Monitoring
Cameras
An interesting project
that I was involved in for more than 20 years was remote control of cyclotrons.
The use of video cameras was mandated, but high radiation resistant cameras and
lenses were extraordinarily expensive in those days. Even with their immense
expense, they were not guaranteed to work in the high flux fields we had to
endure.
The solution was
multiphased and took quite a long time to evolve.
At first we used standard
vidicon tube based industrial cameras (this was way before CCDs), fine quality
zoom lenses and Pelco pan-and-tilt motors.
After several cycles of
operation, the cameras failed, and were replaced with fresh units. Being
activated, each failed unit had to cool off in the hot box before I could handle
it on the test bench and determine the failure mode. Telephone calls to the
factory yielded no positive answers, but they did want me to report back to them
when I discovered a fix, as they said they knew I would. Thanks for the kudos,
but would have far preferred a quick-fix!
A problem arose in
that in most cases, after the activation had decayed away (we think the main
culprit was copper), the camera worked normally again. This was frustrating
and not very efficient, since the cameras tended to fail just at the time they
were needed most, during a cyclotron run.
The
solution was to record the video and note the failure modes, then determine the
individual chips that were being affected by activation. Fortunately there were
radiation-rated chips (ceramic vs. plastic) available for the IC's involved.
Swapping the chips out cured the problem, and then we had a truly inexpensive
camera that would work for a long time in the high flux environment.
Next problem was that the
glass in the lenses would slowly turn brown and eventually not be usable any
more. Instead of replacing the whole motorized lens assembly (expensive), we
opted to have on hand the individual lens elements, and just swap them out as
needed, on a schedule. Once brown the glass stayed that way forever. I guess
there might be some truth in the rumors of artificially created "desert glass".
For the most part, we
then had a stable monitoring system, backed up with an identical second system
in the same cell. Note that there never was a problem with passive components
like resistors or capacitors, just the active semiconductors.
When another cell was
being built for an additional cyclotron, we really looked into improving the
monitoring capabilities. Based upon experience we knew there were many critical
areas not visible from a camera mounted in a fixed location, even though it had
had pan-tilt-zoom capabilities. What was really needed was a traveling rail
system to augment the camera motions already provided. The traveling rail would
extend the length of the machine and really allow for optimum inspection inside
the closed off cell.
We had shop-built many
large scale industrial traveling rail systems prior to this project for several
clients, but this application was unique and required some out of the box
thinking.
The solution was a
pneumatic-over-magnet rail. Inside a sturdy sealed tube is located a cylindrical
permanent magnet, sealed with O rings on either end. At each end of the tube is
an air inlet. Through remote controls, the position of the magnet in the tube
can be changed by air pressure on one side or the other, resulting of smooth
positional control. On the outside of the
tube rode a yoke, also equipped with permanent magnets. This yoke followed the
internal magnet and rode smoothly back and forth on the tube. On this yoke we
mounted up the camera system.
This unit worked
impeccably, allowing remote viewing of the entire cell as needed. Still the
radiation problem persisted, being more of a nuisance than a problem at this
point, so a boron-block "doghouse" was constructed at the head end of the rail
system, allowing a safe "parking place" for the optoelectronics package when not
actually being used. The speed at
which the camera can be deployed/parked was remarkable compared to conventional
motorized rails.
Problem
solved.
George
Dowell
NLNL