Fwd: Dilemma on which calibration to use

[Ccja@aol.com]

---------------------
Forwarded message:
From:	BKOBISTEK@prodigy.net (Bob Kobistek)
Sender:	medphys@LISTS.WAYNE.EDU (Medical Physics Listserver)
Reply-to:	BKOBISTEK@prodigy.net
To:	MEDPHYS@LISTS.WAYNE.EDU (Multiple recipients of list MEDPHYS)
Date: 97-09-01 03:35:59 EDT

Bill McCarthy wrote:
>
> This addresses Scott Dube's concern about whether to use NIST's calibration
> or the manufacturer's.
>
> NIST is the federal institution franchised to maintain the National
Standards.
> I believe it follows that: Right or wrong, NIST is right (until NIST states
> they are wrong, [which has happened a few times]).
>
> There is no reason a manufacturer's calibration is any more authoritative
> (if as much) than any other.

I didn't catch Scott's original question, but I think I can infer it
from Bill's answer.  If I am reading it right, and if Scott is talking
about air kerma calibrations, I'd like to add to Bill's remarks.

NIST maintains free-air ionization chambers which serve as our national
standard for the gray (formerly for the roentgen).  As Bill said, this
chamber is right, period, because it is our standard.  But to irradiate
this chamber and read its output, there are a number of uncertainties
that must be considered, and when they are properly combined, NIST
provides calibrations with a 2-sigma uncertainty of 1% for diagnostic
X-ray energies.

Now, the manufacturer, who also performs calibrations, selects one or
more high-quality secondary standard chambers and sends them to NIST for
calibration.  When the manufacturer computes his uncertainty, he
obviously has a lower limit of the 1% assigned to the uncertainty of the
secondary standard by NIST.  After considering uncertainties due to
charge measurement, beam energy, temperature and pressure measurement,
etc., the manufacturer's uncertainty is likely to be 2% or higher.

So the major advantage of NIST calibration is greater accuracy.  The
trade-off is price.  NIST charges ~$1,800 for the first energy, $900 for
each energy thereafter.  ADCL's and manufacturers charge considerably
less.

A caveat regarding manufacturer's calibrations is credibility of their
accuracy claims.  There is an internationally accepted method for
calculating uncertainties, documented in The ISO Guide to Uncertainty in
Measurement.  This is a statistical approach, and adherence to the
guidelines assures that everyone involved is comparing "apples to
apples."  When a manufacturer quotes an accuracy, who knows what method
they use to analyze their uncertainties?  Also, is the manufacturer
being realistic or optimistic when estimating some Type B
uncertainties?  Does the manufacturer participate in intercomparisons
with other calibration labs?  What kind of quality assurance does he
use?  If the lab is accredited, you have some assurance that things are
being done with some amount of uniformity and quality.  THe ADCL's all
participate in round-robins to assure consistency of calibrations.  Also
NVLAP accredited labs have to follow all the right guidelines, as do HPS
accredited labs.

The current issue of Medical Physics has a good article about
uncertainty analysis in air kerma calibrations, by Ibbott, et al.,
entitled, "Uncertainty of calibrations at the accredited dosimetry
calibration laboratories."


Bob Kobistek
National Physics Consultants, Ltd.
Cleveland, Ohio