4/20ma Sensor scaling/calibration

Dale87

Lifetime Supporting Member
Join Date
Feb 2022
Location
Illinois
Posts
59
5069 IF8 with a 4-20ma 0-100psi, sealed fixed, no calibration, transducer attached. IF8 is scaled 4ma= 0 psi, 20ma=100psi. If over time the sensor weakens and zero psi no longer is 4ma, this can be adjusted on the input card ie. 3.997ma=0psi. How do you calculate the scaling for the maximum value if you have no way of putting 100psi on the sensor. The normal pressure on the sensor is around 51psi. If you zero the low end and take a reading around the halfway point, can you calculate what you need to change the max to ie. 19.89ma = 100 psi.
 
Divide the actual pressure at half scale by the pressure reading, that will give you a value. Multiply your maximum scale by that value.
So if you were measuring 50psi and your PLC was giving 51psi, it is reading a bit high. 50 divided by 51 is 0.98, 100psi times 0.98 is 98psi, that is your new top scale.
Then run the check again to make sure that it is now reading correctly at 50psi.
It you were only able to check at 1/10 scale then the error margin would be too big, but a 1/2 scale I don't see a problem. Of course if you have something incredibly sensitive to pressure changes you might have to rethink and buy a pressure calibrator that can be used with the sensor off the machine.
 
You need to multiply by unity i.e. multiply by 1.

TL;DR

For the nominal calibration, a rise of 100PSI, from 0PSI to 100PSI, is equivalent to a rising of 16mA, from 4mA to 20mA.

So
(100PSI - 0PSI) ≡ (20mA - 4mA)

100PSI ≡ 16mA
Divide both sides by 16mA:
100PSI ÷ 16mA ≡ 16mA ÷ 16mA

(100/16) PSI/mA ≡ 1

6.25 PSI/mA ≡ 1
That 6.25 PSI/mA is the slope of the nominal calibration line, and it is equivalent to unity (one), as shown above.

If the minimum calibration values (mA and PSI) change, and we can't measure new calibration values around the maximum (20mA ≈ 100PSI, then we assume the slope of the mA vs. PSI characteristic calibration line from the minimum reading [0PSI,3.997mA] to the "normal" reading [51PSI,normal mA] must be same as it is from the minimum reading [0PSI,3.997mA] to the maximum reading [max_PSI,20mA]:

(normal_PSI - min_PSI) (max_PSI - min_PSI)
Slope = ---------------------- = -------------------
(normal_mA - min_mA) (max_mA - min_mA)


Except for max_PSI, we presumably know everything:

  • normal_PSI = 51PSI;
  • normal_mA = mA reading taken at 51PSI;
  • min_PSI = 0PSI;
  • min_mA = 3.997mA;
  • max_mA = 20mA.
Solve for max_PSI.
 
You need to multiply by unity i.e. multiply by 1.

TL;DR

For the nominal calibration, a rise of 100PSI, from 0PSI to 100PSI, is equivalent to a rising of 16mA, from 4mA to 20mA.

So
(100PSI - 0PSI) ≡ (20mA - 4mA)

100PSI ≡ 16mA
Divide both sides by 16mA:
100PSI ÷ 16mA ≡ 16mA ÷ 16mA

(100/16) PSI/mA ≡ 1

6.25 PSI/mA ≡ 1
That 6.25 PSI/mA is the slope of the nominal calibration line, and it is equivalent to unity (one), as shown above.

If the minimum calibration values (mA and PSI) change, and we can't measure new calibration values around the maximum (20mA ≈ 100PSI, then we assume the slope of the mA vs. PSI characteristic calibration line from the minimum reading [0PSI,3.997mA] to the "normal" reading [51PSI,normal mA] must be same as it is from the minimum reading [0PSI,3.997mA] to the maximum reading [max_PSI,20mA]:

(normal_PSI - min_PSI) (max_PSI - min_PSI)
Slope = ---------------------- = -------------------
(normal_mA - min_mA) (max_mA - min_mA)


Except for max_PSI, we presumably know everything:

  • normal_PSI = 51PSI;
  • normal_mA = mA reading taken at 51PSI;
  • min_PSI = 0PSI;
  • min_mA = 3.997mA;
  • max_mA = 20mA.
Solve for max_PSI.



On a analog input card where you can change both the 4/20ma and the engineering units, is it better to leave the engineering units alone and adjust the ma. Ie 3.9997ma=0psi / 19.9786ma =100psi or is it better to change the engineering units? Ie 4ma=.0003 psi / 20ma= 99.58psi??
 
whatever you put isnt going to be truly accurate unless you put 100 psi and see what the sensor puts out. you might get close but no way to know unless you hit it with 100 psi.
 
On a analog input card where you can change both the 4/20ma and the engineering units, is it better to leave the engineering units alone and adjust the ma. Ie 3.9997ma=0psi / 19.9786ma =100psi or is it better to change the engineering units? Ie 4ma=.0003 psi / 20ma= 99.58psi??
It probably does not matter for accuracy.

There are 24 bits of resolution in those numbers, and far fewer than that coming from the raw A/D Converter.

So unless the limits prevent extrapolation, the difference would be hard to notice.

In fact, if the module does extrapolation, and you are not bothered about overrange/underrange conditions, then you can simply enter the normal_mA and normal_PSI values (e.g. 12mA and 51PSI for High Signal and High Engineering, respectively), and let the module make the calculation from there.
 
