Response Tests of the PMTs
The camera of Solar 2 consists of XP2280/B photo multiplier tubes (PMT), and a number of tests have been done here at UC Davis to determine the behavior of the tubes and their Winston cones so that a calibration would be possible. The tests were done in a specially constructed enclosed box, and measured the output voltage of the PMT when exposed to light from a pulsing LED (light emitting diode). The box, when fitted with a collimated LED, allowed for horizontal scans of the PMT and Winston cone at different angles.
The Winston cones for the PMTs are in four different sizes, depending on the row which the PMT focuses on. Except where stated otherwise, these tests were done using cone 20.
One of the first difficulties found was a time dependence of the measurements. While the testing apparatus was on, the measured response for identical positions and orientations would decrease at a variable rate. This was likely due to rising temperature in the LED, but, as it remains unquantified, it introduces an error of as much as 10 % to the response plots. However, as this appears to be a property of the testing apparatus, and not the PMT, it has not been fully explored. In the plots shown, curves from the same plot are of data taken in the same sitting, so error on a single plot should be smaller than the illustrated effects. However, error between plots is expected to be fairly large, so comparison between them using the vertical scale may be misleading, or even contradictory.
A somewhat startling result was found when the PMT was rotated 180 degrees around its axis, so that it could be scanned upside-down, or inverted (see sketch). Plot 1 shows the marked discrepancy between the two orientations. The results of rotating the box itself indicated that the most likely cause of this was the Earths magnetic field, as seen from Plot 2. As no fitted magnetic shields were available for the PMT, an old shield from a slightly larger tube was used to see if the effects could be eliminated. Plot 3 shows the results of these tests. No large discrepancy was seen when the PMT was rotated by 90 degrees in either direction, as seen in Plot 4. All of the data taken up to this point was using a blue LED; results of scans with a red LED are in Plot 5 and Plot 6.
Testing of the PMT using the experimental set up, as well as data preparation and presentation, was done by Michael Barnard, a physics graduate student at UC Davis.