Using standard copper as a shunt - FAIL

[BiOzZ]

i made a post a few days back about using 14 AWG house hold grounding solid copper wire as a high current shunt and i would like to fowlow up on my trials

someone posted that the temp co of the wire is simply too high to be useful and the wire will heat up way to fast to be useful

yes and no! ... yes the tempco was so high it was unusable short of calibrating it before every use
no the shunt wire did not heat up at all at 9.5 amps for 1 hour (seeing how it was thinner than the wires used to connect it ... it was not a real concern)

here is the rig that i used to test it



it is all messy so let me explain

the large meter is my reference ... its on the low side and in amp mode

for current regulation i use a SIMPLE mosfet on a fanned heatsinc with SIMPLY a pot on the gate ... i adjust the pot untill i have the current desired ... this was VERY stable not drifting at all in 1000 counts

EDIT: for the power supply i used a standard 250 watt benchtop PSU modified HEAVILY ... big inut chokes, big output chokes and filters, slightly larger caps and ground isolated (shorted to ground here tho ...)

here is the circuit i used



i used a simple instantiation amp that i had sitting in my parts bin that i use on all sorts of things ... but im sure it worked just fine for this


Results

i dont remember what i had the pot set as but the gain should be ~ 200 ish

at room tempature for a full hour the voltage out at 9.5 amps was 2.415v and only drifted by 4 on the LSD

as SOON as i applied heat to the copper it jumped up to 3.2ish volts than slowly bleed down

CONCLUSION

if its not being constantly adjusted for temperature or in a temperature controlled environment ITS USELESS

sorry to whoever i doubted!

[Alana]

How about janking 10A current shunt from those 5$ Chinese DMMs then?

[croberts]

Can you epoxy a temperature sensor to the shunt wire and then use it to temperature compensate the output.

[jeroen74]

Wouldn't it be easier, cheaper and less complex  to just use the proper kind of wire

[C]

If your copper current shunt was the wires needed anyway, like the wires from a motor drive circuit to the motor, it would be dual use and not waste power.
An old car I had did just that for a charge/discharge indicator.   

The problem is if you want/need a more calibrated reading.
If this was a temp sensing resistor and you wanted accuracy, you would use a 4 wire kelvin setup to measure it's resistance.

what if you combine that idea with a way to sense the temp indirectly.

For example, say you took a 10 gage magnet wire and in parallel put two 30 gage magnet wires  The two 30 gage wires in close contact would have almost the same temp. as the 10 gage wire, so the resistance of the 30 gage wires would change almost the same ratio as the 10 gage wire. Connect all three at remote end and cover with some heat shrink. By using 2 wires you can get the voltage at the remote end and use the resistance change to help calibrate out the temp change.

C