Latest revision 15 August 1997

TJD 10670 = August 10, 1997
TJD 10671 = August 11, 1997

At approximately TJD/SoD 10670/60300, a calibration exercise began
which would last approximately 10 days and which consisted of cycling 
the LLDs on the Spectroscopy Detectors.  The LLDs were first set at
very low values and were raised once per day.  As a result, the
DISCSP channel 1 rates from approximately TJD/SoD 10670/60300 to
TJD/SoD 10671/59111 were so high that overflows occurred.  Because
there is only one flag bit in the telemetry for discriminator
overflows, no matter which type of detector or which detector module
caused the overflow, there is only one quality flag (30) which was 
set for effectively the entire interval.  Thus, analysis programs 
such as pulsar analyses which exclude data covered by quality flag 
30 will find very little data for TJDs 10670 and 10671 even though 
they might be concerned only with discriminator data from the 
Large-Area Detectors (DISCLA) and not from the Spectroscopy Detectors 
which caused the overflows.

The daily raising of the LLDs effectively caused the overflows to
go away although some continued on TJDs 10671 and 10672 but for
very short intervals, usually the typical two packets.