Uncertainty Calculations In Thermocouple Simulation

Started by Hawaii596, 03-31-2011 -- 15:49:18

Previous topic - Next topic
Print
Go Down Pages1
User actions

Hawaii596

03-31-2011 -- 15:49:18
If you're simulating a thermocouple mV output from the T/C output of a Fluke 5520A (for example).  Let's say you are doing 100 Deg C Type K into a digital thermometer (such as a Fluke 51/52, etc.).  If you use a K(male) to K(male) interface cable.... so basically>> T/C output from the 5520A via a K to K cable to the input of the Fluke 51 meter>> if you are calculating the total uncertainty, how do you derive a value for the K to K cable?

Stupid question.  I just haven't had to deal with this little detail.  It seems like using the total standard K wire spec tolerance (+/-2.2 deg C) is a little much.  But there is likely some intermetallic offset variability.  And as T/C's are pretty "magic" there isn't a good, easy straight forward method for this.

Looking for some thoughts on this one.
"I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind."
Lord Kelvin (1824-1907)
from lecture to the Institute of Civil Engineers, 3 May 1883

retiredAFinSD

#1
03-31-2011 -- 16:14:39
Hope this helps.

hxxp: en. wikipedia. org/wiki/Thermocouple


metrologygeek

#2
03-31-2011 -- 16:16:32
If there's no temperature gradient there's nothing to compensate for. Your significant sources of uncertainty are your cold junction uncertainty in the 5520 and the DC accuracy of the 5520.

jimmyc

#3
03-31-2011 -- 16:20:42
i found this stuff, didn't really read most of it, but it may help

http://prod.sandia.gov/techlib/access-control.cgi/2004/041023.pdf

dminesinger

#4
04-01-2011 -- 10:37:26
Not having done much Uncertainty Calculations. My question is how do you fiquire the Uncertainty of leads for a DC voltage measurement of say a Fluke 5720A to an HP/34401A. Since your not using it as a thermocouple but leads, does it have to be fiquired in? I do not know. Looking forward to the answer.

PapaBear



Quote from: Hawaii596 on 03-31-2011 -- 15:49:18
If you're simulating a thermocouple mV output from the T/C output of a Fluke 5520A (for example).  Let's say you are doing 100 Deg C Type K into a digital thermometer (such as a Fluke 51/52, etc.).  If you use a K(male) to K(male) interface cable.... so basically>> T/C output from the 5520A via a K to K cable to the input of the Fluke 51 meter>> if you are calculating the total uncertainty, how do you derive a value for the K to K cable?

Stupid question.  I just haven't had to deal with this little detail.  It seems like using the total standard K wire spec tolerance (+/-2.2 deg C) is a little much.  But there is likely some intermetallic offset variability.  And as T/C's are pretty "magic" there isn't a good, easy straight forward method for this.

Looking for some thoughts on this one.
PapaBear

*'It's the Veteran, not the reporter who has given us the freedom of the press.'
'It's the Veteran, not the poet, who has given us the freedom of speech.'
''It's the Military who salutes the flag, who serves beneath the flag, and whose coffin is draped by the flag, who allows the protester to burn the flag.'

Hawaii596

#5
04-01-2011 -- 10:52:26
I just found at euramet.org website a calibration guide for, let me get the exact title: "Guidelines on the Calibration of Temperature Indicators and Simulators by Electrical Simulation and Measurement"  EURAMET cg-11, Version 2.0 (03/2011)  Previously EA-10/11.

It has a whole page on uncertainties, more than I likely will use in the low end item I am doing.

As a fellow sojourner in my struggle to learn uncertainties (one of these days I'll actually get a class on it), let me give it my two cents.  I am pretty intuitive and study a lot; and I think I've got a lot of it down.  But every once in a while, I scratch my head about one of these new little details.  And I'm nervous that Ill get it wrong.

I would say as with any uncertainty calculation, it depends on how much it actually contributes.  I'f I'm calibrating a Fluke 77 at 3 VDC, I know intuitively that it doesn't matter.  But if I'm doing the 1 MHz AC voltage (or many other parameters) on an HP 3458A, it often does.  My intuitive way to determine it would be experimentally.  Maybe do a series of measurements with the double shielded Fluke 5440A series leads, then repeat a number of times with some pomona leads.  If it were me, I might re-insert maybe 10 times with the Fluke 5440A leads, and do readings at 100mV, 1V and 10V.  Then repeat the process with Pomona leads, calculate a mean for each, and a standard deviation for each, and see if the difference adds up to significant change in the measured value.  If so, use that; if not, consider it to be insignificant.  You could do for a full set of measurements on a 34401A, and keep the results for reference.

"I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind."
Lord Kelvin (1824-1907)
from lecture to the Institute of Civil Engineers, 3 May 1883

RFCAL

#6
04-04-2011 -- 10:16:04
the leads would be insignificant, but inspectors want to see it included.Your bigger error will be operator error. Also do not forget environmental conditions, especially with the 3458A.
We had our low end DC Measure done by Suzanne Castrup,( ISGMAX.com ). She charged $1500 for the DC measure portion and included an explanation on CD for any inspector. From those calculations, I branched to the DCI,ACV,OHMS,ACI. Once you have the 10 "A" readings using a good meter calibrator, you can sub the specs for the different measurement areas. We used a Datron 4808.We still have one and it's accuracy is still the best of the Fluke calibrators.
I am pretty good with these Uncertainties, so if you have any other questions,I'll be glad to help.
Print
Go Up Pages1
User actions