FIFO occupancy study of GLAST front-end electronics

Contents of this page

• Method
• Result
• FIFO occupancy distribution (PDF file; 21 kB)
• Acknowledgement

Method

back to top

1. IRF files created with Glastsim
   for noise occupancies of 1.0e-4, 1.0e-3, and 1.0e-2

Three IRF files were created with noise occupancies of 1.0e-4 (baseline), 1.0e-3 (high noise situation), and 1.0e-2 (extream case). Source flux was given by "backgndmix", or a mixture of cosmic rays, albedo gamma-rays, and so on.

For this study, Glastsim was locally modified so that noise hits do NOT create any trigger. So, in this study, noise hits do not increase trigger rate, but do increase event size. This situation corresponds to the one in which noisy strips are masked from a trigger line with a trigger-mask, but left unmasked from a data line.

2. List of FIFO depths created by gsc_shell
   for source flux of x1, x3, and x10 background rate
   and for calorimeter latchup of 20 us, 10 us, and 0 us.

Gsc_shell read an IRF file created above and simulate read-out electronics with source flux at original background rate (which gives about 5 kHz L1T rate), 3 times higher rate, and 10 times higher rate. Since calorimeter "vetoes" L1T after a trigger, it affects FIFO occupancy distribution. To see the effect, calorimeter latchup time was also set to 20 us (baseline), 10 us (shortest possible), and 0 us (extream case). For each event triggering, gsc_shell outputs its FIFO depth, which is the number of events in FIFO after the event in question is filled into the FIFO. It ranges from 1 to 8 in this study.


Gsc_shell was also locally modified for this study so that it can simulate trigger rate increased. Given a factor for source flux it devide a time stamp of an event by the factor to result in increase of source flux by the factor.

3. FIFO occupancies calculated from the list created above



Based on output files of gsc_shell, FIFO-depth distributions were computed for each combination of noise occupancy, source flux, and calorimeter latchup time, with all the 16 towers summed together. From the distribution, averaged FIFO occupancy and maximum FIFO occupancy are also computed.

Those results are listed below. Also, the distributions are available below as a PDF file.

Result

back to top

calorimeter latchup = 20 us (baseline)

Source Flux	noise occ. = 1.0e-4			noise occ. = 1.0e-3			noise occ. = 1.0e-2
	L1T rate (kHz)	Averaged FIFO depth	Maximum FIFO depth	L1T rate (kHz)	Averaged FIFO depth	Maximum FIFO depth	L1T rate (kHz)	Averaged FIFO depth	Maximum FIFO depth
x 1	4.73	1.015	5	4.80	1.013	5	4.73	2.962	8
x 3	11.92	1.049	8	12.01	1.059	8	5.43	4.743	8
x 10	25.30	1.216	8	24.27	1.392	8	5.64	6.126	8




calorimeter latchup = 10 us (shortest possible)

Source Flux	noise occ. = 1.0e-4			noise occ. = 1.0e-3			noise occ. = 1.0e-2
	L1T rate (kHz)	Averaged FIFO depth	Maximum FIFO depth	L1T rate (kHz)	Averaged FIFO depth	Maximum FIFO depth	L1T rate (kHz)	Averaged FIFO depth	Maximum FIFO depth
x 1	5.00	1.023	7	5.04	1.032	5	4.86	3.282	8
x 3	13.51	1.085	8	13.70	1.166	8	5.44	5.697	8
x 10	31.72	1.418	8	28.67	2.429	8	5.64	6.439	8




calorimeter latchup = 0 us

Source Flux	noise occ. = 1.0e-4			noise occ. = 1.0e-3			noise occ. = 1.0e-2
	L1T rate (kHz)	Averaged FIFO depth	Maximum FIFO depth	L1T rate (kHz)	Averaged FIFO depth	Maximum FIFO depth	L1T rate (kHz)	Averaged FIFO depth	Maximum FIFO depth
x 1	5.27	1.059	7	5.34	1.098	6	4.98	3.495	8
x 3	15.72	1.239	8	15.86	1.492	8	5.43	5.635	8
x 10	38.94	2.119	8	29.21	2.900	8	5.66	6.643	8

Acknowledgement

back to top