Current Tracer - for findinding shorts in a PCB
Scrim
valuable multilayer pcb with a short between a power supply line and earth I
can't find. The basic idea is to inject a string of pulses through shorted
circuit and use a small solenoid type sensor to follow the current path
until the short is reached.
Thanks,
Alan
Arfa Daily
"Scrim" <nos...@nospam.nospam> wrote in message
news:1woun.157498$1j3....@newsfe10.ams2...
Sometimes, a low ohms meter such as the Bob Parker ESR meter, can be a
valuable aid to finding PCB shorts. Also, allowing a limited non-damaging
current to flow round the PCB and through the short, can cause a rise in
track temperatures, that can be seen in the dark, with a reasonably
sensitive CCD camera.
Arfa
tm
"Arfa Daily" <arfa....@ntlworld.com> wrote in message
news:rQoun.195146$ti6.1...@newsfe24.ams2...
Stick the board in the freezer. When you remove it and apply current limited
power
to the suspected circuit, the frost will melt showing just where the short
is.
stra...@yahoo.com
Has it ever worked or is this a failure? Are you thinking incomplete
etch or shorted bypass cap?
G²
Doug White
Spray freeze is faster, or if you have a can of freon "duster" around,
turn it upside down (it's the same stuff). Hose down the suspected area,
turn on a current limited supply across the short, and Voila!
Doug White
Scrim
news:442f6f03-dad3-49d1...@q23g2000yqd.googlegroups.com...
G�
Thanks for the ideas. This is a high precision board that used to work.
Ideally I'd like to have a thermal camera, so I like the idea of trying a
digital camera in a dark room. Does it have to be a CCD? I'm not sure what
my web cam is (cmos?), but at least the infrared filter is easy to remove
and it doesn't have autofocus, so I can set it right with the lights on
first.
It's a multilayer board so tracing with an ohm meter seems like a bit of a
non-starter as I have no idea where the tracks go, and because of it's
precision I don't want to cover it in frost, although that's a great idea
for another time.
I've already repaired an area of board under an electrolytic that leaked.
The board was carbonised all the way through, but I think I've fixed that.
I've looked at all the tantalum capacitors I can see and they all seem ok.
Alan
tm
"Scrim" <nos...@nospam.nospam> wrote in message
Only a thermal imager will do that and they are $$$$$$$ not cheap.
Maybe you could rent one?
It would be nice if you had the board layout. A good DVM on the mV
range could find the short. Or the ESR meter suggested by someone else would
work.
Wild_Bill
be very helpful in trying to find a low resistance path on a board without
an schematic.
If you're lucky, and/or patient, you might find one cheaply on eBay or
elsewhere.
Toneohm is one brand name for shorts locators, and I bought one of their
older models years ago for about $20 on eBay.
The circuit-tracing types of shorts locators generally have a digital
display, but the helpful feature is the tone ouput, which allows the
operator to test circuit locations without watching the display, as the tone
changes as the test points get closer to the short.
The models I've seen use 4-wire Kelvin probe sets, which allows the circuit
tester to ignore circuit component resistance values.
--
Cheers,
WB
.............
"Scrim" <nos...@nospam.nospam> wrote in message
news:1woun.157498$1j3....@newsfe10.ams2...
stra...@yahoo.com
> <stratu...@yahoo.com> wrote in message
I've replaced lots of physically leaking 'lytics and where there is
one, there are usually more. If you don't initially see the 'puddle',
look for brown 'mist' around 'lytics. You might try powering it up and
see if you can read the Voltage variations around the board. You might
find a single cap that decided to be a wire. Tantalums are known to do
this and I replaced one (32 years old) just last week. Monolythic
ceramics fail this way sometimes too. Once in 40 years of doing this I
had a 'tytic that shorted out.
G²
Fester Bestertester
If it were me, I'd want to fix it via any means possible and worry about
calibration later.
Robert Macy
To find such a short magnetically, you should use a "point" sensor
[see below], not a coil. You really need to get in close to the path.
Don't go too high, else the GND planes will start to attenuate. The
1kHz to 2 kHz range is fine. Pulsing at a high enough rate will move
you out of the low 1/f noise range, and these sensors are much smaller
than any coil you can put in there, short of using a SQUID.
These people make such a sensor and are knowledgeable:
NVE Analog Sensor Catalog, NVE Corporation, 11409 Valley View Rd, Eden
Prairie, MN 55344
http://www.nve.com.
For the pusling supply, you might even get away with using something
as simple as a function generator. When shorted, like clip lead to
clip lead that constructs a 1 inch diameter they can supply around 100
mA, and I've measured that at distances of over 3 feet away.
Depending on how skilled you are [and determined to make this cheaply
on your own] you could do this using your computer's soundcard and
coils.
Jim Yanik
news:0001HW.C7E0B273...@news.eternal-september.org:
hey,you can frost it,then bake it dry in an oven.
--
Jim Yanik
jyanik
at
localnet
dot com
Baron
> On Apr 5, 9:49 am, "Scrim" <nos...@nospam.nospam> wrote:
>> Is there a DIY circuit out there for a good current tracer? I have a
>> valuable multilayer pcb with a short between a power supply line and
>> earth I can't find. The basic idea is to inject a string of pulses
>> through shorted circuit and use a small solenoid type sensor to
>> follow the current path until the short is reached.
>>
>> Thanks,
>>
>> Alan
>
> To find such a short magnetically, you should use a "point" sensor
> [see below], not a coil. You really need to get in close to the path.
