RE: 120-ml and tool monitor

["William E. Cox" <COXWX@naesco.com>]

>Glen Vickers wrote:
>Shouldn't the "probability of detection" be 1.7 sigma, which would equate to
>95% confidence?  We use the 95% confidence level to be consistent with the
>quality levels of our formal Ge unconditional release counts.  We use this
>level of confidence for all free-release monitors.

The tool monitors use adjustable probability of detection and probability of
false alarm values, both of which combined with the actual background and alarm
level dictate the count time and provide a go/nogo alarm. I was not suggesting
that a specific value should be used, only what my recollection of a particular
setup was as it pertained to an approximate geometry.  The actual values one
uses for these settings must have a justifiable basis.

One could argue that a frisker is used for unconditional release surveys and it
only has between a 50% and 80% probability of detection. Therefore different
settings may be appropriate for different contamination monitoring applications.
Another example are whole body personnel contamination monitors. It has been my
experience that typically these are not set to a 95% detection confidence level.
I've seen values between 0.8 and 1.3 sigma used combined with probability of
false alarm values between 2.4 and 3.1 sigma in these applications.

I stand by my original statement that if one were to use a tool monitor for this
application it could be done, however specific testing would need to be
performed, some gamma mix assumptions made and that one may not be able to meet
ODCM specific limits within the maximum 100 second count time allowed by the
firmware.

Bill Cox
Seabrook Station HP
coxwx@naesco,com


************************************************************************
The RADSAFE Frequently Asked Questions list, archives and subscription
information can be accessed at http://www.ehs.uiuc.edu/~rad/radsafe.html