This manual describes RAW level TRB analysis, on the TRB unpacker level (before HodoRaw or WallRaw).

change hsc.sh to use your path in the following line export MYHADDIR=/u/#username#/trb_analysis/

Set up your environment: . hsc.sh

compile macro: make

now check hodo_mitcuts.txt

WallRefHitPar and HodoRefHitPar are given there in the following scheme

Module	!UnpCutLow	!UnpCutLow	!HitfCutLow	!HitFCutLow
	(in chn)	(in chn)	(in ns)	(in ns)
0	-1e9	1e9	-1e9	1e9

Change the second and third column for the cut the UNPACKER is doing on the data while filling HodoRaw Level. These number above there mean no cut at all. Numbers are in TDC channels (100ps).

In our case one line should read: 1   34000  34800  -20   10 the last two numbers are cut in ns by the hitfinder AFTER the calibration.

run anaHodoUnpackerTRB_05:

(1M events starting with 0 and hodo_mitcuts.txt as parameter input)

./anaHodoUnpackerTRB_05 filename.hld 1000000 0 hodo_mitcuts.txt |tee logfile |grep Ev

the code will create a lot of output:

in logfile you will see all TRB errors and more. grep 805 logfile|wc will tell you how many error for TRB 805 you had in that file

the main root output file is written to /scratch.local/filename.hld.root It will contain HodoRaw, HodoCal, HodoHit (and same for Wall) categories. In addition there are debug- histogramms created.

this is now important

The hists are written to the root file and printed into postscript files in the directory where you stared the macro.

TRBDiagRaw_805.ps

This is the picture which i always plotted (logbook) BUT it is in TDC channels not in detector channels. Ploted is time_in_window versus channel. It runs from 0 to 127 for unprocessed data and from 128 to 255 for data which was accepppted by the time cut of the unpacker. This times are without Trigger Time Reference substration.

TRBDiagDif_805.ps

Now this is the same picture but now for the times after Trigger Time Reference substration. A fixed number (40000) is added to all times to get a positive number. This means that the Time Reference will be at 40000(!).

• example start raw:
  

Note: only channels which have a corresponding detector channel will be in the 128-255 range. This means the Time Ref itself will never show up there. If you change the lookuptable you can copy Time Ref to a dummy Detector channel, but still then it might be (very likely) out of the cut windows.

• example start dif:
  

Things to check:

Is the cut on the prompt peak O.K.? If not you have to change it and rerun the analysis!

Yes, then check if the Raw Time after Cut is "touching" the boarder of the window (which you can see on the left side of the picture in the raw time without cut). If yes, the windows is too small. If there is still place, windows can be changed. Take care about the Time Ref signal. The "width" of the blob is determinded by the CTU timing and the size is 2000chn!!!! Which is bad, because this means your data has to away 2000chn from the time window. Means: Data after cuts should not start before channel 2000 and the windows should continue 2000 channels afterwards. The Time Ref Signal has to be completely in the time window, too, if no seperate TDC is used for that. In Apr07 TRB 805 has its own TDC for time reference. Here the Time reference is allowed to be out of the windows for the TDC C.

other examples:

• example veto raw:
  

• example veto dif:
  

-- BjoernSpruck - 18 Apr 2007