What to do with correction factors overlooked by an accredited standard (6.4.11)?

[Gus]

Hello Colleagues,

i have bumped into one of those things "not mentioned in the test standard... that should be" and would very VERY much appreciate your input.

so, maybe 6.4.11 inst exactly worded for my particular case but i dont seem to find it anywhere else in the standard, so... to get to the point at hand:

I am in charge of a test lab that runs tests on fan cooled electric equipment, and Test site altitude DOES affect thermal equilibrium results on said equipment*, but current "certifiable" standards make no account for altitude corrections... the IEEE standard from which the aforementioned "other" standards were born DOES mention (briefly) a method to correct the temperature rise obtained from the tests, and also the IEC equivalent standard (uses a fancier formula but ends up with the same results for correction factors) now, to reiterate, we are accredited for national and international standards derived from these two main standards , but none of my accredited methods take into account these correction factors.

i am not sure how should i approach this to get to apply these correction factors and have my acceditation bodies to accept this "method deviation" (i'd like to avoid using that particular term if possible), does this really consist of a method deviation?... test method would be exactly as required by the standard but the result would be "corrected" for test site altitude

how do calibration labs approach these situations, im sure it happens right?

*This is due to the change in air density at different altitudes

 

[dwperron]

Your calibration procedure for these instruments should require that the correction factors be applied after calibration. That is what 6.4.11 directs you to do: you "shall ensure the reference values and correction factors are updated and implemented, as appropriate, to meet specified requirements".

The only reason not to use the corrections is if you can document that they are not a significant contributor to measurement uncertainty.

 

[Gus]

well yes, it should, but it doesn't, that's the issue here, am i still in compliance with the test standard if im "correcting" results in a manner not specified by the test standard itself?

and it probably doesn't matter but, its a "Test" procedure, not a "calibration" procedure

 

[Benjamin Weber]

Hi Gus,

here is my oppinion:

First of all your are accredited as a test lab based on ISO 17025. You write "certifiable" standards. CAUTION: An accredited test lab is not allowed to "certify" anything. There is ISO 17065 for certification bodies! But I assume, you are not really certifying anything, do you?

Your accreditation scope lists a set of standard and/or specific test procedures. If your scope covers the standards that do not contain the correction factors you are formally compliant with the standard if you don't apply the factors. On the other hand: If your scope covers the standards mentioning the correction factors you have to take the factor into account.

You should clearly communicate to your clients if you apply the factors or not. If the client explicetly requires you to apply the factors, make a clear statement in the test report, that this is the client's request.

As dwperron already mentioned: How much do the correction factors contribute to the total measurement uncertainty? Or better: How much does the final result change if you apply the correction factor or not? If this change is low in comparison to the total measurement uncertainty, you could easily just add this change as a single contribution to the total measurement uncertainty. Important: There is no hard limit to decide wether a single contribution is low in comparison to all other contributions. You have to make a decision.

I don't know your test procedure. But I can hardly imagine, that the air pressure/density has a major effect on the result in comparison with other contributions like uncertainty of the probes (thermoelement? PT1000?) or contact qualilty between probe and DUT.

 

[Gus]

Yeah, that was my bad on the "certifiable standards" , what i meant is that those standards are the ones recognized by the Certification bodies who have accredited the lab.

To be more clear on the impact of temperature tests, this is something that would impact product certification with UL, since they are the only who really care about temperatures, on the other standards there is a Heat run involved but im determining efficiency and the temperature value doesn't have a major effect on this result

BUT... the standard i mentioned earlier (which is not an accreditable standard but is the starting point from which the efficiency standards come from) tells you to reduce the Temperature rise in 1% per each 100m your test site is above 1000m ASL, our lab is 1625m.a.s.l. so... a reduction of 6% is due on Temperature rise... this means that for a temperature rise of 80... we change it into a Temperature rise of 75, and in some cases this could mean going from a NON compliant status, into a COMPLIANT status... since the product states that it is meant to be operated at 1000m (3300ft) and the customer has been always responsible for taking this into account, so in a way testing at a greater altitude without the correction factor makes R&D work harder for a cooler product than really required... as to the impact of this on uncertainty it seems kind of weird because i wouldn't have a ±U% related to this.... it would only be -U% on the temperature

in this case contact quality of the thermo-elements is not much of an issue, thermoelements are only used as reference to attain stability (that will be confirmed regardless of contact), actual temperature is attained by Change of Resistance method

 

[dwperron]

well yes, it should, but it doesn't, that's the issue here, am i still in compliance with the test standard if im "correcting" results in a manner not specified by the test standard itself?

and it probably doesn't matter but, its a "Test" procedure, not a "calibration" procedure

If you are claiming to use a test standard you must follow it completely, otherwise:

"7.2.1.7 Deviations from methods for all laboratory activities shall occur only if the deviation has been documented, technically justified, authorized, and accepted by the customer"