Arfa,
Thank you for your time. Your information is of great value to me.

I've had a look at your results, and the first thing that I feel that I
must
say, is that the output waveform from your test circuit is simply not good
enough to do any serious evaluations of what is going on. It really does
need to be a sine wave. How can I put it ? What you are doing, is a bit
like
trying to find a steering fault on your car, with one of the tyres flat.
You
either need to have a proper signal generator, where you can properly
control the output level and frequency, or use a second sound card with a
signal generator program running. You can easily build yourself a
generator
using what's generically called a 'function generator' IC or a less
sophisticated one using a phase shift oscillator.


http://www.zen22142.zen.co.uk/Circui...r/testgear.htm

Here is a circuit built around the 8038 that I plan on pursuing.

Yes, that looks fine. Put it in a nice little box, and it will serve you
well in your endeavours.



I am 55, retired steamfitter, piecing together a second income out of
a fascination for electronics and a love of music and gadgets. I know
the theory very well but I have no practical experience whatsoever,
especially in the field of trouble-shooting. I am starting out with
baling wire and bubble gum and a desire to do more, and operating on a
shoe-string budget. I have a friend who knows people with broken music
amps.


OK. Understood on your skill levels.

http://www.partsexpress.com/pe/pshow...19-030&scqty=1

I plan on purchasing an 8 ohm 200 watt dummy speaker load from
partsexpress.com so I can "pump it up" and see what it be like. I will
use my sound card generator/scope to check it out then. I would like
to be able to tap off this load to power a speaker at a low volume
level so I could hear what is going on, as well as see it.

That item looks as though it will do the job, and seems a very fair price. I
would, however, treat its quoted rated power handling of 200 watts, with a
little caution. It is likely to run *very* hot at that sort of level. If you
intend using it for high power tests as an ongoing item of test equipment,
you might consider mounting it in an additional box, with a fan blowing
across it. It is easy to 'tap off' for a small speaker. I have a dummy load
box which I built, and into which I incorporated a crude power meter. I also
fitted it with a small 32 ohm speaker, which I connected straight across the
input to the dummy load, via a fixed resistor, and a pot hooked as a
variable resistor. The series resistor was chosen to give, when added to the
variable resistor set to maximum resistance, a maximum safe drive to the
speaker when the driving amp is delivering a lot of power. On lower drive
levels, the pot can be turned down in value, to increase the drive to the
little speaker. That way, I can get a reasonable listening level from the
little monitor speaker, over a range of a few watts, to a hundred or so. I
also have a BNC socket connected straight across the input to the dummy, so
that I can make an easy connection to the 'scope, which doesn't involve
holding a probe in place. Whilst we're talking probes, I wonder what you are
using to connect to your sound card ? A proper scope and probe presents an
input impedance of at least 1 meg, with a very low capacitance. This may not
be the case with your sound card "scope" and cable, and this could result in
additional measurement errors if the impedance is low.


I'm sure you will agree that starting out with 10 mV for an input is
better than using 10 V. Bear in mind that I am shooting in the dark.
Now I know that I can safely use several volts of input on the power
amp.

Yes, agreed. Always sensible to start at a low level, as long as you know
what sort of typical level you are aiming for at various points in the
circuit. That will come with increasing experience. When you have it all
working correctly, make your own notes of expected levels and voltages on
the schematic, but remember to state what input level you were using, and
how the front panel controls were set. Such notes can be very useful for the
future. Also, when you have located the fault, it's always useful to either
note on the schematic what the symptoms and problem was, or keep a notebook
of makes and models and problems.

I would like to apprentice to an experienced tech. I could learn much
in a very short while. But that is not an option for me at this point.


I was originally apprenticed in the TV repair business some 35 years ago,
and have been self employed for probably 17 years or so now.



The drive waveforms for the bases of the outputs do look pretty
poor, but actually look more like an overdrive condition than anything
else.


Does it seem strange that there is a (possible) overdrive condition
when the input is so low? Is there something to be learned there? I
understand that the input signal is too low to make much of any of
this, except for that I need better equipment.


It's hard to comment given the limitations of the test results with the
equipment used so far. The only reason that I said that it *looked* like an
overdrive condition is that what is appearing is virtually a square wave.
Even given the poor shape of the input waveform, the fact that it seems to
have been turned into a square wave would indicate that the output from the
stage previous to where you are making the measurement, is 'hitting the
rails', which is odd, as you surmise, given the low drive level. There may
be something to be learned in that there could be something amiss with the
bias around the preceding stage, but again, it's hard to say because with DC
coupled amps, a leaky cap in a preamp, can blow an output transistor or two,
four stages further down the line.


Music injected directly into the power amp input sounds distorted and
fuzzy, like an AM station that is not tuned on center. I have done
this with other units so I am reasonably sure my source is ok. I got
it to come clean for a very short while by doing the chopstick drum
solo on the different parts, but it relapsed and I haven't been able
to do that again. It was a cool drum solo, though.


That might suggest that there is a bad joint somewhere on the board. These
are not always visible to the naked (or sometimes inexperienced) eye. It is
also possible for a component to have an internal bad connection. I have
known this to happen with semiconductors where the internal connections are
very fragile little wires that are sonically welded to the die.

