"How to find" / "where is" shorted-to-ground on my PCB?

Thread Starter

booboo

Joined Apr 25, 2015
168
Hi everybody
A couple days ago, I received my PCB. but I have a problem. I have shorted-to-ground problem on my PCB. errrrrrrr!:mad: such problems are sooo tough!
I'm diging whole of the net to find a way to locate the shorted point. via google video I found these:

www.youtube.com/watch?v=F-xI4ceQCco[who has such a device at home!?]
www.youtube.com/watch?v=JAOV_erLpYw[not bad but a bit complicated]
www.youtube.com/watch?v=JTrb3FYsXQM[it's not available to me]
www.youtube.com/watch?v=XhqwC3BFBug[not bad but I don't have this spray]

Via google web found these:

http://electronics.stackexchange.com/questions/16398/find-short-location-on-pcb
http://www.edaboard.com/thread31322.html

In the E.se page, Russell's answer is nice (How about it? is it good in your opinion?). my DMM is VC9805. Is it enough to use this method? and for making current, How to make it and how much? a simple and temporary way(circuit) would be appropriate I think. this is the circuit:



Can you see any shorted circuit?
I have these regulators & ICs in stock:
L7805CV, L7905CV, AMS117 3v3 and 5v, LM317T

Can I make a temporary circuit for current?
 

mcgyvr

Joined Oct 15, 2009
5,394
You are more familiar with your circuit/pcb than anyone here so its much more work for us to find the error for you..

Can't you run net-net error checking in your PCB program?
Got a schematic?
Did you compare the schematic to the PCB layout (again another error check function that's in any decent pcb program)?
Did the PCB manufacturer run any etest/flying probe?
Can we assume blue traces are 1 layer and red are another? But what about green traces at P4/P5/Q2 (I'll assume they are on the red trace layer)?
 

MrChips

Joined Oct 2, 2009
30,720
Is your PCB fully populated with all components soldered on the board?
What is S1? Is it a push button?
Double check the pinout of S1.
Remove S1 and see if the short goes away.
 

Thread Starter

booboo

Joined Apr 25, 2015
168
Thank you guys
I apologize because I forgot to say that My PCB isn't fully populated. this is the PCB:



You are more familiar with your circuit/pcb than anyone here so its much more work for us to find the error for you..

Can't you run net-net error checking in your PCB program?
Got a schematic?
Did you compare the schematic to the PCB layout (again another error check function that's in any decent pcb program)?
Did the PCB manufacturer run any etest/flying probe?
Can we assume blue traces are 1 layer and red are another? But what about green traces at P4/P5/Q2 (I'll assume they are on the red trace layer)?
1- My PCB program is Altium. when I'm seeing the PCB and put the cursor of the mouse on for example GND line, I cannot see any problem on it(short circuit) because just GND line getting bold.
2- yes. but there is no problem in schematic.
3- Not completely. a glance.
4- No, I don't think so
5- Yes. it's just a two layer PCB. the green traces is because of overlapping components' marks.
Is your PCB fully populated with all components soldered on the board?
What is S1? Is it a push button?
Double check the pinout of S1.
Remove S1 and see if the short goes away.
Yes, it's a push button. but I had the problem before mounting it too.
 

MrChips

Joined Oct 2, 2009
30,720
You need to provide more information.
Post your circuit diagram.
What is the part number on the MCU?
Show sharp high resolution photos of both side of your board.
 

JWHassler

Joined Sep 25, 2013
306
Is the short on your populated board or on the naked PCB? Big difference in the required effort.
If it's on the unpopulated PCB, you might just need a microscope and a whole day.
 

Lestraveled

Joined May 19, 2014
1,946
Several of your links describe using a current source and a volt meter. This is a very good method. If you do not have a power supply with current adjustment, then put a resistor in series with your power supply and adjust the voltage. The lower the voltage the closer you are to the short.
 

Thread Starter

booboo

Joined Apr 25, 2015
168
Is the short on your populated board or on the naked PCB? Big difference in the required effort.
If it's on the unpopulated PCB, you might just need a microscope and a whole day.
I don't know if it was on the naked-PCB because when I soldered the components, I thought there is no problem.
Several of your links describe using a current source and a volt meter. This is a very good method. If you do not have a power supply with current adjustment, then put a resistor in series with your power supply and adjust the voltage. The lower the voltage the closer you are to the short.
I have a PSU at home. could it help?
 

Lestraveled

Joined May 19, 2014
1,946
OK, so you have a computer power supply. That will work. I would run the test at around .5 amps.

- Use the 5 volt and ground outputs of your power supply.
- Put a 10 ohm 3 watt resistor in series with the 5 volts. (Or something close.)
- Connect the lines to the shorted path.
- Probe the board with your volt meter, looking for the lowest voltage. You are looking for very small voltage.
 

