Checking 50k Plots
This page describes what to look for when evaluating the 50k replay plots to check HCal's performance.
Where to Find HCal 50k Plots
After each production run a replay of the raw data on 50k events will be performed and the results posted to the HALOG under the title '50k replay plots for run XXXXX'. In this log entry will be an attachment called 'summaryPlots_12628_hcal_50k.pdf' which contains the plots described here.
What to Look for in HCal 50k Plots
This section will describe the plots for the HCal 50k replay and what one should look for when checking them. Shifters don't need to read the full description of the plots, but they should check that they look like the example plots to their right. These plots were made for the SBS-11 kinematic and will look different at other kinematic settings. Note that sometimes plots are added or removed from this output so each plot may not always be exactly where described here.
Summary Plots Page 1
The plots for page 1 of the summary plots are shown in Figure 1A. The top left plot labeled 'Cluster ADC time over all channels : HCAL' shows the fADC times of the clusters found in HCal (an energy cut is applied). The fADC window is currently 160 ns. The events coincident in HCal and BBCal are calibrated to occur in the fADC window at about 60 ns. There should be a Gaussian beginning to form around 60 ns (it's not a large bump at higher kinematics). If this structure is missing there may have been a trigger time or latency change that the experts need to address. However, if there is a larger peak at ~100-120 ns (Figure 1B), the first peak may still be visible as well, this is an LED trigger being captured in the data and does NOT indicate a latency or trigger change.
The top right plot labeled 'Amplitude vs Module (saturation check): HCAL' shows the a spectrum of all the fADC waveform amplitudes (in mV) vs. the PMT module number (from 1 to 288). The fADCs for HCal use a 2V range. Signals above 2V indicate that the module saturated for an event. Saturated events are generally set by the analysis code to list their values at about 4V, so one will see a small band of saturated hits at 4V. It is expected to see the fADCs saturate occasionally, but the bulk of the data should be below saturation. This becomes an issue if any channels are saturating very frequently. Experts may need to adjust the PMT module HVs in these channels or otherwise address the issue.
The bottom left plot labeled 'ADC amp vs int : HCAL' shows the correlation of fADC waveform amplitudes for each PMT module vs the fADC waveform integral per module per event. The fADC amplitude should be correlated roughly linearly with the fADC integral, so you should see a generally linear correlation on the 2D plot. This is another way to check for saturation issues or other anomalies.
The bottom right plot labeled 'ADC amplitude vs module ZOOM : HCAL' is the same as the top right plot except it is zoomed in on the Y-axis. One should see a roughly even distribution of fADC amplitudes between all 288 PMT modules, with more of the events near 0 mV (the fADC pedestal). There may be some variance due to geometry and HV settings, but no channel should have events only at a value of 0 mV. This would indicate that PMT has a bad connection or lost HV.
Summary Plots Page 2
The plots for page 2 of the summary plots are shown in Figure 2. The top plot labeled 'Cluster TDC over all channels : HCAL' displays the TDC time of all the clusters found in HCal. This should look like a plateau with a small peak on top at about -60 ns. This peak is the coincidence events between HCal and BBCal. If this peak shifts there may have been a trigger or latency shift that experts need to address.
The bottom plot labeled 'Cluster TDC vs Module : HCAL' shows the TDC time of the clusters found in HCal vs the PMT module (from 1 to 288). What should be seen is a band of TDC times for each PMT in the region of about 25 ns to -125 ns.
Summary Plots Page 3
The plots for page 3 of the summary plots are shown in Figure 3. The top left plot is labeled 'Number of clusters per event : HCAL' and shows the number of clusters belong to a given event. This plot should look Gaussian and be centered around 35 for SBS-11.
The top left plot labelled 'Number of blocks per cluster : HCAL' shows the number of PMT modules present in each cluster. At SBS-11 this should be centered around 5 and look slightly Landau in form.
The bottom right plot labelled 'Cluster E vs Multiplicity' shows the cluster energies on the X-axis and the cluster multiplicity (number of clusters per event) on the Y-axis. This plot should show that the multiplicity increases with cluster energy.
The bottom right plot labelled 'Cluster Energy : HCAL' shows the energy of the clusters detected in HCal. It should have a Landau-like shape and peak around ).2 GeV for SBS-11.
Summary Plots Page 4
The plots for page 4 of the summary plots are shown in Figure 4. The plot labelled 'Position of best cluster : HCal' shows where the best (highest energy) cluster is located on the surface of HCal for each event. The key thing to check is that all of the modules in HCal have some number of hits and aren't empty. If you see empty (white) blocks that probably means that the HV to some HCal PMTs has switched off (this often happens in groups of 12), or these is some other connection issue in the signal path. At higher kinematics, like SBS-11, this plot doesn't show much useful clustering information. All that can be seen with this low elastic rate and high background is that there is a slightly higher concentration of clusters on the beam side (right of plot) of HCal.
Summary Plots Page 5
The plots for page 5 of the summary plots are shown in Figure 5. The top plot labelled 'Number ADC events vs module : HCAL' shows how many fADC events were recorded for each of the 288 PMT modules (X-axis). The important thing to check is that the number of events for each module isn't zero. This would indicate a channel has a loose connection or the HV is off. It is fine if there are channels with fewer events than others. This is due to geometry and acceptances which vary over HCal.
The bottom plot labelled 'Number TDC events vs module : HCAL' shows how many TDC events were recorded for each of the 288 PMT modules (X-axis). The important thing to check is that the number of events for each module isn't zero. This would indicate a channel has a loose connection or the HV is off. It is fine if there are channels with fewer events than others. This is due to geometry and acceptances which vary over HCal. It is known that channels 84 and 178 are low (this will be investigated during an access).
Summary Plots Page 6
The plots for page 6 of the summary plots are shown in Figure 6. This plot labelled 'Pedestal vs element ID : HCAL' shows the fADC pedestal events for each of the 288 PMT modules (X-axis). The Y-axis is in volts, and we expect fADC pedestals to be around 100 mV. There will be a spread of pedestal events above approximately 100. This is fine as these are events only partially in the leading edge of the fADC window. The key thing to check is that no channel's pedestal has moved significantly.
Summary Plots Page 7
The plots for page 7 of the summary plots are shown in Figure 7. This plot labelled 'BBCal-HCal Cluster Correlation' shows a plot of cluster row location in HCal and BBCal. At higher kinematics like SBS-11 correlations are difficult to see without many cuts. At SBS-11 just check to be sure the events are somewhat concentrated in the center of the plot and there are no empty blocks.
Summary Plots Page 8
The plots for page 8 of the summary plots are shown in Figure 8. This plot labelled 'BBCal-HCal trigger difference' shows the timing difference in ns between the HCal and BBCal triggers. There should be a peak visible around 500 ns since one trigger arrives 500 ns after the other.