Finding leaky components

lesbenito

Member
Joined
Sep 29, 2014
Messages
196
Reaction score
17
Location
Maryland
I have a couple of boards that seem to be leaking voltage. I can't get them up to 5v. One only goes as high as 4.2 and the other about 4.8. In both cases the PCBs are not getting enough juice to boot.

If you have any tried and true ways of finding leaky or shorted parts I would love to hear about them. I have looked for variances in resistance and in voltages but can't seem to find the culprits.

Your assistance is greatly appreciated.


Thanks

Les
 
You may have luck with the EDS CapAnalyzer, but it would likely show all caps on the 5v line as bad and not any specific one. He also makes the LeakSeeker, but I've only had 1 success in using it out of 3 attempts. Somewhere on youtube there's a Randy Fromm video that shows how to use those old 6v lantern batteries to track down a leak that looks interesting.

There's always the old standby of pulling one leg of a component and seeing if the short goes away. First try electrolytics as I believe those fail most often, then it's a crap shoot: a bunch of smoothing caps, resistors, ICs.... Visual inspection will let you know if one of these is fried, but I did have an LS74 go bad and short which was physically damaged on the underside (not visible until pulled), so a logic probe/comparator helps with that.
 
If it's dragging your power supply rail down that much, it's likely pulling a lot of current... which will make it pretty obviously hot. I have a thermal imaging camera that I look at my boards with, but you can also just feel for warm components.

You could also hook up a more powerful power supply that can supply the current, and it'll likely become even more obviously hot, or possibly even pop... in which case you'll find it quickly. ;)

DogP
 
I have a couple of boards that seem to be leaking voltage. I can't get them up to 5v. One only goes as high as 4.2 and the other about 4.8. In both cases the PCBs are not getting enough juice to boot.

If you have any tried and true ways of finding leaky or shorted parts I would love to hear about them. I have looked for variances in resistance and in voltages but can't seem to find the culprits.

Your assistance is greatly appreciated.


Thanks

Les

What game are these in ? I have seen a bad big blue filter cap drag the +5 down.
 
You may have luck with the EDS CapAnalyzer, but it would likely show all caps on the 5v line as bad and not any specific one. He also makes the LeakSeeker, but I've only had 1 success in using it out of 3 attempts. Somewhere on youtube there's a Randy Fromm video that shows how to use those old 6v lantern batteries to track down a leak that looks interesting.



There's always the old standby of pulling one leg of a component and seeing if the short goes away. First try electrolytics as I believe those fail most often, then it's a crap shoot: a bunch of smoothing caps, resistors, ICs.... Visual inspection will let you know if one of these is fried, but I did have an LS74 go bad and short which was physically damaged on the underside (not visible until pulled), so a logic probe/comparator helps with that.



Well the funny thing is that I was able to get my hands on a leak seeker. Didn't really help. I was looking forward to it solving all my problems but unfortunately it did not. It might be a user error but I have tried 2 probe and 3 probe tests to no avail. If you have any tips on how to use it I am all ears.
 
If it's dragging your power supply rail down that much, it's likely pulling a lot of current... which will make it pretty obviously hot. I have a thermal imaging camera that I look at my boards with, but you can also just feel for warm components.

You could also hook up a more powerful power supply that can supply the current, and it'll likely become even more obviously hot, or possibly even pop... in which case you'll find it quickly. ;)

DogP



Yeah I was going to put in a more powerful power supply and see if that helped to figure it out. I'll try that next.
 
Well the funny thing is that I was able to get my hands on a leak seeker. Didn't really help. I was looking forward to it solving all my problems but unfortunately it did not. It might be a user error but I have tried 2 probe and 3 probe tests to no avail. If you have any tips on how to use it I am all ears.

Sorry, no real tips that I can offer. The first board I tried it on, a Defender CPU, I didn't have much luck and gave up quickly. I'll revisit it soon. The second board is a Battlezone CPU board. The game worked fine, but one day I left in on overnight while testing and it locked up, and I think it locked up one more time, but not recently. It still plays fine but somewhere in my testing I found the +5v line had a near short (~150 ohms). The LeakSeeker couldn't "lock on" a window. I emailed Dave Miga directly and he offered some advice but ultimately nothing that was useful, e.g. he told me it was either a bad power supply or failing electrolytic caps when 1) it was on my bench, unpowered and 2) I used the CapAnalyzer to verify all caps were good. He also mentioned bad RAM, but I do not currently have a way to test that.

For the board I got it to work on, a GORF game board, it was a last resort. It had a dead short on the +5v line. I had pulled all tantalums and smoothing caps and none were shorted. I figured I'd give the LeakSeeker a try and within a minute it pointed to a LS74. I pulled it, found it physically damaged and it tested bad with one of those ebay special chip testers. Replaced it and short was gone.
 
Regarding a "LeakSeeker" type of device... I use my LCR meter ( http://www.amazon.com/Signstek-M4070-AutoRanging-accuracy-display/dp/B00GYT5WL8 ) on L for that kind of functionality.