I do want to learn proper fault finding. Can you suggest any books,
particularly as relates to music amplifiers? I have been reading the
faqs for the various electronics newsgroups, including this one.

I do not really keep up with what is available from a tuition-text point of
view, so am not able to really offer any advice or opinions on that one. I'm
sure that there will be others reading this, who may be able to offer you
advice on that score. Have you tried Googling to see what comes up ?



Would an experienced engineer consider a blanket replacement,
especially if the equipment is over 25 years old? I know that
electrolytics can dry out over time, but have also been told that
there are many variables in that. From what I have read, the
replacement may not be as good as the one replaced, in many instances.
If someone brought this amp in, would you do a listening test, input
and scope, or start measuring with your DMM?


Most experienced engineers that I know, would not just start replacing
components willy-nilly, in the fond hope that it will result in a cure, at
the same time as improving performance. Apart from anything else, if you are
doing this professionally, there are commercial considerations in terms of
time spent and cost of parts. Also, as I said before, the 'shotgun' approach
potentially paves the way for causing additional problems. It really is easy
to accidentally put a cap in backwards, or get a tiny whisker of solder
across two pads, or have some old print come away from the substrate as you
desolder, and so on. If you spot the problem as you cause it, that's not too
bad. It's when you don't spot it at the time, and then wind up with a
completely new set of symptoms at best, and loss of the magic smoke at worst
.... Trust me, I've been at this for 35 years, and I don't know of any
professional service engineers who don't have 'accidents' happen to them.
Any who tell you they don't, aren't to be trusted ...!!

Again, as I said, there are places where a blanket replacement of caps is
justified, and is the preferred 'fix-method'.

On the other hand, a blanket re-sweat of all the joints in an area,
introducing a small quantity of new solder at each joint, is a valid repair
method that is much used by professionals, to fix an invisible and otherwise
un-locatable bad solder joint. With a little practice, this can be done very
quickly, but watch carefully for solder running between joints, and forming
a bridge.

If this amp turned up on my bench, I would start by putting music through it
at a reasonable level and just having a listen. Experience would likely give
a good idea of what might be going wrong. I would then inject a sine wave of
about 800Hz to 1kHz, and 'scope the input to the power amp, just to check
that what was going in was still a good sine wave, bearing in mind that most
guitar amps are designed to be able to 'go dirty' with appropriate gain
settings. Having established that the power amp was being driven with a good
sine wave, I would next connect the 'scope across the output load, and have
a good look at the shape of the wave coming out. This would then give me a
good idea of what was going wrong. For instance, a *highly* asymmetric
output waveform would likely indicate something wrong with the drive to one
of the output transistors. this would almost certainly result in a no-signal
DC offset as well. Crossover distortion would indicate a problem in the bias
network for the output transistors. A partially asymmetric waveform would
likely be down to some kind of problem with the bias of the driver stage, or
possibly the pre-driver (assuming that the supply rails are about even). A
badly clipped signal may well indicate something going wrong back at the
front end of the power amp. Possibly a problem with a feedback network.
These are all just generalisations, you understand. It's hard to teach this.
Over the years, you just develop a 'feel' for it.

Having got an idea of what was going wrong, I would then go about locating
the exact stage where the problem was, by moving the 'scope probing-point,
to locate the last place that the signal was good, and the first place that
it was bad. Once I found which stage that was, I would switch to a
voltmeter, and eyeballing the schematic, to see if anything looked wrong
with the voltages around the stage. If there are no voltages marked on the
schematic, then it comes down to a degree of experience to know what to
expect. Usually, once you have found a wrong voltage, it will be fairly
obvious what is (probably) causing it, at which point, you switch to
'resistance' on your meter, and check the suspect component(s).


Do you have any recommendations on procedures and test gear for
evaluating capacitors?


For anyone doing a lot of service work, the single most useful tool for
evaluating electrolytic capacitors, is the ESR meter. Many on here swear by
the Dick Smith unit designed by Bob Parker, who pops up on here from time to
time. I have used one just about daily for some years now, and it has paid
for itself many times over. It is available for a very reasonable price both
in kit form, and ready built. Like all ESR meters, the results that it gives
you require a degree of interpretation, but the fact that it has a digital
readout, and a table of expected values on the front panel, makes this a
little more intuitive with this meter, than with some others that I've seen.
A basic capacitance meter is useful for evaluating lower value
non-electrolytic caps. Many digital multimeters have a simple capacitance
meter built in. They don't really tell you a lot about electrolytics,
though. On many occasions, I have had electros that read near as damn it
correct for value, but have an ESR reading that's out the window and away
with the fairies ...

I think somewhere, you asked whether "OP" was "output". It can be, but also,
on usenet, it often refers to the "Original Poster" or the person who
started the thread with their question.



I thank you very much for sharing your time and experience with me.
Your observations and advice are very valuable to me. I know more than
I did before.

For whatever it is worth, I have posted this question on
alt.guitar.amps. Lord Valve has made the very same points that you
have made.


No problem. It's nice to have someone these days who's interested in getting
into the service business. Most who are still hanging in there, tend to
advise against it, unless you have masochistic tendencies ! I have never had
any direct dealings with 'Lord Valve', but have seen his replies and posts
from time to time, and he seems a suitably skilled and experienced engineer,
whose advice you could bank.

Arfa