Multilayer board, track tracing?

[N_Cook]

Any tips for single ended signal injection and tracing for high resolution
tracing , floppy/HD head as searchcoil ?

 

[Arfa Daily]

Any tips for single ended signal injection and tracing for high resolution
tracing , floppy/HD head as searchcoil ?

I would have thought that any tracks which passed internally between power
and ground planes, would have struggled to radiate enough signal to be
picked up in the way that you are envisioning. Also, where there are
fine-pitch 'bunches' of tracks which run parallel to one another for any
distance, I would think that crosstalk might become a problem to this
technique. If such a board has any (internal) tracking problems, I don't
think that you are going to do a lot with it anyway. Even if it is just
generally faulty in some way, if the board is complex enough to require
multi-layering, I think you are going to be pissing in the wind attempting a
(commercially viable) repair without the benefit of at least schematics ...

Arfa

 

[David Nebenzahl]

I would second that. I doubt the time and effort spent in attemptng
this is time and effort well spent. You'd be better off going for a
nice walk in the countryside instead, after telling your customer
this is not a viable repair. Which you or he wll end up doing
anyway.

I third that. Much as I commend the spirit of trying to fix everything
yourself that is possible, some things just aren't made to be fixed.
Multilayer circuit boards seem to be one of those things.

 

[JB]

I third that. Much as I commend the spirit of trying to fix everything
yourself that is possible, some things just aren't made to be fixed.
Multilayer circuit boards seem to be one of those things.

You might be lucky and only power and ground are in the inner layers along
with a jumper or two. A bright light is about the best you can do. If it's
more than 3 or 4 layers you will need schematic in front of you.

I fourth it, that it better be worth your effort with a contract to do x
number of like units unless you are independently wealthy and like
profitless challenges. Alternatively, you could redesign the board as a
retrofit upgrade for the whole product line.

 

[N_Cook]

It was worth an enquiry. Perhaps someone had hit on a wondrous new
technique, i was not aware of.
eg > 1MHz ultrasonic source touching the exposed metal of a trace and
ultrasonic sniffer picking up the buried path. Polyester and glass damping
the propogation sideways, but free passage along the metal trace. Anyone got
a few tens of thousands to investigate ?

 

[Franc Zabkar]

Any tips for single ended signal injection and tracing for high resolution
tracing , floppy/HD head as searchcoil ?

I've repaired a lot of multilayer PCBs, but I've always had circuit
diagrams or a knowledge of the board.

What about a VCR head attached to a ballpoint pen tube? I would think
that as you rotated the pen, the max/min signal amplitudes would
indicate the direction of current flow.

You could also experiment with ribbon cable. That would give you an
indication of sensitivity to crosstalk, and the insulation thickness
could simulate a layer of PCB.

- Franc Zabkar

 

[JB]

Yep. Just yesterday I rejected a big ol' Pioneer amp with every output
transistor short on both channels, and most of the emitter Rs open. The
mackled-up box that it came in from the customer, suggested that they might
have obtained it from uPay or some such. With a sigh, I put its covers back
on, and explained to the shop which took it in, that it was going to be more
trouble than it was worth. They told me that they 'knew' this was going to
be the case, but took it in anyway ...

I then went out for that walk in the country that Gareth suggested ...

A few years back, I would have fixed it come what may. Now, it seems that my
time is more importantly used on jobs that actually make me some money.

Arfa

You have my sympathies. After replacing $100 worth of high power
transistors or hybrids, it doesn't leave much for labor. You would be
better off fixing cars or washing machines.

It has been my experience with consumer equipment, that if you can fix it,
the customer will either be poor and can't afford it, or will be genuinely
disappointed because they really wanted the newest, latest gimmick. If you
fix it, they will never return to claim it. Or in one case, brought the
whole family over and while I was distracted, one snuck out with TV and
paperwork and all bolted on queue without paying.

Always leave the unit disassembled or at least with the back off with wiring
and circuits exposed and settle the invoice before putting it back together.
Thieves don't want to get electrocuted.

