Maker Pro
Maker Pro

troubleshooting method for micro-controller board ?

R

robb

Jan 1, 1970
0
i have a micro-controller board with 3 big 40DIP ICs, 10 or so 16
DIP support ICs. with 32v/8v/5v transformer supply that connects
to two bridge rectifiers {32v, 8v}a fuse links sit between
transformer and rectifiers, no schematic.

i believe the 32v shorted with 5V line and it blew one fuse
link. the fuse link was replaced with same and when powered all
three fuses blown.

So....where does one start to diagnose such a *big* problem ?


..
Only thing working is transformer outputs and it seems like i get
continuity hits everywhere i check
the board was built about 1985 (no SMT) it is about 14" x 5" has
3 x 40 DIP ICs and 10 or so 16 DIP support ICs, handful of
transistors many resistors,diodes etc,

ideas and help would be greatly appreciated :)
robb
 
E

Eeyore

Jan 1, 1970
0
robb said:
i have a micro-controller board with 3 big 40DIP ICs, 10 or so 16
DIP support ICs. with 32v/8v/5v transformer supply that connects
to two bridge rectifiers {32v, 8v}a fuse links sit between
transformer and rectifiers, no schematic.

i believe the 32v shorted with 5V line and it blew one fuse
link. the fuse link was replaced with same and when powered all
three fuses blown.

So....where does one start to diagnose such a *big* problem ?

.
Only thing working is transformer outputs and it seems like i get
continuity hits everywhere i check
the board was built about 1985 (no SMT) it is about 14" x 5" has
3 x 40 DIP ICs and 10 or so 16 DIP support ICs, handful of
transistors many resistors,diodes etc,

ideas and help would be greatly appreciated :)

Well .... if the 5V supply went seriously high, the chances are it's all fried.
:-(

Have you tried powering the pcb from an external stabilised 5V power supply ?

The very first thing to look for is a clock oscillator (and valid reset signal).
Get the data sheets for those 40 pin chips, identify which one is the micro and
scope the relevant pin(s).

What are the 40 pin chip numbers btw ?

Graham
 
R

robb

Jan 1, 1970
0
Eeyore said:
Well .... if the 5V supply went seriously high, the chances are it's all fried.
:-(
there is a 630 mA Busman type fuses on all the lines i hope that
saved things ? but the fact that i am getting a short detected
between 0V and 5V line has me worried maybe thats good i do not
know hopeful someone can give me an method for finding problems ?
Have you tried powering the pcb from an external stabilised 5V power supply ?

The very first thing to look for is a clock oscillator (and valid reset signal).
Get the data sheets for those 40 pin chips, identify which one is the micro and
scope the relevant pin(s).

What are the 40 pin chip numbers btw ?

one is the Phillips MAB 8051 micro,
then SRM 2017 Static RAM,
and NEC 8547E7 (D23256) ?? <- can't find data sheet

then dual full bridge driver - L298
comparator - LM393 LM339 etc
some nor buffers and and and invertor pkgs

thanks again graham i really appreciate some help,
i can handle small problems but potential catastrophics like this
is a bit over

thanks
robb
 
A

Andrew Erickson

Jan 1, 1970
0
robb said:
i have a micro-controller board with 3 big 40DIP ICs, 10 or so 16
DIP support ICs. with 32v/8v/5v transformer supply that connects
to two bridge rectifiers {32v, 8v}a fuse links sit between
transformer and rectifiers, no schematic.

i believe the 32v shorted with 5V line and it blew one fuse
link. the fuse link was replaced with same and when powered all
three fuses blown.

So....where does one start to diagnose such a *big* problem ?


.
Only thing working is transformer outputs and it seems like i get
continuity hits everywhere i check
the board was built about 1985 (no SMT) it is about 14" x 5" has
3 x 40 DIP ICs and 10 or so 16 DIP support ICs, handful of
transistors many resistors,diodes etc,

ideas and help would be greatly appreciated :)
robb

Clearly there is one or more shorts somewhere on the board. Your
mission therefore is to determine what is shorted, and where it is.

If the ICs are socketed, I'd probably pull them (making careful note of
what was where, if it's not clearly labeled on the PCB). This avoids
any (additional) damage to parts that may be quite hard to replace.
From there, checking components in the on-board power supply section
might be reasonable, as would checking the resistance between
permutations of the various power supply rails and ground for shorts
(preferably with something like an ESR meter that won't damage
semiconductors, especially if the ICs can't be removed.) Shorts between
rails could be caused by things like failed de-spiking capacitors. If
everything looks OK with ICs removed, maybe power up the board without
them and see what happens; if the power supply is stable and current
draw reasonable, the problem would seem to be with one of the removed
ICs, or with a circuit driven by one of them.

