Monitoring of the on-board Triggers

When a transient event triggers the on-board logic, starting the HTR data acquisition, it is called ``on-board trigger''. To date (Dec 2001), more than 16,000 on-board triggers have been collected out of $\sim$ 80% of the whole archive currently available: if one supposes that the on-board trigger rate is not significantly different for the observations still missing in our GRBM archive, then it comes out that there should have been at least 20,000 triggers on board, roughly corresponding to a mean rate of $\sim 10$ on-board triggers per day. Obviously, the majority of these transient events cannot be due to cosmic GRBs: actually, out of them, only $\sim$ 4-5% are due to GRBs, the remaining being caused mainly by passages of high-energy charged particles through the GRBM NaI(Tl) slabs (see section ).

Given that, and since the GRB on-line quest algorithms work only on the 1 s ratemeter data, thus initially ignoring whether at a given 1 s bin, the on-board logic has been triggered or not, it comes out that the majority of the on-board triggers have not fulfilled the off-line search conditions. Nevertheless, it is important to monitor the 1 s ratemeters in every energy band, for every detector unit, at the different on-board trigger times. To this aim, the automatic quest for every on-board trigger, that is automatically discarded, a set of SNRs (Signal to Noise Ratios) is given for each unit, for both the energy bands, like in the table reported below.

In particular, from this table, the case of the on-board trigger occurred at UT 17:58:09, on April 11, 1998, has all null SNRs: in this case, the meaning is the following: since this time bin is in the nearby of a data gap (in this case, the SAGA), the search algorithms could not monitor it. For this, such on-board triggers would deserve to be visual inspected, in order to check whether they are true cosmic GRBs or simply particle events. For instance, on a sample of 15,162 on-board triggers, the automatic GRB quest performed with an integration time of 1 s, could not monitor 1032 on-board triggers, corresponding to $\sim 6.8$% out of the initial set.

[4]

-------------------------------------------------------------------------------
Date        UT	                GRBM1 GRBM2 GRBM3 GRBM4   AC1   AC2   AC3   AC4
-------------------------------------------------------------------------------
11 Apr 1998 13:53:43:32229       -0.0  -0.9   6.1  19.0   1.5   0.9   1.0   1.5
11 Apr 1998 16:42:59:20396       -0.2  -0.2   8.9  15.2  -0.1   0.5  -0.3  -1.5
11 Apr 1998 17:06:44:26649        1.2   0.0   6.0  13.2   0.9   0.9   0.9  -0.1
11 Apr 1998 17:58:09:39145        0.0   0.0   0.0   0.0   0.0   0.0   0.0   0.0
11 Apr 1998 19:34:55:62030        0.7   0.2  16.9   3.4   0.6   1.6   2.0   1.5
12 Apr 1998 00:30:29:74881      152.5  -1.2  -0.3  11.1   5.2  -0.5   0.4   2.0
12 Apr 1998 03:29:58:05174        0.8  -0.7   4.0   6.6   1.0  -0.8   1.1  -1.2
12 Apr 1998 07:13:44:14013        0.3   1.4   4.3   2.4   0.4   0.8   1.1   1.9
12 Apr 1998 08:56:02:01455       -0.8  36.5   0.7  28.6  -0.7   0.3  -0.8   0.3
12 Apr 1998 10:12:59:00598        1.6  56.7   3.7   0.3   0.3   0.1   0.3   0.1
12 Apr 1998 11:38:24:79494      183.4   2.2  -1.1  16.7   6.3   0.4  -1.2  -1.0
12 Apr 1998 18:53:09:34063       -1.2  -1.7   2.3  29.5  -1.2  -1.0  -0.8  -0.1
12 Apr 1998 23:37:39:98215       -1.3  -0.3   4.2  -0.4  -1.4  -0.3   0.1  -0.3
13 Apr 1998 00:26:11:83365        1.0   0.2  -0.3   1.5   0.2  -0.1  -0.3   2.0
13 Apr 1998 05:12:11:20591       -0.0   4.2   4.5   0.8  -0.1   1.5   0.9  -0.9
13 Apr 1998 05:26:21:89340       -0.8   0.2 157.0  11.9  -2.6  -0.4   7.3   0.5
13 Apr 1998 05:54:02:09651       -1.0   3.7   1.1   0.8  -0.3   2.6   0.2   0.2
-------------------------------------------------------------------------------

All the other on-board triggers reported in the above table look like due to particle events, since they show significant signals in only one or two units in the GRBM band, with no significant counterpart in the corresponding harder AC band (actually, they are discarded, because they do not fulfil the HR condition -eq. -). The fact that for many of them there is a clear signal in only one of the GRBM unit is not in contrast with the fact that the on-board logic, requiring that the trigger conditions have to be matched in at least two units, was triggered by it: actually, the signal is calculated only a different short integration time (SIT) on board with respect to the 1 s quest. Thus, although on shorter timescales the on-board trigger thresholds may be overcome, nevertheless by integrating on 1 s time bins, especially when the excess is caused by charged particles, the count excess can be overwhelmed by the background fluctuations.

On principle, among the on-board triggers reported above there are some cases (e.g., those at UT 00:30:29, and 11:38:24, on April 12) for which we cannot immediately exclude the possibility that they could be due to activity from any SGR: in fact, also these events do not look so bright in the AC band as in the softer GRBM band, differently from GRBs. Nevertheless, it seems unlikely that these excesses could be due to e.m. plane waves because, aside from the arrival direction, they exhibit very different SNRs between the GRBM units; for this property, they look like being due to local interactions within some detector units, like particles do.