Pressure transmitter drift

[Bobodopalus 2]

Hello all

this isnt strictly omron/programmer but has anyone used much in the way of pressure transmitters?

we have a wika s-11 pressure transmitter on one of our machines installed roughly 7 months ago where the scaling does not seem to be linear, i only started working here around that time so i dont know how it was installed etc. but its a 0-40 bar 4-20mA transmitter connected to an ad081 4000 resolution connected to a cj2m.

so the conversion should be nice and easy 100 raw data = 1 bar however it seems to be scaling logarithmicly to me as we checked with a 2 different guages and when they were at 2 bar the transmitter was showing 137 raw data and at 8 bar it was showing 737 raw data and this seems to keep scaling in this fashion as we get to higher bar.

is this a common fault on transmitters?

is there anything else that could be causing this?

how often would you normally expect a transmitter to last?

i checked my scaling FB on another transmitter connected to a compressor and it worked correctly so its not the maths used

 

thanks in advance

[panic mode 246]

you only posted two sample points which is not enough to determine curvature. non-linear sensors can be linearized in software, using some sort of fit or lookup table. lookup table is suitable only for large number of points or where nonlinearity is very complex (not monotone). one approximation is to use segmented linearization with only few points and interpolate values that are in-between. i think the most practical solution to linearize is to collect muhc of points, find curve (polynomial if possible) that is a good match and then implement it in code.

so is your product really sporting linear output?

i guess the safest thing to say is that it depends on the product. so check the datasheet.

if i am not mistaken, it is this one:

https://www.wika.ca/upload/DS_PE_S_11_en_us_16562.pdf

and it claims  Non-linearity (per IEC 61298-2) ≤ ±0.2 % of span BFSL

to me that sound pretty linear.

how exactly did you wire it and are you sure your analog input is configured correctly?
 

 

[Bobodopalus 2]

im fairly sure the analog input is configured correctly as there is another transmitter working on it correctly.

the transmitter should be linear yes,those 2 readings are just a couple i wrote down, we did go up and down the range and raw data would get further and further away as the pressure was increased, using a scaling function block that is proved to work, it would be correct scaling around 9 or so bar and then vary more and more as it raised.

when i go back on site i will be testing with an amp meter, unfortunatley i dont have a amp generator or analog output unit i could use to prove the analog unit.

i have tested the same transmitter with the same function block just now and the readings were perfect 1 to 1, 1 bar = 100 raw data

sadly i didnt install this and couldnt say about the wiring, i believe its only 2 wire so i would doubt thats been done wrong, and the analog unit settings look fine to me