(It's not a bad idea to figure out what the output interface circuitry
looks like that's driven by the ICs before powering it up without them;
maybe you'd be wise to tie some of the outputs to a fixed level rather
than letting them float.)

What's the purpose of this board? What sort of board-level I/Os does it
have? Based on the atypical power supply voltages, I'm guessing it must
drive or interface with some external devices; maybe those output
circuits are damaged, or the things driven by them or the wiring to them.

Just some thoughts; hopefully they're helpful.
 
J

Jamie

Jan 1, 1970
0
robb said:
it's all fried.


there is a 630 mA Busman type fuses on all the lines i hope that
saved things ? but the fact that i am getting a short detected
between 0V and 5V line has me worried maybe thats good i do not
know hopeful someone can give me an method for finding problems ?



power supply ?


valid reset signal).


is the micro and



one is the Phillips MAB 8051 micro,
then SRM 2017 Static RAM,
and NEC 8547E7 (D23256) ?? <- can't find data sheet

then dual full bridge driver - L298
comparator - LM393 LM339 etc
some nor buffers and and and invertor pkgs

thanks again graham i really appreciate some help,
i can handle small problems but potential catastrophics like this
is a bit over

thanks
robb
I use a low ohm meter (Kelvin), it's very good at
tracking down the nearest point of the short on the
board.
You may want to look at capacitors. Especially
tantalums, they short very nicely many times leaving
nothing but ash.
 
D

David L. Jones

Jan 1, 1970
0
i have a micro-controller board with 3 big 40DIP ICs, 10 or so 16
DIP support ICs. with 32v/8v/5v transformer supply that connects
to two bridge rectifiers {32v, 8v}a fuse links sit between
transformer and rectifiers, no schematic.

i believe the 32v shorted with 5V line and it blew one fuse
link. the fuse link was replaced with same and when powered all
three fuses blown.

So....where does one start to diagnose such a *big* problem ?

.
Only thing working is transformer outputs and it seems like i get
continuity hits everywhere i check
the board was built about 1985 (no SMT) it is about 14" x 5" has
3 x 40 DIP ICs and 10 or so 16 DIP support ICs, handful of
transistors many resistors,diodes etc,

ideas and help would be greatly appreciated :)
robb


First thing you should do is check for shorts across the various power
rails. If a short is found, using a sufficient resolution low-ohms
meter might let you located the offending component. Sometimes the
bypass caps can go short on order gear like this, replace them all if
needed.

Dave.
 
J

John Tserkezis

Jan 1, 1970
0
Jamie said:
You may want to look at capacitors. Especially
tantalums, they short very nicely many times leaving
nothing but ash.

Or worse. I find many times they go short with no visible evidence at all.

I was pushing the boss for a thermal imaging camera for this very purpose,
just apply power till things get warm enough in the right spots to tell what
is shorted.

I thought I had put forward a very good argument for saving time during
repairs that are relatively rare where we were, but he still sadly didn't go
for it... Oh well, I tried.
 
J

James Sweet

Jan 1, 1970
0
I use a low ohm meter (Kelvin), it's very good at
tracking down the nearest point of the short on the
board.
You may want to look at capacitors. Especially
tantalums, they short very nicely many times leaving
nothing but ash.


Sometimes it's possible to isolate sections of the circuit fairly easily, a
jumper link that can be removed, or in more drastic situations, a trace can
be cut. It doesn't take too many cycles of that to narrow things down.

Keep in mind that 5V TTL boards can pull quite a bit of juice, on larger
boards it can appear to be a nearly dead short when using an ohm meter
between Vcc and Gnd.
 
J

James Sweet

Jan 1, 1970
0
First thing you should do is check for shorts across the various power
rails. If a short is found, using a sufficient resolution low-ohms
meter might let you located the offending component. Sometimes the
bypass caps can go short on order gear like this, replace them all if
needed.

Dave.

Second that on the caps, I recently repaired a Volkswagen ECU that had a
shorted electrolytic capacitor, I think that was the first time I'd seen a
lytic shorted, regulator was getting hot which clued me in. I've found
shorted tantalums in several things.
 