While it's actually meant to measure inductance, it also displays the resonant frequency of the circuit... and the closer you get to the short, the higher the frequency gets (since the inductance decreases as the trace length decreases).

DogP
 
Regarding a "LeakSeeker" type of device... I use my LCR meter ( http://www.amazon.com/Signstek-M4070-AutoRanging-accuracy-display/dp/B00GYT5WL8 ) on L for that kind of functionality.

While it's actually meant to measure inductance, it also displays the resonant frequency of the circuit... and the closer you get to the short, the higher the frequency gets (since the inductance decreases as the trace length decreases).

DogP

Woohoo! Another toy... I mean tool for Les to acquire!
 
Woohoo! Another toy... I mean tool for Les to acquire!
I hope Les finds it very helpful. ;)

BTW, I use a cheap set of pointy multimeter probes with it (these: https://www.sparkfun.com/products/12078 with the safety boot cut off so it'll plug into the banana jack). As you can imagine, it's hard to accurately probe with those alligator clips it comes with.

To give you an idea of the kind of precision you can get... if I have a shorted trace with a through-hole somewhere in the middle... with this meter and pointy probes I can probe the left and right sides of the inside of the plated through-hole and determine whether the short is on the left or right branch of the circuit. On a straight piece of wire, I can tell whether I'm getting closer or farther from the short by moving ~1mm.

Of course it gets a little harder as it becomes less of a short... with a 22 ohm "short", I only get about 5mm resolution on a piece wire (which is more realistic of real use, since most transistors/ICs don't short 0 ohms... unless you're tracking down a solder bridge).

DogP
 
I hope Les finds it very helpful. ;)



BTW, I use a cheap set of pointy multimeter probes with it (these: https://www.sparkfun.com/products/12078 with the safety boot cut off so it'll plug into the banana jack). As you can imagine, it's hard to accurately probe with those alligator clips it comes with.



To give you an idea of the kind of precision you can get... if I have a shorted trace with a through-hole somewhere in the middle... with this meter and pointy probes I can probe the left and right sides of the inside of the plated through-hole and determine whether the short is on the left or right branch of the circuit. On a straight piece of wire, I can tell whether I'm getting closer or farther from the short by moving ~1mm.



Of course it gets a little harder as it becomes less of a short... with a 22 ohm "short", I only get about 5mm resolution on a piece wire (which is more realistic of real use, since most transistors/ICs don't short 0 ohms... unless you're tracking down a solder bridge).



DogP



Thanks for the advice. I appreciate it.
 
If it's dragging your power supply rail down that much, it's likely pulling a lot of current... which will make it pretty obviously hot. I have a thermal imaging camera that I look at my boards with, but you can also just feel for warm components.

You could also hook up a more powerful power supply that can supply the current, and it'll likely become even more obviously hot, or possibly even pop... in which case you'll find it quickly. ;)

DogP



So I used my thermal camera and as you can see in the attached picture, the hottest component on the board is a 82s153 which is a field programmable logic array. There are a couple on the board and they all are about the same temp so I don't think that's it. I'll keep slogging away I guess.

b9b38b2d660e5f78db8af50e3482be78.jpg
 
Weird... are you sure your bench power supply isn't weak, or have too small of wire or a bad connection at the connectors? If the supply is good, and can handle 5V@15A (typical arcade PS), and you're dragging the rail down more than that, it's probably pulling 20A or so. Those extra 25 or so watts should be very obvious on the thermal imager.

Have you checked the voltage right at the output of the power supply? Tried swapping the PS?

DogP
 
Weird... are you sure your bench power supply isn't weak, or have too small of wire or a bad connection at the connectors? If the supply is good, and can handle 5V@15A (typical arcade PS), and you're dragging the rail down more than that, it's probably pulling 20A or so. Those extra 25 or so watts should be very obvious on the thermal imager.

Have you checked the voltage right at the output of the power supply? Tried swapping the PS?

DogP

The voltage at the PS is almost 6v. About 5.8 or so. I did change out the power supplies and got one that does 16a vs. the previous one that does 15a. No change. BTW this is a Time Pilot board. When I plug in only the top board (audio/video) my voltage is fine. As soon as I add the bottom board that's when the voltage drops to about 4.6.

ugh.
 
When you hook up the bottom board, is the voltage at the power supply still 5.8V, or does it drop to 4.6V? If it's just still 5.8V at the power supply, but 4.6V at the board, then you've definitely got a bad connection on your wiring harness, or too small of wires.

DogP
 
if you are using one of those cheap chinese jamma harnesses you probably don't have enough wire to deliver the power on. I run 4x 18 gauge wire on my 5V rail for the test bench and none of the chinese jamma harnesses have anywhere close to enough to do the job.


Large chance power delivery is your issue and not the board or the power supply.
 
Back
Top Bottom