Re: NRC's time-to-count technology

[Keith <WELCH@CEBAF.GOV>]

We use an NRC system employing the time-to-count gm tubes in a 'global'
monitoring system.  The particular task is background (environmental) photon
monitoring.  The only quirks (alas, no quarks) we have seen with them is a
reduced tube lifetime, apparently due to the stress that the bouncing high
voltage has on the gm tube.  Not to nitpick your description of operation, but
the voltage is not varied between "off" and operating voltage, it's dropped to
around 250V after the pulse, then raised back to operating voltage (around
500-600v on our particular tubes) after the pulse is processed.  A rough guess
of mean time to failure for a gm in continuous use - at background - is about
three years. Obviously, that relates to some total number of counts that i
don't have on the top of my head, but I would guess it shortens the tube life
to about a third or half of the life of a "normal" gm tube. 

 As with anything more complex, I guess it's arguable that the system overall 
is more prone to failures, since it requires more gadgetry to keep it going,
but we have had no real problems. We have identified a characteristic "mode" of
gm tube failure. Since we record the data from these monitors in a history file
we can go back and look in detail at the results.  What we have seen is
"spiking" of the output, followed by what appears to be normal operation, then
another spike. The rates then drop to a higher than normal background rate.  
The odd part is that after the spiking, the monitor will calibrate fine 
(using typical calibration dose rates) but the ability of it to monitor
background is shot. It will no longer see small fluctuations in the few tens of
microrem range.

If you want to chew the fat about our experience with NRC and their systems,
feel free to email or call.

Keith Welch
welch@cebaf.gov
804 249-7212