Re: Industrial radiography and computers

["Gary T. Pepper" <gpepper@capitalnet.com>]

At 11:10 AM 3/6/1998 -0600, you wrote:
>Hi All - 
>
>We are going to have a number of exposures taken of floors and walls in one
>of our science buildings.  The purpose of the radiography is to determine
>the location of cables in the post-tension cement beams, to avoid them
>during a drilling operation.  We are contracting a local firm to perform
>the radiography. 
>
>There has been some mention of removing any computer CPU's and magnetic
>storage devices (disk drives) in the near (40 ft') vicinity of the exposures.
>
>Do you know of any references that I can research?
>Has anyone personally experienced problems with computer storage loss by
>radiography exposures?
>
>(The source that will be used is a ~65 Curie Ir-192 source)
>
>Thanks in advance for any responses - please reply to my e-mail below
>unless you thing it interesting to the entire list.
>
>----------
>Ken Smith
>Acting Radiation Safety Officer
>University of California
>Santa Cruz, CA  95064                   
>Voice: (408) 459-3911
>Fax: (408) 459-3209
>e-mail: ksmith@cats.ucsc.edu 
>


Radiation effects on electronics, especially on electronic systems (such as
computers) are difficult to predict, because they can depend on such
factors as radiation type, energy spectrum, dose-rate, total dose, inherent
shielding present, etc, etc.  The effect of ionizing radiation on a
semiconductor device or circuit can also vary greatly, depending on whether
power had been applied to the circuit during exposure.  In a relatively
complex electronic system, such as a personal computer, there are many
different semiconductor devices, representing different semiconductor
technologies (e.g. CMOS, bipolar devices, etc) and fabrication processes.
In general, most semiconductor "families" and fab processes exhibit varying
degrees of sensitivity to ionizing radiation.  Even within a given family
of semiconductor devices, there is often a great variation in device
sensitivity to ionizing radiation.

To sum it all up, there is no "safe" method to predict the total dose (or
dose-rate) for which your computers would fail (i.e. the electronics and
storage media).  Why take a chance?  Remove the computers from the areas
where the exposures are to take place and that way you won't loose any
sleep over it!  ;-)

Note that there are numerous databases which contain experimental data on
the effects of ionizing radiation on semiconductor devices.  In general,
these are of little use for assessing your problem, as they wouldn't
contain many of the devices that currently used in a typical desk-top PC.
You would also have to be very careful when interpreting relevant data from
these databases, because you can easily be led astray.  For example,
suppose you find total dose failure data for part 'X' (from a given
manufacturer) in a radiation effects database.  Suppose the data is 6
months old.  It may no longer be applicable for the same device 'X', that
you buy off the shelf today, from the same manufacturer!  The reason is
that semiconductor manufacturers often "upgrade" their fabrication process
to improve reliability and yields.  Any change in fabrication process can
result in a change of the radiation response of device 'X'.  There are
often differences in the radiation response of device 'X', obtained from
different manufacturers too.

As you can see, it is best to leave Pandora's box closed for this problem;
in this case I think it would be prudent to forget the analysis and/or
search for relevant data and just move the computers!

Best regards,

Gary Pepper, Ph.D.
Radiation Physics Consultant
Ottawa, Ontario, Canada

e-mail: gpepper@capitalnet.com