5 point linearity protocol for level transmitter

Dear,

i was looking at the datasheet for level transmitter from E&H, and they offer 5-point linearity protocol instead of the basic (i guess 3-point linearity protocol),

I can see that 5-point will be better than 3 point when doing the calibration, what do u think? am i right?

thank you!

charbel

I find that people use characterization/look-up/linearization tables for volume conversion in inventory applications, rather than plain level correction.

I think one reason is the difficulty in getting numbers accurate enough use for correction because it requires knowing when the level is at exactly 25%/50%/75%. Does the effort involved warrant the nominal increase in accuracy?

In general, instruments tend to be "accurate enough" for people to use the readings they get for process level measurement.

For inventory purposes, which use volumetric units, they'll sometimes go to the effort to use correction functions.

Charbel said:

Dear,

I can see that 5-point will be better than 3 point when doing the calibration, what do u think? am i right?

charbel

Click to expand...

Yes, this is correct in a calibration, it gives you more data points to check for error and can show you better if the instrument is starting to fail.

The more points the better but there is a caveat. You still must know what curve or line these points are supposed to be fit too. It is possible to fit a 4th order curve to 5 points but that probably be wrong. Usually a second order curve is good or something logarithmic or exponential.

The word "better" is a value judgement, and therefore depends on the application.

If you are just controlling wet well level in a sewage pump station then three point calibration is probably perfectly adequate.

If you are doing custody transfer of a fine bourbon whiskey then five point probably isn't good enough.

In a great many applications repeatability is more important than accuracy. Also, in a great many applications installation errors more than offset the inherent instrument accuracy.

The point is you have to weigh the economic benefit of more accuracy (if any) against the additional cost of obtaining more accuracy (if any).

thank you guys for your valuable replies