Page 1 of 1

Question about uncertainties

Posted: October 12th, 2020, 11:52 am
by Ohio_Guy
I have a simple question about uncertainty specifications for digital multimeters.

The specifications sheet for multimeters lists the absolute uncertainties for various functions. As an example, the specification for measuring DC voltage may look something like this:

Calibration interval: 1 year
Range: 200 mV DC
Uncertainty: ±(100 ppm of reading + 10 ppm of range)

So, for example, if the meter is displaying 150.0000 mV, the uncertainty would be ± 0.017 mV.

O.K., I understand that. But let’s say I only care about a change in voltage, and not the absolute voltage itself. As an example, let’s say I measure a change of 2.000 mV:

Starting voltage: 150.0000 mV
Ending Voltage: 152.0000 mV
Change in voltage: 2.000 mV

What is the uncertainty on the 2.000 mV? What equation or method would I use to calculate the uncertainty on the 2.000 mV change in voltage?

Re: Question about uncertainties

Posted: October 21st, 2020, 6:13 am
by Andrea C
From the specs, you can compute the error for each reading at the 150mV and 152mV levels.
Using your example, that is 17uV and 17.2uV

Do RMS on these to arrive at 24.4uV.
I believe RMS would be the error combining approach.

Worst case, suppose the one reading was at low limit of allowed error and second reading was at high limit of allowed error. That's effectively just adding the errors of each reading together to give you error of 34.6uV on their relationship. But statistically, you'd be having awful luck to have such a distribution.