> Don't go too high, else the GND planes will start to attenuate. The
> 1kHz to 2 kHz range is fine. Pulsing at a high enough rate will move
> you out of the low 1/f noise range, and these sensors are much smaller
> than any coil you can put in there, short of using a SQUID.
A scrap HDD RW head works well.
> These people make such a sensor and are knowledgeable:
> NVE Analog Sensor Catalog, NVE Corporation, 11409 Valley View Rd, Eden
> Prairie, MN 55344
> http://www.nve.com.
>
> For the pusling supply, you might even get away with using something
> as simple as a function generator. When shorted, like clip lead to
> clip lead that constructs a 1 inch diameter they can supply around 100
> mA, and I've measured that at distances of over 3 feet away.
>
> Depending on how skilled you are [and determined to make this cheaply
> on your own] you could do this using your computer's soundcard and
> coils.
--
Best Regards:
Baron.
whit3rd
[about finding a short in a printed wiring board]
> Stick the board in the freezer. When you remove it and apply current limited
> power
> to the suspected circuit, the frost will melt showing just where the short
> is.
Great idea! Any old AC wallwart is a good current-limited power
source,
I've got a bin of those ... around here ... somewhere...
Jeff Liebermann
wrote:
I use a sheet of liquid crystal material:
<http://scientificsonline.com/product.asp_Q_pn_E_3072375>
Make sure the board starts cold or at least at room temperature. The
reason I only have one sheet is that all my friends (and their kids)
have pillaged the rest of my collection. It doesn't work on the
component side of the board, but works quite well at locating hot
spots on the circuit side. Otherwise, an IR thermometer works well.
Try to get the 12:1 optics, not the 8:1.
--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558
# http://802.11junk.com je...@cruzio.com
# http://www.LearnByDestroying.com AE6KS
mike
these situations.
Another way is to stuff current in the power leads and use a high-resolution
voltmeter to find the point of lowest voltage.
A tek 576 curve tracer makes a 4-terminal measurement
and you can see it on the scope screen. And when you get close
you can get lots of amps at low voltage to burn out the short.
High-resolution time-domain-reflectometer can be used, but
you have to probe between caps...and be extremely careful you
don't blow the TDR on residual board volts.
I've used a thermal imager for this. That's the best way if you
can get access to one. Shorts between ground plane areas
are very difficult to pinpoint without thermal sensing.
rush14
A lot of good suggestions and I have one other. In our factory the
most useful tool we had for power/ground shorts was the current
limited power supply and a sheet of temperature sensitive liquid
crystal, same stuff used in the mood ring craze of the 70's I believe.
At the time Edmund Scientific was our source for the stuff.
Set the current limit to a "safe" level and lay the LC sheet on top of
the component side of the board and watch for a color change in the LC
sheet.
We had several methods for tracing shorts, including the Tone Ohm
tracer but the LC sheets were easy to use and worked best for power/
ground shorts.
Good Luck,
Rush
Scrim
a 7 1/2 digit multimeter so I'd prefer not to do anything that might upset
it if I can avoid it, but if all else fails it will go in the freezer! I
look forward to thermal cameras becoming cheap enough to own, but in the
mean time I'm going to try a combination of the other ideas presented here
until solve it.
Thanks for the help,
Alan
DaveC
> <http://scientificsonline.com/product.asp_Q_pn_E_3072375>
> Make sure the board starts cold or at least at room temperature.
Great idea, Jeff.
Which temperature product would you recommend for this use?
Dave
Jeff Liebermann
Note that the above web page says "Locate Electrical Shorts on Circuit
Boards" near the top of the page.
The last remaining one I have hidden from the visiting kids is marked
25-30C. It's about right for testing PCB heating. If unavailable,
the next lower range might be better. Note that you're not going to
see much linear range in temperature indication. There is a rather
abrupt transition at some temperature in the range. Areas of the
board that are hot are bright blue. Most everything else is black. In
between is a smear of red and green. If there's an air gap between
the board and the liquid crystal sheet, the affected area grows
rapidly larger and is difficult to pinpoint. At that point, I would
suggest using a IR gun type thermometer.
<http://www.teachersource.com/Chemistry/LiquidCrystal/LiquidCrystalSheets4x4inch.aspx>
<http://www.teachersource.com/Chemistry/LiquidCrystal/LiquidCrystalSheets12x12inch.aspx>
There are other sources for less $$$.
N_Cook
So as it is a populated board, how do you know it is a in-board short and
not a shorted VTS , if SM can look very inniocent, or something similar?
--
Diverse Devices, Southampton, England
electronic hints and repair briefs , schematics/manuals list on
http://diverse.4mg.com/index.htm
Scrim
"N_Cook" <div...@tcp.co.uk> wrote in message
news:hpnj52$ego$1...@news.eternal-september.org...
In fact I found the shorted component was a surface mount diode connected
directly across the capacitor on the reverse of the board, which I've
replaced. Now to connect it all up again and see if everything's working
again...
Alan
DaveC
> directly across the capacitor on the reverse of the board, which I've
> replaced. Now to connect it all up again and see if everything's working
> again...
>
> Alan
And not one word about how you went about identifying the diode?
Dave
N_Cook
news:0001HW.C7E55460...@news.eternal-september.org...
So a TVS, but uni or bipolar ? and what caused the high A-t transient to
knock it out ? or it will likely reoccur
Scrim
Good point, sorry! I used an old HP 546A Logic Pulser and an HP 547A Current
Tracer. Designed for use with logic circuits, but perfect for this job. You
can pick them up quite cheaply on eBay sometimes. Superb bit of kit, except
the labels - don't try to peel the plastic layer off. It looks like that
protective peel off plastic, but it's not!
Alan
Scrim
Alan