Re: Count By Weight Scale
Without seeing exactly what they reported on, I'll just take a stab at it.
Say you have a device with a calibrated range of 0-1000, with a tolerance of 1. at 0, the error is measured at .4, at 200, .38, at 500, .2, at 800, .4, etc. I am within tolerance, but my error has no linear characteristics. The error reflects a non-linear curve. This is especially an issue with instruments that should reflect a linear scale.
While most instruments employ a zero/span type adjustment setup, when you adjust at low and high, it linearizes the new settings. Thus, you have confidence that you instrument will operate on a point other than the exact point you calibrated at. In the above example, while my instrument is in calibration, what can I reasonable assume is my error at 940? What about 593?
Just something you might want to keep an eye on.