Telemetered vs Processed Events

In addition to the data for individual events, the HRC has scalars that report the rate of MCP triggers and the rate of triggers that pass on-board validity tests; these are known as the "Total" and "Valid" event rates. These rates are useful in determining the detector deadtime, particularly when the event rate exceeds the bandwidth available for telemetry. The telemetered values are recorded in the HRC level 0 hrcf*_ss0.fits file. During level 1 processing, the two 1.0 second samples of the total and valid event scalars for a given HRC science frame are summed and output into the hrcf*_dtf1.fits file in column names TOTAL_EVT_COUNT and VALID_EVT_COUNT. A count of the telemetered events whose time-tags indicate that they occurred during the 2.0 seconds covered by the scalar is also added to the hrcf*_dtf1.fits file as PROC_EVT_COUNT. As long as the rate is low enough that we are not telemetry saturated the value of PROC_EVT_COUNT should be nearly the same as VALID_EVT_COUNT since essentially all the valid events can make it into the on-board FIFO and out in telemetry. This appears to be true for the HRC-I; however, it is not the case for the HRC-S.

There is a significant fraction of events on the HRC-S that are counted as valid by the rate scalar that do not get telemetered. This fraction is a configuration dependent. A good data set to view the difference between the HRC-I and HRC-S and the effect of the HRC-S configuration is a sequence of calibration observations of PSR B0540-69 performed in 2000 June to evaluate possible solutions to the HRC timing problem. Table 1 provides a listing of the observations with the relevant configuration changes.

Table 1: PSR B0540-69 Observations - Sequential Order
DetectorObsIDAntico VetoEdge-blank SpecmodeTrigger Level

Figures 1-3 show the total, valid, and telemetered rates for this sequence of observations. The effect of the detector configuration changes can be seen on the rates: the total rate on the HRC-I drops when the trigger level is raised and it drops on the HRC-S when "imaging-mode is selected and the trigger level raised for Timing mode.

Total Event Rate
Figure 1: Total event rate for 2000 June 20-21 observations of PSR B0540-69 using the HRC-I and HRC-S under a variety of configurations. Dashed vertical lines separate the individual observations. The configurations for the longer observations are briefly indicated by text in the plot; see table 1 for details of these and the shorter observations.

Valid Event Rate
Figure 2: Valid event rate for 2000 June 20-21 observations of PSR B0540-69. Similar to figure 1.

Telemetered Event Rate
Figure 3: Telemetered event rate for 2000 June 20-21 observations of PSR B0540-69. Similar to figure 1.

Figure 4 shows the ratio of the telemetered event rate to the reported valid event rate. For the HRC-I, the ratio is near 1.0 as expected for the observed rates; the different appearance of the HRC-I default comfiguration is most likely due to the event time-tagging problem. The ratio on the HRC-S is much different from 1.0 without any noticible change in the rates. Selecting HRC-S imaging mode and/or raising the trigger level moves the ratio closer to 1.0.

Telemetered/Valid Event Rate
Figure 4: Telemetered/valid event rate for 2000 June 20-21 observations of PSR B0540-69. Note that for the HRC-S a significant fraction of the contributors to the valid event rate do not get telemetered.

Another possible HRC-S configuration change that we can look at is using the edge-blanking function for the selection of the individual MCP segments. Figure 5 shows the telemetered to valid rate ratio for a series of observations of G21.5-0.9 where this was done. We have performed this test three times and the data shown are for the first instance in 1999 September. The data were slightly affected by flaring in the background but the general result that the three MCP segments differ in the ratio is not significantly affected.

Comparison of Telemetered/Valid
Event Rate for the Three HRC-S MCP Segments
Figure 5: Comparison of the telemetered/valid event rate among the three HRC-S MCP segments from the 1999 September 5 observations of G21.5-0.9 (ObsIDs 147, 151, and 157). The individual segments were selected using edge-blanking regions.


One way that I can see to produce the discrepancy in the valid and telemetered rates on the HRC-S is if an over-shoot on sufficiently large pulses causes double counting in the total and valid rates. The fraction that are double counted would drop as the trigger level is raised. There might possibly be differences in the pulse shapes and possibility of over-shoot among the MCP segments that causes the different ratios between the segments.

The difference here would be yet another useful thing to investigate on the flight-spare detector assembly and POC electronics.

Last modified: Fri Dec 7 12:35:11 EST 2001

Dr. Michael Juda
Harvard-Smithsonian Center for Astrophysics
60 Garden Street, Mail Stop 70
Cambridge, MA 02138, USA
Ph.: (617) 495-7062
Fax: (617) 495-7356