Re: Burst of ERP GC events found (fwd)

Kate Scholberg (kate@cithe501.cithep.caltech.edu)
Wed, 30 Aug 1995 10:43:27 -0700 (PDT)

>
> "VXMACB::vxmacb::michpub"@lngs.infn.it writes :"
> >
> > 10804-02747-1- 4-20-194- 8- 3
>
> Well, the tunnel's down so I can't check in more detail, but here we have 4
> bursts over 3 threshold in one SM. I suspect more of the same, ala the last
> few days. Kate, did you say yesterday that you _lowered_ the threshold??
>
> --

I lowered the probability threshold from 6e-6 to 3e-6: so it should
take a lower probability burst to generate an alarm (equivalent
to turning up the multiplicity threshold). I'll lower it some more
today (there was another LUT iteration yesterday, so rates may have
increased yet again).

[Note that a constant probability threshold doesn't necessarily
yield a constant rate of alarms: due to the overlapping nature
of the burst search time windows, the alarm rate depends on
both the event rate and the probability threshold.]

Kate.