Thermocouple RTD single point check 2

[youngblood81]

Does anyone use a "single point check" to verify the output of their thermocouples and RTDs (basically checking to make sure the element is reading something), then use a TC/RTD simulator to check the rest of the loop? Everything I've read indicates that RTD's drift, and thermocouples drift even moreso, which leads me to believe that this method is practically useless. Can anyone suggest any published standards for TC/RTD loop calibration?

Thanks

 

[hacksaw]

single point checks are for circuit commisioning and absolutely essential to get a system up and running, the issues that you've raised are rarely evaluated in end-use applications

the isa is a good source for such specialized issues as are rtd-t/c manufacturers

 

[mgtrp]

We have plenty of contractors who - when left to their own devices - will try to get away with only a single point check, but we typically insist on a complete check using an oil bath to heat up the RTD. This practice was consistent across all companies that I've worked for/with in multiple countries, although all electric utilities.

Using an oil bath may not be so easy on a thermocouple, which may be measuring many hundreds or even thousands of degrees.

 

[youngblood81]

Thanks for the input guys! It is much appreciated! Everything piece of information I've come across leads me to believe that the "single point check" tells us little to nothing about the percent error in the loop, especially with the elements being the "weak link". Thanks again, and take care!!

 

[electricpete]

Everything I've read indicates that RTD's drift, and thermocouples drift even moreso

I'm not familiar with that. In fact I have always assumed that thermocouples and RTD's are very stable as far as the sensors themselves (other than things like loose connections which tend to increase resistance). Can anyone provide more info to support the quoted statement? (just curious)

=====================================
(2B)+(2B)' ?

 

[rasevskii]

I agree with electricpete. Many years ago I worked for a very well known company involved in process control and commercial temperature control systems, as a field engineer.

Drift of thermocouples was/is not an issue. The MV output is a function of two dissimilar metals at the given temperature. The same with RTDs: the ohmic value is a function of temperature, period.

If anything has gone wrong with the TC or RTD (burn-out, ground fault, etc) there will be a sudden large change in the reading and the measuring instrument will show this. In fact with TCs there was always a burn-out circuit in th instrument that drove the reading up to the end of the scale (safely shutting off the heat source) in case the TC went open circuit.

But with the actual measuring instruments themselves, drift can occur and therefore such instruments were calibration checked at regular intervals, anywhere from once every month up to once per year, depending on the criticality of the application.

Unfortunately I think youngblood has gotten the wrong information from somewhere.

rasevskii

 

[youngblood81]

electric pete

I've attached my references that initially supported my suspicion that thermocouples and RTD's drift. Can anybody attach other references/data to the contrary, because I would seriously love to see them?

Thanks,

 

[youngblood81]

Sorry, I thought I posted two references, but I suppose it only allows one file per post. Let me try this one again...

 

[youngblood81]

The first reference is a 10-page document, but I've highlighted the pertinent material. With this material in mind - including the fact that the first reference refers to temperature elements (TCs/RTDs) as the "weak link" with regard to accuracy - I don't see that a "single-point check" would do anybody any good, when a certain degree of accuracy is desired. Can anybody produce any info to the contrary?

Thanks

 

[DRWeig]

I you search around enough, you can find manufacturer's data for RTDs with stability specs. Most common is 0.05°C per five years, but high-purity versions are available ($$$) that get down to 0.005°C per year. With standard accuracy and a good initial calibration / commissioning, drift over time shouldn't be much of an issue... But then, most of what I work with does not require such close tolerance.

Good on ya,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[electricpete]

Thanks yb81 and Dave for responding to my question. I was not aware of those type of specs. LPS for taking the time to educate me.

It looks like RTD's have very good temperature stability. If we look at Dave's number "common" RTD's 0.05C per 5 years... would take 100 years to drift 1C. It is really irrelevant for the applications I'm concerned with (primarily motor bearing and winding temperature sensors).

I see much higher rates mentioned for thermocouples, primarily related to oxidation/corrosion which are temperature dependent and naturally accelerate at higher temperatures. Mentioned temperatures of 800F are nowhere near what I'm interested in (again, motors). It's not clear to me exactly how big the drift is in the range up to let's say 250F which is about the highest winding temperature I ever see.

So as expected it probably depends on your application (max temprature) and requirements (is 1C change significant?).

I don't see that a "single-point check" would do anybody any good, when a certain degree of accuracy is desired. Can anybody produce any info to the contrary?

I can see the logic coming from the perspective I had before that thread. That is: each RTD or thermocouple comes with a calibration table that tells us resistance vs temperature (rtd) or voltage vs temperature (thermocouple). If the stated accuracy for the table is well within our requirements, then the main thing we want to double check is that we have the right sensor installed... which single point check will do. Sure, you get better check with two points, but still not perfect if you have a non-linear table lookup function for thermocouple. More points is always better, you can still question the in between points to some extent not matter how many points you check. How valid are the questions depends on how exact you want to be or need to be.

I can also see where coming from your viewpoint (that the drift is significant for your application), this approach would not be valid.

Thanks again. There is always something to learn here.






=====================================
(2B)+(2B)' ?

 

[electricpete]

Since this is a motor forum, I'll talk about winding RTD's in particular - there are very limited options for calibration check of the sensor itself. You would be lucky to get a single point check when motor is disassembled for refurbishment at relativley uniform temperature. I wouldn't ask to vary the temperature of the whole stator to accomplish RTD check. Also during operation, we can cross-check the RTD's to see how closely they agree with each other. Differences may represent actual differences or indication errors. As long as the spread is less than 10F, I personally don't get concerned.

=====================================
(2B)+(2B)' ?

 

[youngblood81]

The applications I'm thinking of are more along the lines of critical process temperatures loops that are "calibrated" annually, and have calibration tolerances associated with them. Throughout this research I've been asking myself how a "single-point check" can satisfy our needs as some have suggested, and how can we know that we don't have issues after the check is performed. The more I read the more I've convinced myself that we cannot be convinced. I'm not suggesting that for non-critical applications it wouldn't be useful to test for some degree of functionality (or just to make sure that noone has stolen the element, lol). I'm just not convinced the "single-point check" should be used for critical temp loops.

Thanks