E

Eeyore

Jan 1, 1970
0
robb said:
there is a 630 mA Busman type fuses on all the lines i hope that
saved things ?

I'm afraid fuses are merely for preventing fires, not preventing damage (except
for the case of 'extreme' fuse types like the ultra-fast types).

but the fact that i am getting a short detected
between 0V and 5V line has me worried maybe thats good i do not
know hopeful someone can give me an method for finding problems ?

No, I'm afraid that's actually bad news. I assume therefore it literally 'wont
do anything' at present and any power supply connected to it goes into current
limit ?

one is the Phillips MAB 8051 micro,

OK. That's a real easy one to check AND replace (at least with a modern CMOS
version).
Data sheet here.
http://www.datasheets.org.uk/search.php?t=0&q=mab8051&manystr=&sub.x=0&sub.y=0

If it was working I'd check for activity (pulses/wvaeforms) on the Xtal1/2 pins,
ALE and PSEN first.

then SRM 2017 Static RAM,

OK. Not sure how easy that would be to replace.
Data here for the pinouts..
http://www.datasheets.org.uk/search.php?t=0&q=srm2017&manystr=&sub.x=0&sub.y=0

and NEC 8547E7 (D23256) ?? <- can't find data sheet

I expect that's a 256k bit EPROM. Does it have a glass window ?

Important note. This is the most important part on the board. If it's fried so
are you. A dead EPROM will mean that the whole machine can't be repaired (unless
you can get the source file). Treat it with utmost respect.

then dual full bridge driver - L298
comparator - LM393 LM339 etc
some nor buffers and and and invertor pkgs

These are 'commodity' parts that are easily replaced. Not a problem here. First
you need to get the 'big' digital parts functioning.

General tip. Don't apply power to any high power I/O switching chips like the
L298 until the digital stuff is back to normal again.

Here are some ideas. You say you have a short across the 5V logic supply.
Firstly you need to clear that and find the cause. Are the chips socketed ? If
so it's easy to remove them and check for individually shorted chips. Check
across the power pins. Normally these are opposing corner pins like pin 20 and
40 on the 8051 for example ( or 7/14 8/16 and 10/20 on 74 series chips) but best
to get data sheets for anything unfamiliar. If you're lucky you may find one (or
two or threee) failed 'sacrificially' - saving the others but you need to find
the failed chips and get replacements for them.

If they aren't socketed now is the time to do this ! Removing ICs from PCBs
without damage is a skilled task so I would sacrifice anything that's easily
replaceable by cutting the pins at the IC body with fine cutters and removing
those pins one by one from the PCB. That way there's very little chance of doing
much damage. Be careful to fit sockets that firmly grip the leads especially if
this is some industrial controller used in areas where there's significant
vibration. That means DON'T use turned pin sockets, the pins will 'walk out' of
them.

How about you give is a complete list of the ICs on this pcb, plus make a scan
of it and post it where we can see it ? That'll help enormously.



Graham
 
R

robb

Jan 1, 1970
0
Jamie said:
I use a low ohm meter (Kelvin), it's very good at
tracking down the nearest point of the short on the
board.
You may want to look at capacitors. Especially
tantalums, they short very nicely many times leaving
nothing but ash.

Currently the only diagnostic tools at my immediate disposal are

Fluke 177 DMM, a Tek 2236 oscope and a usb/PC based all purpose
logic anal/spectrum anal/oscope

i am hopeing these will be sufficient to solve / track this
fault(s)
if low ohm meter is not terribly expensive then i'll try to get
one if it helps more than my current tools
 
J

John Tserkezis

Jan 1, 1970
0
robb said:
Currently the only diagnostic tools at my immediate disposal are

Fluke 177 DMM, a Tek 2236 oscope and a usb/PC based all purpose
logic anal/spectrum anal/oscope

i am hopeing these will be sufficient to solve / track this
fault(s)
if low ohm meter is not terribly expensive then i'll try to get
one if it helps more than my current tools

You're missing a hammer.

That might be a lot funnier if we weren't talking about 32v on a 5v line,
but at this stage, a hammer certainly seems a viable tool.

Last time I had nearly 12v on a 5v line, (most of the board was 5v logic) I
didn't waste time diagnosing. It was faster to change all the silicon and
worry about shorted tantalums later.
Checking the semis afterwards, I found nearly half of them were stuffed, and
the other half I wouldn't have trusted anyway.

That was at nearly 12 volts. Not 32.

