Duration time in each file (not run!).
There is some period where the prescaling was applied (around 1.8e7 UTC (red points)). The duration time for the period is simply reduced by factor of 2.
Duration time in each run.
The maximum duration time for a run is 29000 [s] ~= 8 [hr].
The cumulative duration time. The effective time increases linialy with time, and ~61.7 days in total (so far). There is some period where the prescaling was applied (around 1.8e7 UTC). So, the slope is different for the period. Flasher data are not included (run# 108895, 108896, 108911 and 108912).
Event # Vs Run ID
The event rate. It's about 1.3 Hz. There is some data with low rate... There is a correlation between event rate and the duration time (See above plot.(left)). This has to be checked. As I wrote above, this low rate was due to a prescaling to 2 at that time.
The event rate for each Npe cut.
black: no cut
green: log10(total Npe) > 3
blue: log10(total Npe) > 3.5
red: log10(total Npe) > 4
Strangely, the rate for high Npe is also lower than the other usuals.
This is again due to the prescaling.
I simply multiplied the rate by factor 2 for the prescaled
period (red points).
Now, the rate during the prescaling becomes higher than others. (It turned out this is because I didn't apply the prescale cut correctly.)
The event rate for each Npe cut after simple prescaling correction (multiply factor 2).
black: no cut
green: log10(total Npe) > 3
blue: log10(total Npe) > 3.5
red: log10(total Npe) > 4
Prescale correction for the prescaled period (green points).
Now, the rate during the prescaling is consistent with others.
The event rate for each Npe cut after prescaling correction .
black: no cut
green: log10(total Npe) > 3
blue: log10(total Npe) > 3.5
red: log10(total Npe) > 4
Time difference between events. The slope value is consistent with the rate. There is some tail (mainly due to data of prescaling period). The slope is -1.25.
Same as left. With log10(total Npe)>3. The slope is -0.123. This is also consistent with the rate.
Same as left. With log10(total Npe)>3.5. The slope is -0.0057. This is also consistent with the rate.
Same as above without prescaled data. (cut 1.76e7 < UTC <1.83e7). The tail becomes smaller. (Although, I should probably clean data more.) The slope is -1.30.
Same as left without prescaled data. With log10(total Npe)>3. The slope is -0.128.
Same as left without prescaled data. With log10(total Npe)>3.5.
The slope is -0.0059.
Time difference between events.
Black: all data
Green: prescaled data (1.77e7 < UTC < 1.82e7). (slope: -0.69)
Blue: normal data (UTC < 1.77e7, UTC > 1.82e7). (slope: -1.30)
Time difference between events.
I cut the prescaled data and run#108895, 108896, 108911 and 108912 that showed low rate in above plots.
They are tagged as good by the verification group, but
the online filtering page says the DB had some problems at that time.
It tunrs out they are flasher runs.
After the run removal, the time difference distribution agrees well
with the poisson distribution.
