finding a short on motherboard

I am doing a test to trying to find a short on a motherboard and I have a regulated power supply to use, using this method I will be freezing the board and adding power, My questions are what current and voltage would be safe to use? and how would I wire the positive and negative up to the motherboard to give it the power it needs ?

may be it would be easier after building something like this :
**broken link removed**

Hi,

I assume the short is in a supply node:
* limit voltage below the rated voltage
* limit current that the max. disipated power is less than 1/4 W

* connect the supply to the board

* use a millivolt meter and follow the traces. Look for the lowest voltage.

The lowest voltage reading will be close to the short.

Klaus

And do bear in mind that if this is a computer motherboard, the current consumption can be tens of Amps so the measured resistance across the supply might be very low anyway.

Krypton2035's link shows a low value resistor measurement and Klaus uses the resistance along the trace by following voltage drop. The methods are similar but not quite the same. If you are sure the PSU is safe under overload current, and you have a track layout, Klaus's method is probably quicker at finding the short. Whatever you do, NEVER apply a voltage higher than normal to force more current, if the short does blow out, everything else will follow shortly afterwards!

Brian.

I'll tell you my anecdotal experience. I was repairing a motherboard of a vintage computer (mid-80s). First I was trying to power it with a quality name-brand ATX PSU, but it kept shutting off because I guess it was sensing a short and tripping. I measured the resistance on the 5V rail on the motherboard and it was low, but I couldn't find why.
I figured since the board is not working anyway I don't have much to lose. I connected it to a no-name Chinese ATX PSU, and that one did not trip. There was smoke. There was a lot of smoke. One axial ceramic decoupling capacitor completely burned off on the motherboard.

After that the board works fine. I didn't even have to replace the one burned decoupling capacitor (there are others in parallel).

It's bad advice to give, but if you have a minor short and you're just about ready to give up, you can just let the short burn. My uncle has a thermal camera at work, so finding a short that way could be less extreme, but you'd have to sell some organs to afford one of those.

whack said:

It's bad advice to give, but if you have a minor short and you're just about ready to give up, you can just let the short burn. My uncle has a thermal camera at work, so finding a short that way could be less extreme, but you'd have to sell some organs to afford one of those.

Click to expand...

Wow, you must have some pretty used up organs. ;-)

You can get one of these for way less than the cost of a quality organ.

I was just curious about the cost of thermal cameras, as I recall a relative showing me something similar to the Seek and they aren't necessarily going to spend internal organ prices for such a camera.

Many test methods exist.

- measure millivolt or microvolt voltage rise with 1A current limited lab supply set to 1V
- Use a pulse 50 mOhm RdsOn switch to discharge a 1V big plastic low ESR cap at 1MHz and low low duty cycle from pulse gen. then sniff with probe shorted into loop for the strongest then weakest signal transition where the bridge loop current follows. or even 10kHz with 1ns rise time is easier. The current resonant pulse ripple will be easy to see using coax or twisted pair from source with CM choke if needed and shorted scope probe loop, 1" diameter and up. It only picks up nearfield noise pulses in rf range of scope and Joules energy is low.

Watch out for shorted MOBO layer vais and caps and shorted CMOS with ESD diodes burnt 0 Ohm from improper handling.

Measure actual board resistance and if <100 mOhm short consider a capacitor discharge large enough mF to fuse open the link with <+Vcc but if much less,, it may be FUBAR. Beware of ESD risks doing this.

darren30 said:

I am doing a test to trying to find a short on a motherboard and I have a regulated power supply to use, using this method I will be freezing the board and adding power, My questions are what current and voltage would be safe to use? and how would I wire the positive and negative up to the motherboard to give it the power it needs ?

Click to expand...

I would use a computer supply. that way you are not exceeding any voltage levels. use a meter to determine which voltage is shorted. some times the hot part is not the bad one.
If you find the regulator chip is hot, ohm out it's output. If that is low, then use a meter to see if it gets lower, away from the chip. we are talking 0.01 ohms the last digit on most meters. But you can find a short tracing to least resistance.

The problem with using a computer supply is they have overload protection. The very problem of a short on the board prevents them being used.

Brian.

If a 20A 3V supply with 1m Ohm ESR Caps were to be applied to a 10 Mohm short then it may fuse open with a 300A pulse before shutdown with a 900 W pulse.

betwixt said:

The problem with using a computer supply is they have overload protection. The very problem of a short on the board prevents them being used.

Brian.

Click to expand...

yes, but which power supply crow barred?

old days, you at a +12v, -12V and a 5V

Now, they probably have a 3.3V...(I haven't looked at a computer power supply for sometime)

still what ever voltage is shorted, you can drive that line with a value up to that voltage with external supply.