In your shoes, (depending on the scale of the baord and availability of
replacement boards) I'd either outright replace it, or, if this is the only
one left in existence I'd turf the silicon, and hope the EPROM or flash or
wherever the code is stored still works.
 
R

robb

Jan 1, 1970
0
[trimmed lots of helpfull advice]

responses to helpfull questions in separate post
How about you give is a complete list of the ICs on this pcb, plus make a scan
of it and post it where we can see it ? That'll help enormously.

hi Graham thanks again for the help,

i posted pics of the board on
(alt.binaries.electronics.schematics)

there was no obvious visual evidence of any smoked parts that i
could see , only the middle busman fuse was blown after the short

and that seemed to be related to power going to the L298

thanks for all the help,
robb
 
T

Tom2000

Jan 1, 1970
0
there was no obvious visual evidence of any smoked parts that i
could see , only the middle busman fuse was blown after the short

and that seemed to be related to power going to the L298

thanks for all the help,
robb

Tough problem with all the ICs soldered in.

Chances are that the 32 volt supply is only used for the L298s, as you
suggest. It also appears that you can unsolder these without damaging
any other parts on the board. (Even if you have to cut them out,
they're still available and easy to replace.)

If your short doesn't disappear after pulling the 298s, try
unsoldering any of the caps near those chips.

Leave unsoldering any ICs as your very last resort.

I'm now bowing out of this discussion and leaving it to the experts.

Good luck, Robb.

Tom
 
B

BobG

Jan 1, 1970
0
Look for date codes. If its only a couple years old, some smart EE
might have actually put Transient Voltage Suppressors in some
important places to guard against incompetent operators. Can anyone
rattle off three TVS brand names from memory he can look for?
 
M

Michael Kennedy

Jan 1, 1970
0
Eeyore said:
Funny. That's where I'd START !

Graham

He probably means to avoid damaging them by poor soldering skills, but who
knows.
 
R

robb

Jan 1, 1970
0
John Tserkezis said:
You're missing a hammer.

well i purposefully left the hammers out because i did not
realize their usefull ness in this situation.

i actually have quite array of hammers 5 oz. to 5 lb rubber,
brass, nylon etc

i just hope the time comes that i do not need to use them.... on
the man or the machine
That might be a lot funnier if we weren't talking about 32v on a 5v line,
but at this stage, a hammer certainly seems a viable tool.
no it is still funny, in nervous and tearful kind of laugh
Last time I had nearly 12v on a 5v line, (most of the board was 5v logic) I
didn't waste time diagnosing. It was faster to change all the silicon and
worry about shorted tantalums later.

uhggg what a project my small project has become
Checking the semis afterwards, I found nearly half of them were stuffed, and
the other half I wouldn't have trusted anyway.

That was at nearly 12 volts. Not 32.

In your shoes, (depending on the scale of the baord and availability of
replacement boards) I'd either outright replace it, or, if this is the only
one left in existence I'd turf the silicon, and hope the EPROM or flash or
wherever the code is stored still works.

i really hope i actually make it to those problems considering
what i have read so far.

thanks for the help and ideas,
robb
 
R

robb

Jan 1, 1970
0
Tom2000 said:
Tough problem with all the ICs soldered in.

Chances are that the 32 volt supply is only used for the L298s, as you
suggest. It also appears that you can unsolder these without damaging
any other parts on the board. (Even if you have to cut them out,
they're still available and easy to replace.)

If your short doesn't disappear after pulling the 298s, try
unsoldering any of the caps near those chips.

Leave unsoldering any ICs as your very last resort.

I'm now bowing out of this discussion and leaving it to the experts.

Thanks Tom,
i really appreciate any comments and help

i unsoldered one of the L298 and it has no shorts so now the next
then i will do a trace of the 5V lines until i can learn
something usefull about the interconnections.

i hate working blind and re-tracing things when i could just look
on a schemat.

so thank you ,
robb
 
E

Eeyore

Jan 1, 1970
0
robb said:
Thanks Tom,
i really appreciate any comments and help

i unsoldered one of the L298 and it has no shorts so now the next
then i will do a trace of the 5V lines until i can learn
something usefull about the interconnections.

i hate working blind and re-tracing things when i could just look
on a schemat.

I suggest you remove the 5V voltage regulator.. That'll remove the connection
from the power supply to the digital section and NOW check for a short across
the 5V and ground.

The regulator chip may have an internal short you see and it is at least easy to
remove from the pcb.

Graham
 
Top