Resistance measurement accuracy with SCXI

[AEI_JR]

I'm looking into using the SCXI-1102 and SCXI-1581 to make resistance measurements of less than 100 ohms.  I calculate that the best accuracy that I can acheive with this setup is 0.332 ohm (through the SCXI-1600).  Is this correct?  Is there a way to easily improve the accuracy using SCXI?

[Doug_L._Bear]

I calculated it out to be ±0.491 Ohms, work shown below.  To get better accuracy, make sure you are operating in the 25° C range, and you could perform an external calibration on it and take your measurements within 24 hours of that.  Other than those though you are pretty much stuck with this accuracy.

1581 Accuracy of source (from spec pg A-1)  - ±0.05%

1102/1600 Accuracy (from ni.com accuracy calculator) -±0.0489 mV at 10 mV input, single point reading

100 Ohm * 100 µA (the source of the 1581) = 10 mV

0.0005 * 10 mV =  ±0.005 mV 1581 Accuracy

Total Accuracy = sqrt(±0.0489^2 + ±0.005^2) = 0.0491 mV

0.0491 mV / 100 µV = ±0.491 Ohms

[Henrik_Volkers]

If self heating isn't a problem, you can think of increasing the current.

Use an external current source (1mA gives 10x more signal) and spend one (differential) channel for an external known reference resistor (again look for self heating) 

If you have a stable current source and you use only one preamp and one digitizer your result depend only of the linearity (and resolution) of the digitizer and the stability of your R_ref

Powering the resistors only for short for measurements is the next step to reduce mean current/power loss