thanks for insights!

As to what fries it: loop out/in and pream in are often fried by strong external signals, usually some speaker out there and blows it to pieces.

Can you elaborate what you mean with "external signals, usually some speaker out there"..?

It´s advisable to add a series resistor and a couple protective diodes to send dangerous voltages away. In your case Rs may be 4K7 and diodes plain 1N4002 .

Do you advise that this be done to both IC4 and IC5 (with series resistors for both sides of the opamps), or only IC5..? So far IC4 hasn't caused any problems, but in principle is it in the danger zone in the same way as IC5..?

One way to implement this that occurs to me is making a small daughterboard with the added components and the chip, which is then pressed into a DIP8 socket on the main board.

1) external signals means exactly that, something coming from outside the amplifier.
How can something from outside get inside?
Through a "door" which in this case is the loop return jack.

Somebody, even yourself, could plug some amp speaker out, which is a *very powerful*  signal into the loop return jack; by mistake of course.
IC 5a is expecting, say, 200mV at a very feeble current, whatever a pedal puts out, and gets a whopping 20000mV, 5 ampere signal (a speaker out) and of course dies.

Nobody will do that on purpose, but mistakes happen, specially on a dark, noisy, smokey stage late at night, with people nervous, in a hurry, with maybe some solid, liquid or gaseous chemicals disrupting their brains.
It takes just one "oops !!!!!" to fry it for good.

To make that input foolproof, Rs limits the current , and those diodes do not let voltage go above +15V or below -15V.
Clever idea, huh?

2) You do NOT need a daughterboard at all, just:

a) desolder the right end of C 35 and lift it away from the board.
Now you have a capacitor with a leg "in the air" and an empty hole in the PCB.
Add a 4700 ohms, 1/4 W resistor from that hole to the floating leg.
b) on the bottom side of the PCB solder a diode (I suggest a 1N914 or 1N4148 because of the size and thin legs) from the hole where you just soldered the resistor to +15V or pin 8 of IC5a , ärrow" pointing to +15 side.
c) solder another diode from that same hole to -15V or IC5a pin 4 , now arrow pointing away from -15V.
Just look at the little schematic I posted above.

Won't look as nice as if factory added but it will work.

Check you didn't short some IC pads together.

*IF*  you wish, you may also add protective diodes to the preamp input, on IC1a, which also has a "door" (the input jack).
There use a 22K to 68K input resistor, and 2 diodes, from leg 2 to legs 8 and 4 as before.

FWIW many amps add those diodes; Peavey on most inputs and unjustly despised Crate **everywhere**, on inputs and outputs.

Maybe they have a dark view about what Musicians can do

[Roly]

It's -300mA.

I do hope you mean mV.  The offset may be of either polarity, just important that it isn't too much or the speaker may be damaged.


Points 3 and 4.  The Send and Return sockets are bridged by a contact when nothing is inserted (at JS4).  This contact quite often gives trouble by not making proper contact and a shot of Deoxit or metho followed by a scrub by inserting and removing a plug several times to work the contacts will often fix it.  Your comments about R42/3 and under point (4) tend to suggest that your problem might actually be this contact being dirty/intermittent.  The object of bridging Send and Return with a known good lead is to bypass this contact and see if it clears the fault.


Repeated failures of a given op-amp may be due to a design fault, but often more likely to be due to its input or output being externally accessible and getting hit with excessive voltage at some point.  It is not unknown for an amp output to get connected to another amp input (or output) either deliberately, or more commonly by accident in the heat of setting up for a show, even perhaps a static discharge if there is a synthetic carpet and air con, and if the circuit doesn't have sufficient protection you can get repeated failures.

In this amp IC5a is fully exposed to any excessive voltage coming in via the Return input, so JM's suggestions apply.

The supply voltage rating is the voltage that 100% of devices will survive.  In practical terms this normally means that there is a bit of a safety margin until the smoke actually comes out so you might find that 90% of parts will survive say 20 volts, 80% 25 volts, and so on.  16-17V on a nominal 18V limit is not a problem.

Maybe they have a dark view about what Musicians can do

And being musicians ourselves we have a very good idea of what musicians can do.   :duh :lmao:

[jjonas]

What a great forum!!

@J M Fahey: Thanks for clarification, it was helpful! Exactly the kind of stuff I wanted to know. Also thanks for the implementation tip, probably I'll do just that.

@Roly: Yes I meant millivolts. Also thanks for further explanations!

Edit: A few more questions. I noticed when I was making the modifications for IC1a, that the 10nF C1 that I'm supposed to "float" the other leg of, has been replaced (not by me) with a 42ohm resistor. I remember that once when this amp was being serviced several years ago the repair guy said that he fixed a few "typical Marshall faults" or something.

Other caps (all 10nF) that have been replaced with 42 ohm resistors are

- C8 (in the schematic, top segment, 5th upward wire off the 0V rail from the left)
- C22 (top segment, middle part, above TR2)
- C32 (middle segment, to the right of the reverb unit connection)

Any idea what significance these modifications have?

Anyway I replaced the 42 ohm resistor that was in place of C1 with a 51k resistor (plus added the diodes), I hope that's ok.

[bluesky6]

I did a repair on a Vox Pathfinder 15R last weekend where there was no reverb. Replacing the 4558DD fixed the problem. The op-amp out to the reverb unit was working but the op-amp input from the reverb had died.

This may be a common issue. Some of the other amps I've looked at use a 4560 op-amp pair instead of the 4558.

[J M Fahey]

Some capacitors, notably used by Marshall, come in a shape and colur that *looks*  like a resistor ... yet they are caps.
So I don´t think those are resistors at all but caps.
A detailed close up picture would be great, but start by posting all colour bands and any other code printed/painted on them.

[jjonas]

The colour bands are yellow-red-green-brown. Red could be orange too, it's hard to tell for sure, but I'd say it's red. The component body is light green. No text or markings can be seen. Posting pics is no problem.

[J M Fahey]

If you read them Br-Gn-Or-Ye according to
http://www.pmel.org/Handbook/HBpage26.htm
they are 15000pF or 153 or 15n x 400V, a very reasonable value , specially in the tone control.
Standard bass capacitor there is 22n and a slightly smaller one such as 15n would increase midbass somewhat, a reasonable goal in an SS amplifier which probably was perceived as "less full" or thinner than equivalent tubed ones.

[jjonas]

Wow, another thing learned. I've never seen caps like that before. I've attached a pic I took just for the benefit of others who haven't seen caps like that.

BTW in the end I did solder the 51k resistor in series with the "42ohm resistor that was a cap" instead of replacing it, just to make sure. Glad to get confirmation.