What does "UKAS calibrated" mean?

[try]

Hello,

when looking at the following ad

https://www.ebay.de/itm/UKAS-Calibrated-Agilent-HP-34401A-6-5-Digit-Multimeter-KEYSIGHT-HP/264312546608

I discovered that the calibration certificate was issued by Trescal but at the top it carries a UKAS logo.

From my limited understanding this DMM is calibrated by Trescal and not by UKAS.
Or is Trescal issuing that certificate on behalf of UKAS?

The few calibration certificates I have seen that were issued by accredited cal labs here in Germany never carried a PTB logo.
If I red a term like "PTB calibrated" I expect the device to be calibrated by PTB directly.

Are there different habits in the UK?

Regards
try

[mzzj]

UKAS is the accreditation body/authority. "UKAS calibrated"  is bit misleading at best.
National calibration lab comparable to PTB would be NPL

German 17025 certificates would have DaKKS logo as it is the national accreditation organization in there.

'Murican "NIST calibrated" or "NIST traceable" is the most misleading one, usually it just means some "sort of" (claimed) traceability to NIST.

[dl1640]

it means the calibration procedure/calibrated item is accredited by accreditation body such as UKAS in  UK.
however the standards used must traceable to SI thru direct or indirect comparison to (inter)national standards, normally thru NMI or manufacturer e.g. Fluke.

[CJay]

It's just a lazy way of saying it has been calibrated by a UKAS accredited laboratory

[Neganur]

I can't read the address, does the zip code start with KY11 WFZ UK  (DUNFERMLINE ?)

Trescal definitely has some accredited laboratories but I can't find that particular cal shop in the official UKAS register.

Oh found it: https://www.ukas.com/wp-content/uploads/schedule_uploads/00001/0013Calibration%20Multiple.pdf

The capabilities are listed under location code "Donibristle".

This document tells you exactly what they are capable of offering in terms of accredited calibration.

And yeah the UKAS logo tells you that their cal shop has been officially accredited to ISO/IEC 17025:2005 standard by the appointed national accreditation body of the UK.

[Magnificent Bastard]

I can't read the address, does the zip code start with KY11 WFZ UK  (DUNFERMLINE ?)

Trescal definitely has some accredited laboratories but I can't find that particular cal shop in the official UKAS register.

Oh found it: https://www.ukas.com/wp-content/uploads/schedule_uploads/00001/0013Calibration%20Multiple.pdf

The capabilities are listed under location code "Donibristle".

This document tells you exactly what they are capable of offering in terms of accredited calibration.

And yeah the UKAS logo tells you that their cal shop has been officially accredited to ISO/IEC 17025:2005 standard by the appointed national accreditation body of the UK.

Wow... that lab has some very good CMCs--- k=2 uncertainty for 10V Zeners is 0.12ppm-- about as good as you can get without owning an on-site JVS (using a large bank of Zeners for the house standard).  I'm guessing that they are one step below NPL.  Some of their other specs are near primary lab numbers.  Not bad at all!

[tkamiya]

'Murican "NIST calibrated" or "NIST traceable" is the most misleading one, usually it just means some "sort of" (claimed) traceability to NIST.

Why do you say that?  NIST tracable is exactly that.  Equipment used to calibrate your equipment is listed on the certificate.  If you look at certificate for THAT equipment, ones used to calibrate THAT is listed and so on, all the way back to NIST itself.  What caliber of equipment must be used is clearly specified in NIST traceability specification.  Test condition and uncertainty is also listed.  I don't know how else it can be done, except take everything to NIST itself or buy your own test suite and directly test them with NIST.

[mzzj]

'Murican "NIST calibrated" or "NIST traceable" is the most misleading one, usually it just means some "sort of" (claimed) traceability to NIST.

Why do you say that?  NIST tracable is exactly that.  Equipment used to calibrate your equipment is listed on the certificate.  If you look at certificate for THAT equipment, ones used to calibrate THAT is listed and so on, all the way back to NIST itself.  What caliber of equipment must be used is clearly specified in NIST traceability specification.  Test condition and uncertainty is also listed.  I don't know how else it can be done, except take everything to NIST itself or buy your own test suite and directly test them with NIST.

Because you can slap "NIST traceable" text to any piece of paper without any control or standard.

Calibration uncertainty seem to be the weak point even on many ISO 17025 (A2LA) accredited calibration labs, seems like some american labs are still stuck with the 4:1 TUR and haven't really grasp the idea of  total calibration uncertainty.
Or how does 0.03cel  calibration uncertainty(k=2) for 0.1cel resolution fridge thermometer sound like? 

[dl1640]

IMO it is good and simple to keep TUR 4:1, however in real world it is very challenging, sometimes you get TUR 2:1 and that is guardbanding come into play.

[mzzj]

IMO it is good and simple to keep TUR 4:1, however in real world it is very challenging, sometimes you get TUR 2:1 and that is guardbanding come into play.

I'd rather not to see mention of TUR anywhere in my 17025 calibration certificates as there seem to be about 50% probability that cal cert doesn't fullfill ISO17025 requirements in that case. 
TUR is a huge pain in the ass if you are dealing in the 17025 "ecosystem", if the calibration certificate doesn't include total calibration uncertainty then its a guesswork to calculate  it from incomplete data and entire idea of traceability is questionable. 

The 0.03cel  calibration uncertainty(k=2) for 0.1cel resolution fridge thermometer example was from one of these. 

BIG difference between 17025 and "TUR calibrations" is that the the uncertainty calculation per ISO 17025 or EA4/02 needs to include also the uncertainty contribution of the calibrated device(DUT).

If the DUT resolution is 0.1 volts it already makes 0.1/sqrt(3) = 0.06 volt standard uncertainty and 0.12 volt expanded  (k=2) uncertainty
If the DUT reading last digit is jumping around ie +-0.1  that would add another  0.2/2/sqrt(3) to the calibration uncertainty.
So claiming 0.03 volt calibration uncertainty (k=2) would be very suspicious if the meter resolution is only 0.1 volts (or whatever units)

[dl1640]

Hi mzzj, if your certificate does not report *total* measurement uncertainty then it is not in compliance with 17025. The uncertainty of the standard used is one of the contributions for one particular calibration just as you know.

[mzzj]

Hi mzzj, if your certificate does not report *total* measurement uncertainty then it is not in compliance with 17025. The uncertainty of the standard used is one of the contributions for one particular calibration just as you know.

Yeah.  I have had bad luck with cal certs coming from (smaller) US based 17025 accredited labs.  Fluke and other "big boys" seem OK.
I don't see them too much around here but uncertainties in 5 out of 20 cal certs have been FUBAR in a way or another.
Like TUR given but no total uncertainty. Or reference uncertainty given instead of total uncertainty 
Or just plain wrong like 0.03 cel total uncertainty for fridge thermometer with 0.1cel resolution.

[dl1640]

Well, if the thermometer has 0.1C resolution then the measurement uncertainty reported on the certificate shall correspond to that resolution, the reported 0.03C is not in compliance with 17025 either. The 0.03C uncertainty is probably the calibration standard alone such as a known temperature source or a thermistor probe.