You might be looking at the wrong end of the horse here.

Why is the pressure transmitter drifting? PSIG or PSIA? If it is PSIG then it should have a breather tube. Is the tube plugged or damaged? If it is PSIA then it will go up and down a little bit from changes in the barometric pressure. Nothing you do about that.
 
I will support post 7 - easily verifiable, probably fixable.

Regarding other "scientific" "research":
1. I will assume that the sensor works on the piezoresistive principle.
Feel free to find out:
- sensor design
- mathematical formulas for the dependence of the deformation of the membrane, the dependence of the output signal of the piezoresistive element on the deformation, and derive the final formula, taking into account the design features of the sensor (connection of the piezoresistive element with the membrane)
- what approximating transformation is used in the transducer of the sensor
2. Make an effort on yourself and take a look at the sensor manual. Perhaps there will be the words "accuracy", "temperature error", "linearity", etc.

So
1st
4 ma - 3.997ma = 0.003 ma
0.003mA * 100 psi / 16 mA = 0.01875 psi
this is 0.01875% of the measuring range
(I know what relative error is)
2nd
It is not known what exactly is the cause of "sensor weakens"
3rd
Missing control measurements

Given the facts above, do you really think applying a linear correction to the calculation is a good idea?
 
The only reason I'm asking is we are in the middle of a slc500 upgrade and some of the SCP min max have been altered from 3277 and 16383. They may be 3340 and 16277. So I'm trying to work that backwards into the scaling properties on the newer 5069 analog input cards. I was just curious as to how the person that altered these numbers came up with them. Obviously 3277 was altered to get a zero reading on the HMI when there was no flow/pressure/weight on the sensors. But I didn't understand how they where changing the 20ma-16383 value on things like 1000lb scales, 400gpm flow meters, or 200psi pressure sensors. Short of bringing in 1000 pounds of test weight, or taking a sensor out and applying 200psi of test pressure.
 
They probably just put a known weight on it and then adjusted the max until the plc matched what the sensor said. so zero the reading and set the min. put some reading on the sensor and adjust the max to match what the sensor or a gauge says.
 
So I'm trying to work that backwards into the scaling properties on the newer 5069 analog input cards.
I am not sure this is the best approach, even though it would work.

The problem I anticipate is when a card fails and needs to be replaced. Now, in addition to physical replacement of the card, the tech, at 3am mind you, needs to find the latest calibration values (this is probably the toughest part) for each replaced channel (pressure, weight, flow, etc.) and make sure they are in the module's scaling configuration for the new card.

Better to have a single calibration/scaling configuration for the card (nominal module scaling config is 4mA is 3277 EU and 20mA is 16383 EU), and then the SCP does the rest. Also makes it possible to do calibration via the HMI, if desired, because now those 3277-16383 "EU" values can be viewed without hooking up a Studio 5000 laptop to the PLC or network and going online.

Just one less (or several fewer) things the tech does not have to deal with, or you being called in at 3am to do.
 
I am not sure this is the best approach, even though it would work.

The problem I anticipate is when a card fails and needs to be replaced. Now, in addition to physical replacement of the card, the tech, at 3am mind you, needs to find the latest calibration values (this is probably the toughest part) for each replaced channel (pressure, weight, flow, etc.) and make sure they are in the module's scaling configuration for the new card.

Better to have a single calibration/scaling configuration for the card (nominal module scaling config is 4mA is 3277 EU and 20mA is 16383 EU), and then the SCP does the rest. Also makes it possible to do calibration via the HMI, if desired, because now those 3277-16383 "EU" values can be viewed without hooking up a Studio 5000 laptop to the PLC or network and going online.

Just one less (or several fewer) things the tech does not have to deal with, or you being called in at 3am to do.




I started it out that way with a SCP AOI. But in an effort try and clean things up I just went to scaling on the card. 5069 if 8 comes default with 4ma=0EU and 20ma=100eu. But at least if you had an SCP instruction in the program, even if it had been slightly altered for scaling, the tech should pick up on what's going on and get 3277 and 16383 back into the cards scaling properties for min max EU. It would be easier to convert from logix 500 with a SCP AOI in l5k. Just match exactly what's in the 500 program on the SCP AOI in L5K. I also thought about leaving the card scaling, and where I mapped my analog channels to PLC tags just write a short description of how the scaling is set up. With min max EU and MA. V30 L5K rung comments would remain even on an upload I believe.
 

Similar Topics

Hey guys, I have a 4-20ma current sensor 4ma = 0 amps, 20ma = 100amps I'm using automationdirect remote field IO. (T1F-16AD-1) I get a...
Replies
1
Views
2,268
I'm new to S7/TIA. There's no devices connected to the PLC yet. I've got a laser measurement sensor (Micro-Epsilon optoNCDT 1220). From the...
Replies
18
Views
2,864
Hi; I have installed a temperature and humidity sensor Autonics THD-R-C which gives 4-20mA signal out. The signal is connected with Fatek...
Replies
3
Views
2,708
Hi Guys, How is Analog 4-20mA Level sensor Programmed in CLICK software of Automation Direct? Which Instruction set.
Replies
5
Views
2,163
For example I have pressure sensor 0-6Bar (4-20mA). I have PLC that have to relay pump based on pressure and I also have digital indicator that...
Replies
8
Views
2,179
Back
Top Bottom