6.1.1 An Initial Failure of the Camera Electronics

  Our first images to come from the quick-look telemetry indicated that the normal CCD dark current pattern and familiar blemishes were not present. It was clear from the erratic character of the signal that there was a serious failure somewhere in the signal processing chain or the CCD itself. To our great relief, we discovered that if the camera was not operated continuously for more than a minute, the malfunction ceased. Evidently, if things had a chance to cool down (or, alternatively, something didn't acquire some slowly increasing electrical charge), the problem was suppressed. Thus, to overcome the camera failure, we changed our observing strategy from an initial plan of taking a long series of exposures followed by a long series of data dumps, to one where the exposures and dumps were interleaved. During each data dump the camera was turned off and allowed to rest.

The change in operating strategy had no adverse effect on our overall efficiency. We had originally planned to take all of our exposures during orbital night, with the data dumps occurring on the daylight side where the Lyman-$\alpha$ backgroud is higher. However, for the bright stars that we ended up observing, we found that this background was insignificant. Thus, we carried out observations during both day and night, with an increase in the availability of many our most important targets.

After the flight, we found a defective line driver chip that transmitted waveforms from the sync generator to the video processor circuits. Its failure in laboratory conditions exactly duplicated the problems we experienced during the flight.