Good bench PSU for my specific application? Blowing up shorts.

[vze1lryy]

Hi,

I am looking for a PSU with one primary priority - I want it to be able to stay on and drive a 0.5 ohm load, or even a 0.01 ohm load. The rest, I do not care about. Right now I am looking at the Agilent 6542A used.

I have read many posts on this forum about bench PSUs, and many seem related to ripple voltage, quality of output, and current limiting capability. I have one primary usage for mine - blowing out shorted components on complex PCBs. My priority is not in line with many of the others asking, which is why I am posting! My apologies if this has been answered before but I could not find any info.

I work on PCBs that have over 400 components on one single bus, in a very small area. If one is shorted to ground with 0 ohms on 3.4v bus, I take my current "bench PSU", put 1.7 or 3v on that bus, and attach the ground of it to the board. Whichever component is shorting 3.4v bus to ground gets smoked instantly, I can see it, and I can replace it.

The "bench PSU" I have been using is... a lenovo. Yes. http://www.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=8748692&CatId=2627 Something like this. The 3.3v and 5v supplies on this machine can get directly shorted to ground without dying for a minute, which is long enough to find a short. The CPU VCORE can be shorted to ground almost infinitely. CPUs usually have a 0.5 ohm load if even that, so the CPU vcore power supply is perfect for this. Maybe it is because the PSU in this machine is amazing, or maybe it is too cheap to have any short protection. For whatever reason, it works better than any other power supply I have tried to date that either smokes itself or turns off instantly trying to drive these loads.

I went out for lunch and accidentally left this on for fifteen minutes and it finally died. I was going to use this as an opportunity to get a more "professional" setup, but I want to make sure whatever supply I buy will actually work. I have tried a few other computers for this, and ATX PSUs, and they just turn off instantly if they see a 1 ohm short to ground.

Can I get something between $100-$200 that can go between 1-12vdc, 5-10 amps, that can handle a 0 or 1 ohm short to ground for short periods of time? If not I'll just be buying another lenovo since I know it works, but i feel god awful silly doing it! I do not care about ripple, I do not care about accuracy of power output, I singularly care about durability when being abused driving 1 ohm loads. This will never be used for projects, it is solely for blowing things up in a controlled manner for troubleshooting purposes.

Thank you!

[Monkeh]

Just about any current-limited supply will be able to deliver its rated current into a load near enough continuously.

[rx8pilot]

Interesting method of troubleshooting. My experience with blown up components that were shorted is trace and pcb damage along with the broken component. you are also assuming the parts are failing to a shorted condition where many failures are open and will not flow any current. How will you find those?

[tautech]

Welcome to the forum vze1lryy
The addition of your country flag in your profile will help with local recommendations.
A SMPS may be your cheapest option.
Have you considered adding extra capacitance to the PSU output for a larger current spike?

[AG6QR]

Any bench supply with adjustable current limiting will happily drive a short, but it will do so by limiting the current.  It will dissipate zero power (at least theoretically, slightly more in reality) as it drives its rated current through a short circuit.  The voltage will be very low.

It sounds like you want something that will deliver a high enough current through a short in order to try and blow things up.

Maybe if you just wired up a really big capacitor bank in parallel with the output of any old adjustable voltage supply?  The capacitors will happily deliver lots of current into a short, possibly with rather spectacular results.

I won't pass judgment on the method of troubleshooting, but it's probably not one I would choose.

To find shorts, my preferred technique is to pass a constant current through the device, and use a millivolt meter to trace where the current is going.  Any bench supply with current limiting can be used for this technique.

[Seekonk]

I bought a Xantrex XHR 7.5V 80A just for current purposes like that, under 20V doesn't have any other purpose other than high current tests on my bench.   I use the mv method of tracking on the chips.   I don't like blowing up a chip before I have a chance to read what is written on top!

[retiredcaps]

To find shorts, my preferred technique is to pass a constant current through the device, and use a millivolt meter to trace where the current is going.

Would your typical mV setting on a hand held multimeter be enough resolution or do you need something more?

[jlmoon]

I bought a Xantrex XHR 7.5V 80A just for current purposes like that, under 20V doesn't have any other purpose other than high current tests on my bench.   I use the mv method of tracking on the chips.   I don't like blowing up a chip before I have a chance to read what is written on top!

oh those poor 3.3v parts won't like you

[electronic_eel]

I have a beefy lab psu which goes up to 50 amps and it works well to blow up shorted parts or traces. But if the part is really well shorted, often not the part itself blows up, but the traces towards it. So it may be fun, but is not a good way to find shorts for repairing something.

I also have good multimeters and sometimes use the milliohm-approach, but making good enough kelvin contact takes time. Also you just get a rough direction and you still have to individually test a lot of parts and points.

The best method in my opinion is using a HP 547A current tracer together with the HP 546A logic pulser. With these you can easily detect where the current flows and just follow that till you found the short. Matter of seconds and no destruction. Both aren't sold new anymore, but you can often find them on ebay.

[nali]

I won't pass judgment on the method of troubleshooting, but it's probably not one I would choose.

To find shorts, my preferred technique is to pass a constant current through the device, and use a millivolt meter to trace where the current is going.  Any bench supply with current limiting can be used for this technique.

It used to be common practice where I worked back in the 70s in production testing - the shorts were then caused by under-etching and often so thin they were  naked to the eye so got through visual inspection. IIRC we used a huge 47,000uF capacitor and most of the time the offending short was so slight it wouldn't even spark.

Agree with the constant current / millivoltmeter technique which I've used to find some shorted tantalums before now. A DMM on its lowest range will just about do.

