Accuracy at lower than full scale

[r.tollund]

Hi. I am using a NI-USB 6255 for differential voltage measurements. I am running it at the lowest voltage scale: 0.1V. From the specification (https://www.ni.com/docs/en-US/bundle/pci-pxi-usb-6255-specs/page/specs.html) I can see how to calculate the absolute accuracy at full scale. However, I am mostly measuring voltages down to 0.1mV, which is just one thousandth of the full scale. At these voltages the absolute error is relatively huge. Is there some way to calculate an absolute accuracy at a lower voltage to get a better accuracy? I would expect the absolute accuracy to be lower at lower voltages.

Thank you.

[santo_13 · Accepted solution]

In short, at worst, your reading in 0.1V range is +/-52uV

• TempChangeFromLastExternalCal = 10 °C
• TempChangeFromLastInternalCal = 1 °C
• number_of_readings = 100
• CoverageFactor = 3 σ

To get more accurate abs error for a reading in 0.1V, here is an example for 0.05V reading in 0.1V range,

• GainError = 150 ppm + 13 ppm · 1 + 1 ppm · 10 = 163 ppm
• OffsetError = 150 ppm + 21 ppm · 1 + 60 ppm = 131 ppm
• NoiseUncertainty =  15uV * 3 / 100 = 4.5 µV
• AbsoluteAccuracy = 0.05 V · (GainError) + 0.1V · (OffsetError) + NoiseUncertainty = +/- 25.8 µV