Help with uncertainty calculation in the calibration of a Pyrogen Test Processor

M

MetroBioCen

Hello everybody

I'm here again with another question for you.

I have to determine the uncertainty in the calibration of a Pyrogen Test Processor. This device is simply a digital thermometer with a thermocouple sensor. Well, in fact, with multiple thermocouple sensors (about 75) and the device scan every 15 minutes for all thermocouples.

The calibration procedure that I'm using is as follow:
I put the thermocouples in a stable bath (stability about 0,01 ?C)
When readings are stable, I take reading of all the thermocouples and the standard that I use.

My question is: For a type B uncertainty calculation:
Should I consider the uncertainty of the thermocouples and the digital indicator independently or;
Should I consider the uncertainty of the sensor-indicator as a whole.

Thanks in advance.
 

Stijloor

Leader
Super Moderator
A Quick Bump!

Can someone help?

If your questions were answered here, please consider helping others.

Thank you very much!
 

Hershal

Metrologist-Auditor
Trusted Information Resource
Each should be considered independently as a Type B contribution. Also, the uncertainty of the bath (if purchased) from its last calibration. You should also consider the uncertainty of thermocouples in the temp range they will be used. Also, if going through different environments (much hotter or much cooler) you should consider the effect of that on the uncertainty in the range they will be used.

Hope this helps.
 
M

MetroBioCen

Thank you for your help

Is there any paper that I can read about this?
 
M

MetroBioCen

Reading again the post by Hershal and some papers that I searched on the web I have found some dicrepancies. I have not found any paper saying that the TC and the digital indicator should be considered independently. This get me confused.

Help please!!!

These are the papers that I found this weekend.


http://www.ukas.com/library/Technic...hnical-Articles/Pubs-List/M3003_Ed3_final.pdf


http://www.upscale.utoronto.ca/PVB/Harrison/GUM/02_DigitalInstruments/02_DigitalInstruments.pdf


http://www.cem.es/sites/default/files/th-001e.pdf (sorry, in Spanish)
 

dgriffith

Quite Involved in Discussions
Here are some of the uncertainty contributions to be considered:
  • • Bias in the reference SPRT.
  • • Error associated with the repeat measurements taken with the SPRT.
  • • Error associated with the SPRT Scanner.​
  • • Error associated with the repeat measurements taken with the thermocouple.​
  • • Error associated with the thermocouple scanner.​
  • • Error associated with the calibration bath.
Each of these 6 may themselves have multiple components, like the bath:
  • Stability
  • Uniformity
  • Display resolution, etc.
Short term repeat readings will capture these; however, if a single reading must be taken these bath characteristics must be included.


Even if a t/c and readout were to be married together to minimize overall error, they still have their individual uncertainty components; error and uncertainty are not the same thing, don't forget.
*So my answers to your questions are :
"My question is: For a type B uncertainty calculation:
Should I consider the uncertainty of the thermocouples and the digital indicator independently or;
Should I consider the uncertainty of the sensor-indicator as a whole."
Yes, and Yes. Because, after you do the first, you will combine them (along with other uncertainty components) to arrive at the second.
 
Last edited:
M

MetroBioCen

Thank you very much dgriffith

Please, forgive my insistence, but this is very important in my conditions.

Suppose the case in which I do not have the chance to test them separately, but as a whole. But I have the uncertainty, reported by the manufacturer of both (in this case is the same manufacturer).

So, I must repeat the two questions again.

If I calculate the uncertainty independently (as I think I should), the result could be up to 0,3 ?C. The Pyrogen test is intended to detect a temperature variation of 0,5 ?C, so this equipment shouldn't be suitable, as the uncertainty could mask errors. But on the other hand, this equipment was built especifically for this test, by a specialized manufacturer, so, I think I'm doing something wrong.

Hope to have explained well, sorry, English is not my first language.

On the other hand, following this line of thought:

Would that be the same for a digital balance? Should I consider independently the uncertainty of the piezo-electric detector and the A/D converter?

Thanks in advance and apologise for my repetition.
 
Last edited by a moderator:

dgriffith

Quite Involved in Discussions
Not sure I can be of more help. If it was purpose built as a system, then the manufacturer should have quantified the error/toleranced the system as a whole. What errors/tolerances are discussed in the manual?
Thermocouple limits of error are unique to the thermocouple type (they don't all have the same error). It can be as much as a few degrees depending on type.
 
Top Bottom