Non linear variance in voltage with current

[ravekanth]

[7-10-2014 edited by moderator as requested]

I am trying to measure resistance by calculating the voltage across my test sample using NI9205. I have a constant current supply and about 30 cm copper wire across as connectors to my sample. I am trying to explore various current levels from 20-500 mA.

However the resistance or voltage/current ratio do not seem to be constant over current values. Increasing current seem to increase resistance by .01ohms.

I do change ranges in my daq from 200 -1-5 to 10 V.(+-) Heat due to current and changes in the sample thereof are not a factor. Is there a offset voltage when changin ranges which affects this? I am using the daq assistant with no custom scaling. I do however, take 1000 samples at 1000hz and average them.

[johnsold]

What size wire?  The excess resistance you report is almost exactly what AWG 20 copper wires would introduce.

Lynn

[Henrik_Volkers · Accepted solution]

Tried using 4 wire measurements?

0.01 Ohm in 30cm copper wire ... seems reasonable .... 

so: two wires for the current and two wires for the (differential) voltage measurement.  Guess in the picture below which screws are for current and for voltage  😉

[ravekanth]

I am still trying to figure out the size. Its litz wire though. And the variance is around .07 ohms. The .01 was a typo.

Any resistance i try to measure shows a drift with current.

Thanks for responses though. 

I was wondering though if there is any input offset for every voltage range for the daq.

[johnsold · Accepted solution]

Litz wire will almost certainly have more resistance.

As Henrik said, a 4-wire mesurement is probably what you need.

NI does not specify different offset voltages for each range. Amplifiers can have an input offset voltage and an output offset voltagewhich are independent of each other The total offset voltage measured at the output is the algebraic sum of the input offset voltage multiplied by the circuit gain and the output offset voltage. So the measured result will vary as the gain changes even if the effective input offset voltage is constant.

If the 4-wire measurement technique does not solve the problem completely, you may need to do some calibration runs to determine what effects the range changes have.

Have you looked at the 1000 sample before averaging to make sure that you do not have a significant amount of power line frequency interference or other time-varying signals present?

Lynn

[ravekanth]

I thought I had already responded, but thanks for the help 🙂 . I think variations seem fine now after I checked with another precision voltmeter as well.