[ RadSafe ] Neutron rem-ball calibrations

JPreisig at aol.com JPreisig at aol.com
Wed Mar 8 22:53:54 CST 2006


     This is from:    jpreisig at aol.com     .

      Hello Radsafers,

           Enjoy the Spring once it arrives.

           Basic information on neutron rem-balls and rem-meters is given in 
the book
"Accelerator Health Physics" by Patterson and Thomas and in the
Accelerator HP coursebook by Coissairt.  Under-responses, etc. are 
discussed in these books.  Basic detector calibration information is given in
articles by Eisenhauer and Schwartz and articles by Hunt in the journal
Health Physics and in other radiation protection journals.

          Hopefully, the neutrons you are looking at with a rem-ball do not
vary spatially or in time very much.  A rem-ball uses a single detector to
arrive at one data point per measurement.  I guess when making such a
measurement the rem-ball is integrating over the whole energy range it is
looking at.  A rem-ball is not trying to produce a neutron spectrum over
many energies.

          Still, if a rem-ball were trying to produce a neutron spectrum over 
25 to 30 energy bins (i.e. energy intervals), one would be doing so with only
one measurement.  The problem would be very under-determined

         A Bonner sphere set of detectors produces roughly 7 measurement
(detector) data points used to solve for a neutron spectrum of 25 or 31 energy
bins.  This problem is still under-determined.  A solution is arrived at by
mathematical unfolding (see Patterson & Thomas' book).  Some neutron
workers, I think, have found it necessary to arrive at a solution using
deconvolution, but my work hasn't gotten that far yet.  Deconvolution is
challenging to do properly, but is routinely done in scientific fields other
than health physics.  Seek answers in the literature, or on the web.

         I have written a Fortran program (boy does this statement date me)
with a name similar to CONVOLUTION, but it doesn't do deconvolution
or convolution at all.  I'd tell you all what it does one day, but NOT TODAY.
Maybe the folks at Brookhaven can nose around in my old software from the
Lab and discover what I was up to.  Neutron work in the near future could get
real lively again.  Back to the topic at hand.

        If you have one rem-ball/rem-meter, I hope its detector is still in 
condition.  In calibrations, the more detectors you have to calibrate, the 
total time it takes to calibrate them.

         Patience is necessary when performing calibrations.  Thus endeth
this e-mail.  Good Luck and keep an eye on room temperature and
humidity, season and perhaps even time of day.  If not, your calibrations may
not come out so good.

        The next Chandler Wobble/Annual Wobble peak will start in 2042
(+/- 7 years; the peak is rather broad).  Sunspots and the Milankovitch cycle
are superposed???  Northern Hemisphere polar bears may run out of some of
their ice sheet in 2042 or thereabouts.

        Have a nice day!!!!!

        Regards,      Joseph R. (Joe) Preisig, Ph.D.

More information about the RadSafe mailing list