cd /elog/
find . -name '*.txt' -type f -print0 | xargs -0 -e grep -n -e prescale

./2004/03/11/20860_1.txt:21:> effectively prescaled by 15). On friday I was quite convinced that the
./2004/03/11/20866_1.txt:12:Note that the Coda configuration "physl2t" is old, contains wrong prescale factors, and should not be used.
./2004/03/15/21080_1.txt:4:Changed the phys6 prescale to 1, from 30.
./2004/03/15/21080_1.txt:5:modified file: ts2_pro2004_pri_ps7.crl, the prescale[6] array element now set to 0 from 29. Then compiled:
./2004/03/26/21624_1.txt:14:Right now the prescale factor for singles is still 10, which is unfortunately high
./2004/03/16/21102_1.txt:2:I have changed the prescale factor for TOF singles (PHYS6 or trig==7) from 30 to 10.  The singles were only about 10% of the recorded events.  This change should add less than 30 events/sec to the total rate at 100mA, with a small decrease in the live time.
./2004/03/16/21105_1.txt:33:changed the singles prescale factor from 30 to 10, so we should get more 
./2004/04/06/22225_1.txt:9:6) Karen has made newphys_kd_BATS.settings.  It downloads and all MLUs pass verify.  PHYS4 counts BAT-TOF coincidences.  This trigger also restores the multi-TOF with Cerenkov events that we have not included since 3/4/04 (see elogs).  If there are multiple TOFs in a sector (where one is BATs), and at least one TOF in the other sector, it is still PHYS0.  So, if we want to run with BATs coinc triggers, we can.  Not sure what the prescale factors should be for hydrogen, we'll have to see.
./2004/04/06/22230_1.txt:19:_ note that the prescale factor for singles is still 10. If the deadtime 
./2004/04/06/22230_1.txt:20:  is <20% then please try to decrease the prescale factor for singles. 
./2004/04/06/22230_1.txt:23:* to put the single (trig 7) prescale factor to 3 do
./2004/04/06/22230_1.txt:26:* to put the single (trig 7) prescale factor back to 10 do
./2004/04/06/22230_1.txt:30:_ Hopefully we can run with a lower prescale factor and the BATS on
./2004/04/06/22231_1.txt:4:deuterium.  For this we can use a prescale factor, as long as we know 
./2004/04/06/22231_1.txt:30:>_ note that the prescale factor for singles is still 10. If the deadtime 
./2004/04/06/22231_1.txt:31:>  is <20% then please try to decrease the prescale factor for singles. 
./2004/04/06/22231_1.txt:34:>* to put the single (trig 7) prescale factor to 3 do
./2004/04/06/22231_1.txt:37:>* to put the single (trig 7) prescale factor back to 10 do
./2004/04/06/22231_1.txt:41:>_ Hopefully we can run with a lower prescale factor and the BATS on
./2004/04/07/22285_1.txt:1:As the deadtime at the begin of a fill is only 15%, I'm trying a prescale factor of 3 instead of 10 for trig==7, according to elog note 22230, beginning with next run.
./2004/04/07/22289_1.txt:1:below trigger distribution with prescale factors 10 and 3
./2004/04/07/22294_1.txt:11:Runs 6342-6344 6346-6347 6349 with prescale factor 10 for trig==7 (PHYS6)
./2004/04/07/22294_1.txt:12:6352-6353 prescale factor is 3.
./2004/04/07/22299_1.txt:11:prescale factor has been decreased from 10 to 3.  Ratio of events which 
./2004/04/09/22433_1.txt:29:singles have to be prescaled right now.
./2004/05/05/23592_1.txt:2:I have made a new RunControl configuration called cosmic.  It is just like l1trig (level 1 trigger, common strobe) but all prescale factors are set to 1.  This will let Chris Crawford get plenty of data where a TOF and a neutron bar in the same sector fire.  Those show up as PHYS6 (or PHYS5 for TOF11-15).
./2004/05/05/23594_1.txt:1:first run with new cosmic configuration (l1_trig without prescales)
./2004/05/15/24003_1.txt:1:made new trigger "newphys_kd_COSMICS" from "newphys_kd_BATS", with just the last two singles lines (CC+tof) removed.  a better idea would be to prescale it by a factor of 10.
./2004/05/26/24498_1.txt:32:-      20% dead at 130 mA so may want to increase prescale on singles
./2004/08/25/28118_1.txt:1:Townsend pointed out a problem with the Compton polarimeter this morning in which the ADC spectra became dominated by a spike at the pedestal position (channel 223) as shown below.  I replayed one of the suspect runs (10383).  The trigger rate for this run looks normal (~30 kHz) as shown in Compton analyzer's EPICS strip charts.  However the prescaled fraction of pulse shapes recorded in the digitizer don't look anything like a CsI pulse.  The scope "histogram" should look like a consecutive sequence of 256 CsI pulses, with each pulse sampled every 10 ns after a trigger for 160 ns.  Scope histogram shown below (from problematic period during run 10383) looks nothing like this.
