NIST Traceability Numbers: Dense people, needle in a haystack

We all in the metrology community know that NIST traceability numbers are only NIST's equivalent of job order numbers, and can't be used as evidence of traceability, but there are a few holdouts, a few really dense people left out there who insist on beating the skeletal remains of this dead horse by asking that we provide NIST trace numbers because we claim to be "NIST traceable".

So, the problem I have is providing proof that NIST trace numbers aren't valid for the purposes of traceability. I found a good thread on NIST traceability numbers on this forum at NIST Traceability Numbers - Confused Yet?, but it's dated 1999, and besides that, I'm looking for "official" information, not people's opinions on discussion boards (it's like preaching to the choir).

I know NIST, NCSLI, etc. have all published papers on NIST trace numbers, where exactly can I find them? Searching for "NIST numbers" on the NIST website does no good, I get thousands of hits with "NIST" or "numbers", many of them dead links. I know ISO 9001 and 17025 have clauses on NIST trace numbers, does anyone know off the top of their head what they are? What about the positions that accreditation bodies like A2LA and NVLAP have on NIST numbers? It's like searching for a needle in a haystack.

Periodicall over the years I have emailed various people at NIST for questions on numerous topics. My experience is that they have always been good at writing back and doing their best to be helpful. If anything, they know so much more than I on what ever the topic, that I get a headache (a compliment to NIST).

You might try sending an email to someone there.

NIST's policyon traceability is found at :

http://ts.nist.gov/traceability/

The answer to your specific question is at:

http://ts.nist.gov/Traceability/SupplMatls/suppl_matls_for_nist_policy_rev.cfm#FAQ_Establish4

LAB (one of the ILAC MRA Signatories in the US) has a nice explanation here:

http://www.l-a-b.com/content/measurement-traceability

It turns out the particular customer that wanted to see the NIST numbers was using MIL-HDBK-52A, "Evaluation of Contractor's Calibration System" which dates back to 1964, and was last updated in 1984. Talk about vintage... NBS has long since been reorganized as NIST, and MIL-STD-45662 has long since gone the way of the dodo bird. We had a long meeting with this customer, and they were very adamant about it... We went out of our way to show them the traceability to higher up labs (e.g. Starrett, which is NVLAP accredited), and the fact that we are both ISO 9001 and ISO 17025/A2LA accredited, but they said that that meant nothing to them. We patiently tried to explain NIST's A2LA's, etc. position on NIST numbers and they didn't believe it, they said their other calibration provider was giving them NIST numbers and wondered why we couldn't do the same. We countered with "How do you know they're not just making the numbers up? Did you ever audit them to the scruitiny that you're auditing us?" By this time, they were getting rude... Our company president said if he would have been in that meeting, he would have shown them the door, as their account isn't big enough for the amount of hassle we got from them, and losing that one customer wouldn't affect our bottom line.

Another dive into the NIST traceability, found a little while ago, that NIST wants you to verify not just the chain of standards back to NIST, but all the related standards back to NIST.
Example: the calibration was done at 70deg.F. at 55%RH.
In this case the standards, which derived the temperature and relative humidity need to be tracked back to NIST. If a time piece of any kind was used, then we need to know that tracked to NIST also.
OMG, Does anyone actually do this?
And there seems to be no clear plan for this. Maybe a good spot for a new ISO-Standard

Wow, what a thread!

All measurements must be traceable to the appropriate measurement units of the SI, through a national metrology institute (NMI). In the USA, the NMI is NIST but others can be used under certain conditions.

If a calibration laboratory is accredited to ISO/IEC 17025, they are required to calculate their measurement uncertainty and show traceability for each contributor. So if temperature and humidity are contributors to the uncertainty, then those instruments must also have an accredited calibration with uncertainties. So yes, they should be doing it.

I really hate it when customers use obsolete standards documents. However, they are entitled to use whatever requirements they want. I would, like ScottBP, try to educate them but in the end they are the customer. Or former customer. An NCSL committee I am on recently spend many years, with the assistance of NIST, educating a federal regualtory agency on the modern arrangements for measurement traceability. That agency finally has it almost right in their requirements. (At least they are not as bad as another agency which insists that "calibration" always requires adjustment of the UUT.)

Graeme

I cannot speak for all ABs, but for IAS, the NIST numbers - absent SPECIFIC information about a SPECIFIC calibration at a SPECIFIC time/date - are NOT, repeat, NOT, considered valid evidence of traceability. This is also in accrodance with the NIST Policy on Traceability.

NIST numbers - MAY - be traceable - IF - there is SPECIFIC information to link the number to a SPECIFIC calibration of a SPECIFIC item on a SPECIFIC time/date.

Hope this helps.

Irrelevant. ILAC takes the position that if the calibration is traced back through an accredited calibration lab with better uncertainties then the intent has been met.

ABs take this approach, and ILAC position supports it.

george1weiss said:

Another dive into the NIST traceability, found a little while ago, that NIST wants you to verify not just the chain of standards back to NIST, but all the related standards back to NIST.
Example: the calibration was done at 70deg.F. at 55%RH.
In this case the standards, which derived the temperature and relative humidity need to be tracked back to NIST. If a time piece of any kind was used, then we need to know that tracked to NIST also.
OMG, Does anyone actually do this?
And there seems to be no clear plan for this. Maybe a good spot for a new ISO-Standard

Click to expand...

Thanks Graeme, you are right on the money.