SPS test details

In this section each of the tests run by SPS is detailed.

1. The Run Header Check
   A difference between the run header time, written to the data structure from the Mac (sharc?), and the first event time taken from the 10MHz clock, may indicate a problem with the 10MHz clock. If the difference between these two times is greater that then value specified in TSPS, a warning is given This check is implemented in the routine, sps_time_check.for.

2. Event Separation Check
   This check fails if adjacent events are separated by less than 410ns in time using the 50MHz clock. Events should not be closer in time than this due to the width of the trigger window, and such a scenario would incicate a problem with trigger lockout. This could affect the timing of events and the calculation of livetime. This check is implemented in the routine,

   sps_time_check.for.

3. 10MHz Clock Check
   This check fails if the 10MHz clock time does not step forward each time the event GTID increases. For disordered events (GTID decrease) a check is made between the 10MHz and 50MHz clock times. A count of the number of times the 10MHz clock stops and the number of disordered events for which the clock disagrees is written to the output file. This check is implemented in the routine, sps_time_check.for.

4. The Rate Check
   T he integration period for calculating the rate is specified in TSPS. It is possible to specify two different integration periods, one for the first x seconds of the run, where x is also specified in the titles bank, and another integration period for the rest of the run. The threshold for an unreasonably high rate is specified through titles, as is the allowed fraction of the run above this threshold. The rate check will fail if the number of integration periods which are above threshold, divided by the total number of integration periods is greater than this fraction. If the run time is less than the first integration period the rate check will fail. This check is implemented in the routine, sps_time_check.for.

5. The PGT Rate Check
   This check fails if the average PGT rate deviates by more that a given fraction (as specified in TSPS ) from 5Hz. As with the rate check two integration periods are specified, but this time the period is specified in terms of the number of PGT events. If the run time is less than the first integration period the rate check will fail. This check is implemented in the routine, sps_time_check.for.

6. The Sync Clear Check
   This check is implemented in the routine, sps_sync_clear.for. It checks for synch clear errors in induvidual channels which could be indicative of electronic problems but are expected for the first integration period of a resynched run.

7. The QLX bit Check
   This fails if the QLX bit is not set to long for all channels. This check is implemented in the routine, sps_gtid_rollover.for.

8. The Occupancy Check
   Over an integration period specified in TSPS, the tubes with zero occupancy are recorded. The allowed number of zero occupancy tubes of each type, OWL, BUTT, NECk and Normal, are also specified in the titles bank. If any of these thresholds are exceeded in an integration period, the test will fail. The output of the test gives the number of failures (ie the number of integration periods for which the failure occured) and also the maximum number of zero occupancy tubes in one period. This check is implemented in the routine, sps_occ_check.for.

9. The Trigger Check
   This check tests the ESUM high trigger occupancy. This check has two main purposes. First, finds the mean crate occupancy for ESUMHI triggers. If a crate is a certain percent (where this percent is specified in TSPS) then the check fails. This portion of the check is implemented in the routine, sps_trig_check.for. Second, the check finds the event rate of ESUMHI triggers. If this rate is too low (which again is specified in TSPS) then the check fails. This portion of the check is implemented in the routine, sps_time_check.for. Currently, the trigger check looks for abnormalities in the ESUMHI trigger only.

10. The AMB Check
    This check fails if the AMB integral pedestal is more than a given percentage of the average rate (both values input in TSPS). A warning is produced if the pedestal is between a given warning threshold and the failure threshold. This check is implemented in the routine, sps_time_check.for.