./2004/09/27/29574_1.txt:12:	Karen will increase the singles prescale factor while we're down, and then 
./2004/09/28/29607_1.txt:2:Took cosmics run 11538.  Usual trigger (newphys_kd_BATS.settings).  RunControl configuration is cosmics, which is level 1 triggers and no prescale on any event type.  I think this is the way we want to take cosmics for timing, but I'll check with Michael and Chris.
./2004/09/28/29620_1.txt:2:Changed prescale factor for singles from 3 to 9 in ~/pro2004/coda/ts2_pro2004_pri_ps7.crl and rebuilt trigger supervisor readout list.
./2004/09/28/29620_1.txt:4:Run 11537 is the last (non-cosmics) run with the singles prescale set to 3.
./2004/09/28/29622_1.txt:24:	I have changed the singles prescale factor from 3 to 9, so the singles 
./2004/09/29/29655_1.txt:2:Increasing the singles prescale from 3 to 9 (singles are now 25% of the events instead of 50%) has decreased our DAQ dead time.  See first plot -- run 11529 (prescale factor of 3) is in black, run 11556 (prescale factor of 9) is in red.
./2004/09/30/29689_1.txt:2:Tried comparing cross sections for ABS running, before and after changing prescale factor.  Used d(e,e'p) and QE cross sections reported by show_deep_asym_all.C for this.  All runs from 9/29 showed 21.9 ev/C for Deep, 7.6 ev/C for QE.  This is 16% lower than the B shift from 9/27, before the prescale change -- 26.0 ev/C for Deep, 9.0 ev/C for QE.
./2004/09/30/29689_1.txt:9:These are within 3% or so.  Conclusion is changing the prescale factor has not changed the cross section (i.e., we handle the busy-inhibit of the charge properly).
./2004/11/10/31020_1.txt:21:for the input bits in the XMLU. (exept for the heavily prescaled singles 
./2004/10/01/29725_1.txt:4:11529 (before singles prescale changed; fill to 150 mA) -- black
./2004/10/01/29725_1.txt:5:11556 (after singles prescale changed; fill to 150 mA) -- red
./2004/10/01/29725_1.txt:6:11588 (after singles prescale changed; fill to 175 mA) -- green
./2004/10/03/29810_1.txt:10:are single-arm events (Trigger 7) and thus scaled down by the prescale 
./2004/12/01/31748_1.txt:2:Changed the prescale factor for trigger 2 (e,e'n) event from 1 to 2.  Run 13007 is the last run with a prescale of 1.  Since we're running hydrogen and trying to get more statistics on (e,e'p), the (e,e'n) event was deemed expendable at today's weekly BLAST meeting.
./2004/12/01/31752_1.txt:1:missing from elog but good run. Last before prescale factor change (trig2)
./2004/12/01/31756_1.txt:2:Starting with run 13009 karen changed the prescale factor for trigger 2. Elog was not running since yesterday after a change in blastrc. Problem fixed, elog runs well right now.
./2004/12/02/31784_1.txt:1:After prescaling trigger 2 (e,e'n) by a factor of 2, DAQ livetime at 175 mA has gone from about 83% to about 88%.  This is consistent with expectations based on trigger rates: trigger 2 was running at about half the total rate of 240 Hz, so the prescale dropped the overall rate by about 60 Hz.  The readout of an event makes us dead for about 800 usec, so 60 Hz would be about 5% dead.
./2004/12/02/31784_1.txt:3:See attached DAQ live time plot versus current.  Black is run 13007, just before the prescale change.  Red is run 13009, just after the change.  Green is run 13019 from early this morning, at a higher fill current.  With this change, runs last about 1.5 hours at 195 mA fill current, and the red needle is typically at least 430 C/hr.
./2004/12/03/31805_1.txt:29:- discussion on changing the prescale factor on trigger 2. Currently it 
./2004/12/03/31805_1.txt:56:- prescale trigger 2 by a factor of 2 to reduce deadtime
./2004/12/09/31967_1.txt:58:- prescale factor of 2 on trigger 2 introduced last Wednesday and 
./2004/12/28/32228_1.txt:6:TSL1A rate is about 50 Hz.  At that rate, without buffering the DAQ deadtime would be about 4%.  Ratio of BDCCT/DCCT is well over 99%.  Look at the data tomorrow for any loss of synchronization.  Also can try a different trigger file so that the TOF singles aren't prescaled, get a higher rate.