IceTop Low Level Verification

IceTop data for the period of  Jun 3 - Jun19 2006 have been processed
and analyzed using the latest i3 files.

The following verification plots are produced by selecting events with
IceTop Simple Majority Trigger.

The runs used for the plots are

DOM HV and Gains

  IceTop DOM HV/Gains  (for Jun3-Jun15 data)
         -- verified to be the same as in March 15 domcal6.2.4 calibration.
  IceTop DOM HV/Gains  (for Jun15-Jun19 data)
        -- verified to be the same as in June 15 domcal6.2.4 calibration.
        -- also shown in table the gain change with respect to March 15 calibration 

Note that
1- Most gains are way off  from their nominal setting.
     HG should be 5E6, LG should be 5E4.
     DAQ to do: adjust HVs to keep the gain at nominal values
2-  Some DOMs show more than 20% gain change between March and June calibrations.
     DAQ to do: run domcal more often

Charge Threshold

Although set to 10 PE in DAQ, the effective HG DOM thresholds are around 30 PE.
LG DOM thresholds were supposed to be set to 1000 PE, but they are around 4000 PE.
Why is this?
DAQ to do:  lower the LG thresholds to 500 PE

Charge correlation  in same tank

The scatter plots compare the charge from low and high gain DOMs both looking at the same event.
If the calibration is perfect the linear region between the DOMs upto ~3000 PE should
lay on the line y=x. However, LG DOM thresholds being ~4000 PE does not allow us to see the linear
overlap.

Charge Spectrum

The plots show the charge spectrum of each IceTop DOM.
The black histogram is the full spectrum. The red curve shows the contribution to the spectrum
of the waveforms from ATWD channel 0, the green is from ATWD channel 1, and the blue is from channel 2.
Note the dips around the channel cross over !
This is due to the baseline correction method used. The effect of the worst method is shown here.
The charge comparison of the ATWD channels clearly shows the problem.
Note the 60PE offset between channels 0 & 1, ~300PE offset between channels 1 &2.

So far the best method we found is to shift the baseline by the average of bins 87-126.
In this case the charge spectrum is continuous, and the  offsets between the ATWD channels
is reduced to ~few PE.

However, some DOMs have double lines.
This is not due to the bipolar behaviour of the baseline, because we adjust for it.
This must be due to calibration of the different ATWD chips. We are investigating this.