G
Sorry, I don't have an excel template. These are hard numbers to nail down, as people will measure them different ways. For Linearity you typically do a linear fit on the data, then calculate the max deviation from the fit. Finally, you divide by the full scale to get a percent full scale. Bias is usually pretty simple. Take data for some time (say 100 - 1000 samples). Calculate the average value, that's the bias. Stability is very tricky. In the sensor world, one of the standard ways to measure stability is using an Allan variance calculation. See attached for a short description.
Edit -- after looking around the forums a bit, it looks like you're probably talking about measurment errors? Look through some of the related topics, I think this has been covered before.
Edit -- after looking around the forums a bit, it looks like you're probably talking about measurment errors? Look through some of the related topics, I think this has been covered before.
Attachments
-
393.6 KB Views: 1,508
Last edited by a moderator:
