Do i ever need to calibrate PT100 probes?

Hi.

Almost every time when working with PT100 probes the customers ask me "have the probes been calibrated?"
The reason for the customers to ask for it is that they often use their own probes also and can see that their values is slightly different from mine so their values is ofcourse the correct one.
I often claim that the values is different because the probes are installed different and meassures from not the exact same position etc.
That's my guess but i dont know for sure if my claim is correct.
(I'm not responsible for the installations so nothing i can do there)


My understanding is that there should never be any needs to calibrate them and i have never seen an option in any of the softwares that i use to "calibrate" them.
I mostly use siemens plc (1200 and 300) and for PT100 i use RTD analog modules and just reads out the value i get from the module and presents it as the real value. (No scaling required when using rtd)
I could ofcourse use some offset value to change the process value but should i?

So in short: Can i calibrate the probes and how do i respond to the customers in a correct way when they ask about calibration?

Years ago when I was on maintenace, if PT100 was reading out, we dump it in a kettle of boiing water and measure it.

If it was out, it got binned for anouther one.

I don't remember ever needing to calibrate it.

RTD elements/probes can be checked against a calibrated standard, usually a reference probe with its own calibration traceable to a national standard.

The probe is checked at X number of temperatures in a dry block calibrator or a regulated oil/water bath, over a temperature range specified by the customer.

This whole calibration procedure is done in a calibration laboratory who then issues a calibration report, including an ID of the specific RTD device. The calibration report states the deviation of the probe-under-test from the calibrated reference standard.

It is then up to the customer to decide how much error or deviation from the standard is tolerable.

There is no calibration adjustment on an RTD; it is what it is.

If its performance is close enough to the calibrated standard it is available for service, if its performance is not close enough to the calibrated standard, then it is not available for service.

I think your claim is basically correct. Also, any difference within a factor of 1.5 or so times the sum of the accuracy specifications of any two probes is insignificant.


TL;DR


If pedagogy is appropriate, another way of making the same claim is saying that a probe measure the temperature of the tip/sensing element itself and nothing else. If they do not accept/grok that argument then ask if they can guarantee an isothermal process with negligible heat flow along the probe bodies; at that point they should either get it, or look at you and blink.

I assume you are using the correct coefficient (.385, .375, .392, etc.) for your probe.

Finally, put crushed ice into a bucket of water (never the other way around, and used distilled water for both) and put both probes in there, then do the same with boiling water, and use that to "calibrate" the probes (is there any non-linearity in these things?).

Its a platinum probe (that's the "PT" in PT100) that changes resistance with temperature, no calibration available for the probe. If it doesn't read 100ohms at 0°C (the "100" in PT100) then the probe is bad and needs to be replaced.

What may throw off the reading is anything that changes the resistance of their circuit compared to your circuit. (bad wires, bad connections, splices, etc.)

It is possible that due to quality asurance policies, _every_ analog signal must be calibrated. And that the calibration must be done by a certified person or company.
I have experienced that. It made for some hilarious discussions about how to calibrate sensors that were extremely unlikely needing calibration, werent critical to the end-product, and were extremely difficult to calibrate, but STILL were to be calibrated ... due to policy.

Other than that, testing with freezing water and/or boiling water could be enough. Or if that is not possible, comparing during production with a portable sensor.

Real calibration in the true sense of the word is not a simple matter, even for a simple sensor.

For all temperature probes on machines I usually have in the HMI a calibration setting the the temperature can be adjusted +/- 10.0. The PLC adds this number to the reading and places it in the PV tag. If the calibration goes off over 10 then time to change the probe.

This is for RTD's and thermocouples.

The rtd sensor doesn’t get calibrated.
But look into a “decade box” for calibration of your controller or display
Only us old timers know about those things

Google it

JesperMP said:

It is possible that due to quality asurance policies, _every_ analog signal must be calibrated. And that the calibration must be done by a certified person or company.

Click to expand...

For ISO we had to have all of ours verified if not calibrated every year, even if they could not be adjusted they would still hook up and check against theirs so they could put the little sticker on it, basically everything that had a analog value would need to be calibrated and anything that was use to check process... heck even our tape measures had a sticker

You should connect same probe (or test equiment) first to your customer PLC input and then to your own PLC input.
That way you see if difference is from probe our from analog inputs and wiring.

An RTD cannot be calibrated. What people confuse for calibration is a offset after the R/I. Just dcam chiming in. Almost all temp inaccuracies are termination issues (sometimes ferrules which is trending here currently), RTD thermal well issues (degraded or no paste), or PLC/PAC/Instrument scaling issues.
I have a link to a good vid on wheatstone bridges and RTD understanding. Think its on work computer, Ill post it later.

“My candle burns at both ends;
It will not last the night;
But ah, my foes, and oh, my friends—
It gives a lovely light!”

― Edna St. Vincent Millay

You can check an RTD resistance at triple point (water, ice and vapour at the same time) and at boiling point. The issue is what minerals are in the water - EC and the like? This can slightly change the triple point reading but is more likely to change the boiling point of the water. No calibration as others have mentioned. It is often possible to calibrate the measuring instrument though. I know with older Omron analogue cards there used to be a calibration function - gone now with the Ethernet IP I/O.

It's not the RTD you calibrate, it's the transmitter/display/input card that drifts and may need adjustment.

R/I placement

More dcam madness- the RTD signal itself is not robust and sensitive to transmit line characteristics (Belden).

I see quite frequently plants hanging R/I pucks in panels hundreds of ft from the instrument.

If you use RTD's with seperate R/I converters for 4-20mA input modules you will get better results if you place the R/I in close proximity to the RTD and send the more robust 4-20mA current loop the long distance. If you want the MOST accuracy, eliminate all unnecessary termination points.

My last part of my rant is, I use headless IFM RTD's with built in R/I. I love them.

dcam said:

More dcam madness- the RTD signal itself is not robust and sensitive to transmit line characteristics (Belden).

I see quite frequently plants hanging R/I pucks in panels hundreds of ft from the instrument.

If you use RTD's with seperate R/I converters for 4-20mA input modules you will get better results if you place the R/I in close proximity to the RTD and send the more robust 4-20mA current loop the long distance. If you want the MOST accuracy, eliminate all unnecessary termination points.

My last part of my rant is, I use headless IFM RTD's with built in R/I. I love them.

Click to expand...

I concur. I have used these in a freezer (I think they're rebranded IFM):

https://www.automationdirect.com/ad..._(with_m12_cable_connection)/rtd0100-06-030-h

With a short pigtail to put this transmitter just outside the freezer:
https://www.automationdirect.com/ad...(rtd_input)/quick_disconnect/ttd-20-n40160f-h