ErnieM

Joined Apr 24, 2011
8,377
Yes your PSU could help, and here is how:

A short is a very low resistance connection and thus hard to localize with just an ohmmeter. The voltmeter part of your DVM is much more sensitive but you need to push thru some current to see the rEsistance. So you current limit your supply and connect it across the short. With your voltmeter pins you poke to find the lowest voltage difference, which is the short. This method does need a little luck to work.

There have been times where the short could be fused open by zapping a large current thru it. Usually one has to add a largest cap it up the zap current available. Don't touch a live probe to your board, make the probe connections first, then complete the circuit away from the board. This insures you don't leave a scorch mark on the board itself.

Obviously zapping in a method of last resort.
 

dl324

Joined Mar 30, 2015
16,846
Have you visually inspected the soldering around the LQFP package?

Regarding the freeze spray; buy a can and keep it on hand. Also keep an eye out for an HP547A current tracer probe. When used in conjunction with a current source, you be able to isolate shorts more quickly (over the voltmeter or ohmmeter approaches). They're a bit expensive now, but the prices wax and wane. I picked up a spare set (of HP545A, HP546A, HP547A, and HP548A) for under $75 4-5 years ago on eBay; they're several times that currently... This seems to be the best deal on eBay.
 
If you have access to a FLIR camera like an E8 model or better....apply current limited power, perhaps a 1/2A. The 'short' will produce a heat bloom fairly quickly and you have found it.

I use this method to trace faulty components as well , such as caps with poor ESR and transistors going bad.
 

Thread Starter

booboo

Joined Apr 25, 2015
168
Thank a lot guys
OK, so you have a computer power supply. That will work. I would run the test at around .5 amps.

- Use the 5 volt and ground outputs of your power supply.
- Put a 10 ohm 3 watt resistor in series with the 5 volts. (Or something close.)
- Connect the lines to the shorted path.
- Probe the board with your volt meter, looking for the lowest voltage. You are looking for very small voltage.
Why 5v? it has 3v3. I'm worry that 5v would hurt my MCU and other 3v3 components on the board. this is the specs of my PSU:



And for 3wat resistor, can I hook up four 10 ohm 0.5watt in parallel?
 

mcgyvr

Joined Oct 15, 2009
5,394
I would simply use a multimeter on continuity mode/ohms (beeper turned on).
Stick one probe on ground and then probe about till you find a pin/trace that isn't supposed to be connected..

This board looks so simple there is no need for anything "fancier" (FLIR camera..current source,etc... there is no need).

Judging by the very poor quality solder job I see already I'm gonna say the issue is more than likely due to that.. Start with the fine pitch devices first.. Like that micro..

Of course the PCB looks to be of low quality too so I wouldn't be surprised if that could be an issue too..
 

dl324

Joined Mar 30, 2015
16,846
Why 5v? it has 3v3. I'm worry that 5v would hurt my MCU and other 3v3 components on the board.
3.3V is fine; some PC supplies (old ones) won't have 3.3V.
And for 3wat resistor, can I hook up four 10 ohm 0.5watt in parallel?
That would give you 2.5 ohms at 2W. 5V with 10 ohms would have given you 500mA max; with 3.3V, you'd want a 6.2 ohm 2W resistor.
 

nerdegutta

Joined Dec 15, 2009
2,684
Stick one probe on ground and then probe about till you find a pin/trace that isn't supposed to be connected..
I make my own PCBs, and before I even consider to start soldering, I'm checking the PCB for shorts. I put my DMM in "beep mode", and have one probe on GND and then I probe around. Sometimes I need jumper wires to connect the different ground planes. Those are the first thing I solder on, and then I probe again.

I guess with a board that's professional made, it could be hard. Especially if the board has conformal coating.
 
Well, the FLIR method is a near instant locator of heat generating issues such as a short. If the OP has access to one, it is a superior method of locating heat generating internal board faults, component faults, heatsink bonding faults, current overload faults, or even shorted faults hidden beneath multi pin components due to solder balls etc. Poking about fine pitch parts for faults can be very tedious.

I had a fault once that wasn't a dead short.....incomplete solder paste remelt bridging UNDER a 0805 part loading the LM317 regulator into thermal shutdown. So the resistance was a few ohms. The FLIR method identified the problem w/o having to de-solder anything to check.
It showed the traces leading to the 0805 part a few degrees warmer than others.
 

ian field

Joined Oct 27, 2012
6,536
Several of your links describe using a current source and a volt meter. This is a very good method. If you do not have a power supply with current adjustment, then put a resistor in series with your power supply and adjust the voltage. The lower the voltage the closer you are to the short.
If using it on a populated board; I'd suggest a silicon diode shunting the current source output to protect any sensitive active circuitry.

It'll give a standing conduction reading, but should still easily distinguish a short.
 
Top