A17.6 Pile-up Effects in the LAN CD Analog Array

Pulser tests were performed on the LAN flight unit using the test set-up shown in Figure A17-13. The 8020 pulsers were set for narrow (<=100 ns) pulses to simulate current pulses from the detectors. The Wavetek pulser was simply used to convert the triggers from the random pulser to the appropriate level for triggering the 8020 pulsers. This arrangement recovers from a triggered output within 100 ns and therefore allows for rapid re-triggering, a necessity for the investigation of pile-up and pile-down effects. The random pulser can put out triggers with random spacings down to approximately 600 nsec.

Figure A17-13 LAN pulser test set-up

Case I: Pile-Up

The 8020 pulsers were adjusted until a spot in W5 near the W5-W7 border was obtained in the DE by E plot. The results are shown in Figure A17-14. Since W5 and W7 are assigned different priorities, the relatively small number of pile-up events in W7 appear enhanced. Note that there are a significant number of events exhibiting pile-up in D with no corresponding pile-up in C. Pile-up in C begins only when the D pile-up exceeds a certain minimum value.

Figure A17-14 Random pulser test results for pile-up

These effects can be explained assuming the situation depicted in Figure A17-15. Here it is assumed that a "normal" event occurs at t=0 followed 1.75 msec later by a second event. Recall that the shaping time constants for the C and D channels are 0.5 msec and 1.0 msec respectively. The result, as shown in Figure A17-15, is that the C pulse tends to pile down while the D pulse piles up. It can be seen that a second event occurring anywhere in the interval 1 ms less than or approximately equal to t less than or approximately equal to 2 ms will tend to show the same general effect. Since peak detectors are used, the final result will be a range of pile-up values in D, and no effect on C. Note also that if the second event occurs <1 ms later, C will begin to exhibit pile-up as well, but only after D has exceeded a minimum value.

Figure A17-15 Pile-up in D, no effect in C

Case II: Pile-Down

The 8020 pulsers were adjusted until a spot in W5 near the W5-W4 border was obtained on the DE by E plot. The results are shown in Figure A17-16. Since W5 and W4 are assigned different priorities, the relatively small number of events in W4 were enhanced by fixing the priority to W4 during part of the test.

Figure A17-16 Random pulser test results for pile-down

These results will be analyzed in two parts. First, consider the situation depicted in Figure A17-17. Here, a "normal" event has occurred at t=0 and a second event is triggered 4.3 msec later. The peak detectors hold the first event for approximately 4 msec and are then reset so that they are ready for a new event at, say, 5 msec. Figure A17-17 indicates that for second events occurring 4.3 msec less than or approximately equal to t less than or approximately equal to 10 msec later, the D pulse will tend to pile-down noticeably, while C is only slightly affected. This tendency probably accounts for the cloud of points immediately below and to the left of the primary pulse in Figure A17-16.

Figure A17-17 Pile-down in D, little effect in C

Referring again to Figure A17-16, a trail of points can be seen emerging from the "cloud," moving essentially horizontally to the left, and then gently sloping downward through region W3. To explain this trail consider the situation depicted in Figure A17-18. Here the second event occurs 3.5 ms after the first. Again, the peak detectors are busy holding the first set of peaks until the 5 ms mark, when they are released from the reset condition to acquire a new peak. As shown in Figure A17-18, the C channel would tend to acquire the final remnants of the positive lobe of the second event, while the D channel would acquire a piled-down version of the D pulse. This would tend to give a data point toward the extreme left of the DE by E map. Note that as the second event is delayed over the range 3.5 ms less than or approximately equal to t less than or approximately equal to 4.3 ms, the positive lobe of the C pulse is rapidly "exposed" for acquisition at the 5 msec mark, while the piled-down D pulse only increases slightly due to its broad nature. The result would be a set of C values over the full energy range associated with D values over a fairly narrow range-hence, a gently sloping trail of data points.

Figure A17-18 Pile-down in D (similar to Figure A17-17) associated with a low value of C from the falling edge of the main lobe

Referring again to Figure A17-17 for the case where the second event occurs 4.3 ms later, it can be seen that the peak of the D pulse occurs after 5 msec while the peak of the C pulse is at 5 ms. Thus, for second events occurring at times slightly less than depicted in Figure A17-17, the D amplitude will remain virtually unchanged while the C amplitude changes significantly as it is acquired from various points on the slope of the positive lobe--hence a horizontal portion of the trail.

Conclusion: The above discussions have dealt with general trends which were noticed in preliminary pulser test data, and possible explanations for these trends. Although not discussed, the relative rate of occurrence of the pile-up/pile-down events seems consistent with these explanations.

Return to Appendix 17 Table of Contents

Return to HISCALE List of Appendices

Return to Ulysses HISCALE Data Analysis Handbook Table of Contents