 

[Franc Zabkar]

It's a thought Franc, but you would need either a damned strong signal
source (see my earlier thoughts about crosstalk in prallel track bunches),
or a very high gain and very low noise amp on the end of the video head. I
say this because from what I've seen of video heads over the years, they
have very few wire turns on them, relying for their output on the fact that
they are rotating at very high speed relative to the tape, and in extremely
tight physical contact with it.

Wouldn't it be easier in the end, to just use your DMM on its lowest ohms
range with a continuity beeper, to hook onto one end of a track where you do
know that it finds its way to the top or bottom surface, then go hunting for
the next place where it appears by initially listening for the beep, and
then checking for the near-zero ohms reading that you will be getting, if
you have found a direct connection ?

Either way, a very tedious process, and definitely not commercially viable
unless, as someone else commented, it is an expensive board for which there
is no service info available, and that someone wants to contract you to mend
a large number of.

Arfa

The 15" x 15" multilayer boards I used to repair typically had 300 TTL
chips.

ISTR that someone once suggested to place a sheet of aluminium foil
over one half of the board to test for continuity and then narrow down
the precise location with successive binary searches. Alternatively
you could make a chip-sized metal pad and probe an entire chip at a
time.

I agree that looking for a small signal may be fruitless. My first
thought was to adapt the method used in some digitizing tablets, but
these use a relatively high current pulse.

As you suggest, I do all my reverse engineering with a DMM on the low
ohms range. I might try the Al foil method next time, though.

- Franc Zabkar

 

[Bob Larter]

JB wrote:
[...]
I fourth it. I don't bother with repairs that're likely to need that
kind of time & effort. You'll invariably lose money on such jobs.

It has been my experience with consumer equipment, that if you can fix it,
the customer will either be poor and can't afford it, or will be genuinely
disappointed because they really wanted the newest, latest gimmick. If you
fix it, they will never return to claim it. Or in one case, brought the
whole family over and while I was distracted, one snuck out with TV and
paperwork and all bolted on queue without paying.

Ouch! That's a new one on me.
My own worst story was the time a pissed off customer pulled a knife
on one of our counter staff when he came to pick up his 'warranty'
repair laptop & was told that we were going to charge him for what had
turned out to be a software problem.[0] I called the cops, then went to
the counter to rescue the counter girl & try to calm him down until the
cops got there. And no, he didn't get away without paying. ;^)

Always leave the unit disassembled or at least with the back off with wiring
and circuits exposed and settle the invoice before putting it back together.
Thieves don't want to get electrocuted.

Good tip. We used to keep finished jobs on shelves about 5' behind the
service counter, so people couldn't just grab & run. That obviously
wouldn't be very practical with TVs, of course.

[0] When the guy first booked the machine in for warranty service, I'd
warned him that the problem would almost certainly be software, & that
he'd be charged if it was.

 

[Bob Larter]

Wouldn't it be easier in the end, to just use your DMM on its lowest ohms
range with a continuity beeper, to hook onto one end of a track where you do
know that it finds its way to the top or bottom surface, then go hunting for
the next place where it appears by initially listening for the beep, and
then checking for the near-zero ohms reading that you will be getting, if
you have found a direct connection ?

I've actually done that, combined with putting the PCB over a strong
light to make it easier to trace the tracks visually. It's a pain in the
arse, but it's not totally unfeasible. The one problem is that if you're
tracing a digital circuit, you're likely to get lots of beeps on power
pins you hit, due to all the ESD protection diodes in the chips.

 

[Heinz Schmitz]

It was worth an enquiry. Perhaps someone had hit on a wondrous new
technique, i was not aware of.

You might want to ask your dentist, vet or clinic to try an x-ray
picture of the board.
If he succeeds to achieve about 30 gray-steps per layer you might
be able to identify the traces with a graphic software after reading
the x-ray-pic with a scanner. Traces passing under a chip will be a
problem, but having done all others you could do these with
conductivity checks.

Regards,
H.