DSM Data
I have written a kumac and fortran program to analyze DSM data written in DAQ events. (kumac,fort).
First data must be extracted from HPSS.
It is then converted to another format via Akio's program daq2trg. Then it is converted to an ntuple for sorting in paw with Trg2ntp.linux. See web log #249 in STAR Trig volume and my example scripts to intermediate file (careful with spaces, directory and file are separate inputs, writes to directory data below current directory) and to ntuple.
Results of DSM analysis for run 4079058
For this run I had set the trigger patch lookup table with the tcl script. I used pedestal=0 and lsb=0 for all channels resulting in sending the lowest 6 bits of 12 to the trigger patch part of the DSM. Here is the data as seen by the DSM.
What is shown is the lowest ~50 inputs (out of 90) to the DSMs. The DSMs take 10 12 bit inputs per board. 6 bits are highest tower from the patch. 6 bits are the patch sum. A 2 D histo of each is shown.
Next is the same histos modified a bit. Before this run Falk had identified some channels that seemed to be floating at full scale (ones at 63 in above histo). He masked these in trigger. In the plot below I have assigned 64 to channels that are unused this year. I have assigned 65 to channels that Falk masked. What is seen is that the second board is obvioulsy screwy.
One thing to notice in the trigger patch spectra is that the pedestals are much noiser than one might expect from the spectra we have seen or from the high tower spectra. This warrants further investigation. In particular we need to simulate what we exept this spectrum to look like from DAQ ADC values. Renee started this.
First STAR data triggered on EEMC
Early 3/22/03 Falk and the shift crew set up to run using a high tower trigger (testeec). This corresponds to triggering on anything >9 in the above spectra. He did however mask out all of the second DSM board. The rates were about 45Hz with no prescale factor. There is a very low fraction of minbias included. Our run #2003. STAR run 4081031. (our runs 2000-2002 and STAR 022-031 are tests of the trigger. Probably only the latest are really useful.) Here is data from minidaq. We should analyze from STAR files when possible.
Crate 5 spectra Note: Looks like a sticky bit in upper channels gave a lower threshold.
