mosfet power amp transistor replacement and debugging

[Lee Collier]

Hello,

I've got a Rauch PB44 power amp that uses the 'classic' J50/K135 paired
Mosfets. One channel is perfect, but the other just gives me a loud mains
hum that does not change with the front volume pots, and doesn't have the
input signal present at all.

I've tested the PSU and all seems to be well (haven't checked it on the
scope yet but DVM says +-70VDC which is the same as the good channel).
I've tested the Mosfets with the DVM using the method at
http://www.4qd.co.uk/serv/mostest.html comparing good and bad channel, and
have come to the conclusion that all four J50s on the bad channel are
duff. I can source replacements at RS but they're about GBP 10 a pop.

I therefore have a few questions:

1. I don't want to waste money on replacement parts if I'm wrong in my
guess. How much damage am I likely to cause if I swap the J50s from the
good channel? I understand they're static sensitive, is it enough to
ground the soldering iron tip and touch the (earthed) chassis frequently?
I have read that solder pumps are bad for static so plan to buy some wick.
If the problem is before the Mosfets, is it possible to kill the good lot
and end up with no amp at all?

2. The Mosfets are Japanese J50/K135, the RS parts are UK 2SJ50/2SJ135.
If I replace the J50s with 2SJ50s, will I need to replace the K135s with
the different parts, or are they drop-in equivalent? I haven't been able
to find this out anywhere. I think 7 new Mosfets (I happen to have 2)
might make the repair uneconomical.

3. Any hints on other debugging I could try? I'm proficient with
solderig iron, DVM and fairly happy with my scope but not experienced with
fixing power amps.

Many thanks in advance. This is a beast of an amp so it would be fun to
have it working...

Lee

 

[Guest]

Hi Lee

Apparently you are not blowing fuses so it would seem your J50 outputs
failed open. Check to see if your Mosfets have source resistors as well as
gate resistors. Check every one, as well as everything else on the board. I
don't have access to a schematic for this amp but I would recommend getting
one. Fixing an amp like this can be a lot harder than it looks without a
schem - even with a schematic for that matter! What you don't want to do is
put new output devices (or the good ones from your working channel) in there
and destroy them due to possible un-detected problems upstream in the
circuit.

I have not yet managed to ruin a mosfet by careless handling despite all the
warnings for static handling - though if you have a wristrap- use it. They
will destroy very easily though if you hook em up wrong.

If you have a scope & other basic test gear and a current limiting light
bulb to insert in mains, and you are brave:

In lieu of a schematic, I usually do a thorough component check, and then
attempt to run the amp without outputs installed (do not attach a spkr or
dummy load yet!) if the circuit will let you do that and observe the output
waveform via scope probe with a sinewave @input. Sometimes you can operate
an amp like this and sometimes you will need to install at least one comp.
pair of output devices to satisfy biasing & feedback requirements to get the
amp up. Do none of this stuff without a current limiting bulb in series with
one side of your mains. A 60 w bulb (120v mains) will usually allow you to
get such an amp up yet limit current enough so that you wont sizzle anything
while you are troubleshooting. If your light bulb glows real bright you
likely have a problem somewhere.

The idea is to determine if the preamp & driver stage is working right
before installing the expensive outputs - which you can about 90% determine
by o'scope probing and following the above method. If you observe a normal
looking amplified version of your input waveform then chances are your
preamp & driver stages are in good shape. Put in the outputs, or the rest of
the outputs and check again. If you have any appreciable DC (more than
about 100mVdc) at the output while the input is shunted - you have problems
somewhere up stream. When you are confident your sine wave looks good - and
you have all the outputs installed, hook up your spkr load and check again.
(Say a blessing prayer first).

If you have a bulb Ilimiter and follow these two rules - you should be able
to fix this thing:
1)do not try to drive a spkr load without the outputs installed.
2)do not install the outputs until you are confident that the preceeding
stages are working correctly.

See this link for some very clever advice on replacements for the obsolete
J50/K135 devices:

http://aga.rru.com/TechTips/disp.cgi?file=RobertBraught/PowerFETRepl.tt

BTW, I would recommend replacing all of the outputs.

Good luck
Todd

 

[Woody]

Hello,

I've got a Rauch PB44 power amp that uses the 'classic' J50/K135 paired
Mosfets. One channel is perfect, but the other just gives me a loud mains
hum that does not change with the front volume pots, and doesn't have the
input signal present at all.

I've tested the PSU and all seems to be well (haven't checked it on the
scope yet but DVM says +-70VDC which is the same as the good channel).
I've tested the Mosfets with the DVM using the method at
http://www.4qd.co.uk/serv/mostest.html comparing good and bad channel, and
have come to the conclusion that all four J50s on the bad channel are
duff. I can source replacements at RS but they're about GBP 10 a pop.

I therefore have a few questions:

1. I don't want to waste money on replacement parts if I'm wrong in my
guess. How much damage am I likely to cause if I swap the J50s from the
good channel? I understand they're static sensitive, is it enough to
ground the soldering iron tip and touch the (earthed) chassis frequently?
I have read that solder pumps are bad for static so plan to buy some wick.
If the problem is before the Mosfets, is it possible to kill the good lot
and end up with no amp at all?

