I thought this was an interesting thread:
-----Snippo-----
>>> Greg Gogates 23/12/1998 17:36:41 >>>
Date: Wed, 23 Dec 1998 10:17:07 +0200
From: L Hards
Subject: Standardisation of procedures
With regard to the international recognition of calibration certificates (in terms of MRA's / MOU's, etc) and harmonisation of measurements, laboratories being accredited in terms of ISO Guide 25 by a recognised accreditation body and these measurements are usually accepted as being traceable to a National or Internationally recognised measurement standard. This is an ideal situation, but I tend to feel that the wheels fall off when it comes to the actual measurements performed on UUT (Unit Under Test). Manufacturers recommended calibration procedures, in house (laboratory) procedures exist, but are there any internationally recognised standards / criteria / procedures that an accredited laboratory should perform to ensure that the UUT meets the manufacturers accuracy specification or the clients requirement. The situation arose when comparing the calibration certificate of the same instrument calibrated by twenty or so accredited laboratories. All the laboratories claimed they were using their professional expertise in selecting what measurements to perform, hence seventeen different selections (3 labs followed the manufacturers recommended procedure as required by the client). How is the problem addressed internationally?
Kind regards
Luke Hards
icon.co.za
-----Snippo-----
Subject: Re4: Standardisation of procedures (fwd)
Date: Thu, 24 Dec 1998 10:15:53 -0500 (EST)
From: Eddie Tarnow
Subject: Re: Standardisation of procedures
Hi All,
This is something about which I have very strong views so I'll try to keep my reply short and to the point.
1 Any measurement having a direct impact on product quality requires accuracy traceability. This is irrespective of any accreditation or certification.
2 Accuracy traceability is transferred to an instrument/product through the process of calibration/measurement using calibrated references and calculating an applicable uncertainty of measurement.
3 Accuracy traceability can however only be satisfactorily transferred in one of two scenarios.
A) If the measurement is performed at the exact point at which the instrument was calibrated the calibration uncertainty can be used in the uncertainty budget.
B) If the instrument is used to perform a measurement at a point other than that at which it was calibrated, then certain assumptions have to be made about the performance of the instrument at points not directly calibrated. The uncertainty in this case is usually assumed to be the instrument manufacturer's accuracy specification.
Now if the instrument has been calibrated according to the manufacturer's recommended procedure, the abovementioned assumption is very low risk.
However if the calibration has been performed according to a procedure designed based on the laboratory "professional expertise", this assumption could be placing the instrument user at great risk.
The practise of calibration laboratories performing "partial calibrations" at a significantly reduced price due to the time saving in only doing half a job, will continue until instrument users and ISO 9000 auditors alike become aware of this implication and consequently ensure that calibrations are fit for purpose.
The normal argument justifying the added risk for the benefit of the lower calibration cost was partially acceptable in the days when measuring instruments were far more accurate than the product tolerances necessitated. Unfortunately, as technology advances, measurement instruments no longer have test uncertainty ratios of 10 or more to 1 due to technological and financial constraints. The risk therefore is now greater than ever before.
In conclusion, customers, ISO 9000 auditors and accreditation bodies should in my opinion be placing pressure on calibration laboratories to ensure that "contract review" has taken place and that the documented technical requirements of the customer are met by the calibration laboratory.
While responsible calibration laboratories will continue to calibrate test equipment manufactured by reputable companies, according to their recommended procedures, what do they do when not even the manufacturer has gone to the trouble of resaerching and providing an acceptable calibration procedure????
Kindest regards and best wishes for a Blessed Christmas and peaceful and prosperous new year to all you metrologists and quality practitioners.
Eddie
Eddie Tarnow
Manager: SADC Resource Centre for Metrology Education (SRCME)
NML CSIR
South Africa
Tel: +27 12 841-3138
Fax: +27 12 841-4458
Web:
http://www.nml.csir.co.za
Standardization of Procedures
Linear Mode