If it's a fraction of an ohm upwards then the offending part will warm up anyway so just let your fingertips do the sensing...

[AG6QR]

To find shorts, my preferred technique is to pass a constant current through the device, and use a millivolt meter to trace where the current is going.

Would your typical mV setting on a hand held multimeter be enough resolution or do you need something more?

For me, the mV range on a hand held meter has been enough.  Of course it depends on the amount of current you're pumping through the circuit, and the resistance of the traces.  If you need more resolution, you can either get a better meter or crank up the current.  Conveniently, when the voltage drop is so small it's hard to see, that's a sign that the traces have low resistance, so chances are they probably won't be damaged by a bit more current. 

[Howardlong]

I agree with rx8pilot, it's not just the part, it's the PCB is usually buggered in my experience after a short, and in my case that's with bus powered USB devices and so are 1/2 amp limited. It happens in about 1 in 2000 devices, we just throw them in the junk pile, no point in trying to repair a six layer board.

[retiredcaps]

I found a link that helped me understand this a bit more.

http://areksnotes.blogspot.ca/2012/09/how-to-find-short-circuit-on-your-pcb.html

[MadTux]

So you want badass PSU that can also be used as electrode welder? I also like badass PSUs, so here's a list:

Oldskool triac preregulated:
6260B => 0-10V @ up to 100A
(the higher voltage/less Amps modells sometimes have problems with regulation with short circuit applied and stay @ 0V/0A

New SMPS:
HP 6671A => 0-14V @ up to 150A
HP 6681A => 0-8V @ up to  580A
HP 6682A => 0-21V @ up to  240A
HP 6683A =>0-32V @ up to  150A

Plz make video of you troubleshooting boards and releasing magic smoke

[vze1lryy]

Thank you for all of these excellent replies! There are times during which I know where the short is based on common faults - connectors and common issues with high side FETs for certain buck converters. A lot of the time I use my brain and knowledge of the circuit, but for others, there are 100 or 200 tiny ceramic caps scattered all across the motherboard and sometimes it is just some random nonsense that will take me forever to find manually. So this method has proven helpful. I start with 1.7 volt and touch the board to my face, and find it this way. No IR cameras or x-ray machines. I only go up tot he voltage the bus is rated for if 1.7 volt will not allow me to find it.

If any decent bench power supply will allow me to drive a 1 or 0.5 ohm load, then I will go with the Agilent 6542A. It's only a little more used than the cheapo circuitworks ones and is an excellent, trusted brand it seems from everything I read.

In my case I've had messed PCBs maybe 1 in every 100, and when the PCB is destroyed, with multilayer PCBs and this many components it is not worth trying to fix.

Thank you again for all the information!

[TunerSandwich]

buy a cheap 20 amp tripp lite 12V supply, and a beefy soak.....dissipate the unneeded power into the soak....since you don't care about efficiency or quality of power.....

cheap....effective....huge knob....mayhem etc etc etc 

[Howardlong]

In my case I've had messed PCBs maybe 1 in every 100, and when the PCB is destroyed, with multilayer PCBs and this many components it is not worth trying to fix.

I would have a discussion with your board house, I don't know how complex your board is, but 1 in 100 isn't very good, and should have been rejected (hopefully) by the board house at the electrical test stage before shipping. It's not uncommon to have board failures, but the board house should be able to filter almost all of the bad ones. We get a few partial panels, where, say, nine out of ten work, and we can if needs be use these too but mask out the bad board(s) from the pick and place machine.

[vze1lryy]

My company repairs damaged Apple laptop motherboards, mostly by liquid. I can't yell at a board house, only the customers, but I can't yell at them for spilling things on them because if they stopped spilling I would stop getting paid.

I got the agilent 6542a. Huge disappointment. The second you try to use it as a short tester in this way, it switches to constant current mode, limits voltage output to 0.6v and sends out absolutely nothing. It is more useless than a $10 PC power supply for this application. Live and learn, $180 unless I use it to power a desk fan in the summer maybe. I'm buying the Lenovo for $80 I should have in the first place, what works works!

[AndyC_772]

Sounds like what you really need is a current limited power supply and a thermal camera. Pass a reasonable current through the fault, and look to see which components start getting warm.

[AG6QR]

I got the agilent 6542a. Huge disappointment. The second you try to use it as a short tester in this way, it switches to constant current mode, limits voltage output to 0.6v and sends out absolutely nothing.

That is precisely what it is designed to do, and what we said it would do.  That's considered an advantage for its application, since many people want to test things without destroying components.  I think I already mentioned that "Any bench supply with adjustable current limiting will happily drive a short, but it will do so by limiting the current.  It will dissipate zero power (at least theoretically, slightly more in reality) as it drives its rated current through a short circuit.  The voltage will be very low."

You can use a meter to find a short using this kind of supply.  Or you can get a supply without current limiting and blow things up.

[Chris Jones]

I am looking for a PSU with one primary priority - I want it to be able to stay on and drive a 0.5 ohm load, or even a 0.01 ohm load. The rest, I do not care about. ....

A D-size NiCd should do that.

If you want adjustable voltage up to 2.5V, then any old power supply in parallel with a Maxwell 3000F capacitor ($65 or so on Digi-key) will deliver up to 1900 Amps within its ratings, and about 8000 Amps if you don't worry about ratings, which I figure you won't. It will take a while to charge up before your troublemaking^h^h^h^h^h^h^hshooting, and you will need some stout cables.

[nixfu]

Hmm.. any decent power supply will be designed to automatically limit current.


If you want something without any current limit.....I recommend a large battery.