Convert a multimeter into an accurate short circuit tracer for $1

[dicky96]

Hi guys
This cheap 'hack' is on Youtube and it seems to work really well. 


There was some discussion whether it could damage sensitive components such as CPU on laptop motherboards, or GPUs as it is using a 4.5V 50mA Constant current source to inject into the device under test.

I am of the opinion that because it is a Constant Current (and is in fact injecting the same voltage as commercial microOhm meters such as VC480C+) i fthe meter was connected to for example Vcore on a GPU or CPU, the 50mA current limiter would instantly reduce the voltage going into the CPU core to practically zero as it would appear as a 'short' o the meter - plus 4.5V at 50mA is only about 0.23W anyway and those devices are designed to draw tens or hundreds of amps at around 0.9V from the VRM anyway.

Also a multimeter on diode mode typically outputs 3V or more with enough current to light an LED (dimly) and we poke around boards with those all the time

I know there are a lot of knowledgeable guys here - so what do you think?  Is there a risk of damaging the DUT?

[kripton2035]

to get a lower voltage on the circuit to test, just build yourself a shorty (link in my signature)

[Kleinstein]

The idea with an extra constant current source is good and 50 mA also sound reasonable. However for just short finding, one could limit the voltage to a lower value, like 0.7 V by a diode.

With 4.5 V open voltage one may damage some circuits, like LEDs and mabe low voltage CMOS chips. With only some 0.7 V there are few parts that may be damaged (e.g. small shottky diodes, small (e.g. RF) tansistors), but this should not happen from the normal supply rails.

The idea with the separte current source to get a kind of 4 wire ohms - only 4 wires to the tips just fails short of the target.  For many cases the 50 mA current source and 0.1 mV DMM resolution is OK, though more resolution could help if possibly failed parts are relatively close by. So a 4.5 digit meter may be preferred.

[RoGeorge]

I've tried something similar before, and it's far from being accurate.  Might work, might not work, depending on the PCB.

In fact, for detecting shorted MLCC on fat PCB traces, it didn't work at all.  I've tried with much bigger currents on a laptop motherboard, and I've ended desoldering good MLCC capacitors.  There were a lot of MLCCs in that laptop.

I've repaired it by downloading the schematic, identifying the solder bridges (there were solder bridges by design, meant to debug the power distribution circuits) and disconnecting some bridges so to isolate the defective area, then only left with a few MLCCs to check, and finally found the shorted one.

[Shock]

You can do it even cheaper if you already have a power source, multimeter and a resistor.

to get a lower voltage on the circuit to test, just build yourself a shorty (link in my signature)

Did you post the voltage and current specs anywhere?

There is also the Vici VC480C+ now, which takes reasonable sub milliohm measurements but is manual ranging and lacks digits.

The YR2050 looks promising but it needs better firmware.

[Gyro]

You really need lower voltage and more current to get decent resolution. Wind your bench PSU down to around 0.5V, set the current limit to 1A or more and connect it across the shorted supply rail and GND. Then with the meter on the lowest mV range, probe around the planes and finally components  to narrow down the lowest reading.

[dicky96]

@shock - this little mod seems to work as well as the Vici and more stable than the YR2050 - The Vici also outputs 4.5V and 50mA (Actually 4.8V and 100mA on the 20mOhm range)

@krypton - I bought all the parts for Shorty display last year on badcaps.net forum thread, but you never had the PCBs ready?  I wanted to buy one or two and try it.How much current/voltage does it use to test?  do you have PCBs now and how can I contact you?

@Gyro and several other - that $1 mod has enough resolution to find a shorted MLCC in a block of 15 capacitors all next to each other (millimeters apart).  How much more resolution does it need to be a usefull short circuit tracer?
At 15 mins/  15:00

 

[kripton2035]

Did you post the voltage and current specs anywhere?

from a simulation of the input schematic, less than a millivolt, and 35mA current.

[kripton2035]

@krypton - I bought all the parts for Shorty display last year on badcaps.net forum thread, but you never had the PCBs ready?  I wanted to buy one or two and try it.How much current/voltage does it use to test?  do you have PCBs now and how can I contact you?

sorry still not done. lots of people send me their protoboard picture and telling it's working nice so the pcb is on the side for now...

[RoGeorge]

Something like this seems very promising for finding PCB short-circuits, but I haven't tried:




Full project:  https://hackaday.io/project/3635-shorty-short-circuit-finder

[dicky96]

With 4.5 V open voltage one may damage some circuits, like LEDs and mabe low voltage CMOS chips. With only some 0.7 V there are few parts that may be damaged (e.g. small shottky diodes, small (e.g. RF) tansistors), but this should not happen from the normal supply rails.

I just tested this $1 meter hack on a standard 5mm red LED - the LED lights up but it doesn't come to any harm.  The meter reads 2.06V

[kripton2035]

Quote from: RoGeorge on Today at 14:20:47

Something like this seems very promising for finding PCB short-circuits, but I haven't tried:

this is the original project I modified by adding a display and named it "shorty-with-display"