2. The Mosfets are Japanese J50/K135, the RS parts are UK 2SJ50/2SJ135.
If I replace the J50s with 2SJ50s, will I need to replace the K135s with
the different parts, or are they drop-in equivalent? I haven't been able
to find this out anywhere. I think 7 new Mosfets (I happen to have 2)
might make the repair uneconomical.

3. Any hints on other debugging I could try? I'm proficient with
solderig iron, DVM and fairly happy with my scope but not experienced with
fixing power amps.

Many thanks in advance. This is a beast of an amp so it would be fun to
have it working...

Lee

If you are getting hum on the speakers, disconnect them and check the d.c.
voltage to ground on the output terminals. If the amp is d.c. coupled -
which it very likely will be - there could be up to perhaps 100mV (either
polarity) present. If an output device has failed short circuit then you
will see either full or nearly full rail (+ve or -ve) if one device has
gone, or an offset greater than 100mV if a complimentary pair have failed,
although in this latter case I would have expected it to blow the rail
fuses.

J50 and 2SJ50 are one and the same - it is just the normal way that letters
are omitted to save space: same for K135 and 2SK135. However do make sure
they are the same encapsulation. Originally they were in the TO3
diamond-shaped power transistor cases, but latterly they were repackaged in
a T-shaped flatpack, albeit with some care these could be made to fit the
same heatsink assuming there is enough room.

As said elsewhere unless there is an obvious problem you will have a
difficult job without a circuit diagram. Your DVM will take so little
current that good devices could still appear to be faulty because of
leakage - a common trap into which to fall.

 

[DaveC]

If you have a scope & other basic test gear and a current limiting light
bulb to insert in mains, and you are brave:

Excellent FAQ on diagnosing and repairing an amp. Just what I needed when
I'll (some day) attack that old 400W boat anchor I want to drive my antique
electrostatic speakers. : -)

Thanks, Todd!
--
Please, no Google links. I wouldn't ask a
question here if I hadn't done that already.

DaveC
[email protected]
This is an invalid return address
Please reply in the news group

 

[Lee Collier]

Hi Lee

Apparently you are not blowing fuses so it would seem your J50 outputs
failed open. Check to see if your Mosfets have source resistors as well as
gate resistors. Check every one, as well as everything else on the board. I
don't have access to a schematic for this amp but I would recommend getting
one. Fixing an amp like this can be a lot harder than it looks without a
schem - even with a schematic for that matter! What you don't want to do is
put new output devices (or the good ones from your working channel) in there
and destroy them due to possible un-detected problems upstream in the
circuit.

Thanks - I wish I had a schematic, but I've spent hours searching to no
avail. It's a Rauch Powerblock 44 from a now-defunct English company.

The idea is to determine if the preamp & driver stage is working right
before installing the expensive outputs - which you can about 90% determine
by o'scope probing and following the above method. If you observe a normal
looking amplified version of your input waveform then chances are your
preamp & driver stages are in good shape. Put in the outputs, or the rest of
the outputs and check again. If you have any appreciable DC (more than
about 100mVdc) at the output while the input is shunted - you have problems
somewhere up stream. When you are confident your sine wave looks good - and
you have all the outputs installed, hook up your spkr load and check again.
(Say a blessing prayer first).

If you have a bulb Ilimiter and follow these two rules - you should be able
to fix this thing:
1)do not try to drive a spkr load without the outputs installed.
2)do not install the outputs until you are confident that the preceeding
stages are working correctly.

OK - the above is really useful, thanks very much. I've got 240V mains
here but I can do the sums.

See this link for some very clever advice on replacements for the obsolete
J50/K135 devices:

http://aga.rru.com/TechTips/disp.cgi?file=RobertBraught/PowerFETRepl.tt

Yeah, I read that. I can buy the originals for about GBP 8 each, and the
IRFP240/9240 for about GBP 5 each, so I'll have to decide whether I want
the money or the hassle.

BTW, I would recommend replacing all of the outputs.

Good luck
Todd

That's all been really helpful, I'll get on to it next weekend.

Lee.

<SNIP>

 

[Lee Collier]

On Sat, 17 Jul 2004 14:16:09 +0000, Woody wrote:

If you are getting hum on the speakers, disconnect them and check the d.c.
voltage to ground on the output terminals. If the amp is d.c. coupled -
which it very likely will be - there could be up to perhaps 100mV (either
polarity) present. If an output device has failed short circuit then you
will see either full or nearly full rail (+ve or -ve) if one device has
gone, or an offset greater than 100mV if a complimentary pair have failed,
although in this latter case I would have expected it to blow the rail
fuses.

The probing I've done so far suggests they've failed open, but I'll give
that a go in a minute.

J50 and 2SJ50 are one and the same - it is just the normal way that letters
are omitted to save space: same for K135 and 2SK135. However do make sure
they are the same encapsulation. Originally they were in the TO3
diamond-shaped power transistor cases, but latterly they were repackaged in
a T-shaped flatpack, albeit with some care these could be made to fit the
same heatsink assuming there is enough room.

As said elsewhere unless there is an obvious problem you will have a
difficult job without a circuit diagram. Your DVM will take so little
current that good devices could still appear to be faulty because of
leakage - a common trap into which to fall.

Good point, although I am comparing good side with bad side - my DVM
is turning on both of each pair on the good side but on the bad side the
N-channel devices are turning on OK but the P-channel are not