DCH-DCT-GLT Combined Calibration

Last modified: 26 January 1999 by G.Sciolla

!!! Warning: these instructions are preliminary!!! 

The instructions you will find in the following are addressed to sub-system DAQ experts or semi-experts (read it: people who took BaBar data at least once in their life!).

If you are not familiar with the BaBar DAQ system used for the Cosmic Run, please read the IR2 Shifter's DAQ Manual available on the web at /BFROOT/www/Detector/Operations/shifts/BaBarRunManual.html.

Where to work:

• Both event level and control level should run on bbr-odf06 or equivalent machine in the same subnet until odf-farm01 is fixed.
• On each window type the following : 

• • cd $BFROOT/detector/online/Production
    source babarScripts/setup.csh

• If you whant to monitor what's happening on the ROMs you have to open the xyplexes on each ROM you want to monitor. They are called

• • odf-dch and odf-pp1 for the DCH
  • odf-lt for the GLT 
  • ....... for the DCT.

  If you are running from the BaBar Console, then you don't have to open all the xyplex windows by hand. Just use the runConfig script after having modified the file babarScripts/RomConfig.txt as needed (see instructions in the Shifters Manual).

How to tun Oep:

• From one of the bbr-odf06 windows type

  • runOep
    

This asks which subsystems you wish to enable for the run, and calculates the appropriate crate mask based on this information. It will print this information in a line looking like:

              ***********
Crate mask is * 0x3800  * <== Type this number in ORC window during Initialize
              ***********

Note the crate mask, as you will need it for Run Control .

OEP will open a separate window titled testOepFSimpleShmReader.

OEP is now ready and waiting for data.

How to tun Oep:

• From another window on bbr-odf06 type the command: 

  • runOrc

  This starts the Run Control processes and brings up the Run Control GUI. The run control should persist for the duration of integrated running.

  The runOrc comand also brings up a cmlog window on screen bbr-con07:0.1 (the middle right screen of the bank of nine monitors that you're staring at), if one wasn't there previously. To enable cmlog to receive error messages, pull down the File menu and select Connect. Then pull down the options menu and select "Update". A dialog box will pop up asking you to enter a string; just click the OK button.

How to Start a Calibration Run

Click on the RUN_CONTROL button. You'll get pull-down menu, and select "INITIALISE". A dialog box will appear asking you to select a crate mask. Type in the value calculated by the runOep script.

Click the RUN_CONTROL button again, and select "MAP PARTITION" from the pull-down menu.

Now select the type of run you wish to take, using the pull-down menu under the RUN_TYPE button. Select "DCH-TRIG-CALIB".

Next click the RUN_CONTROL button and select "CONFIGURE". A dialog box will appear asking you to enter a trigger mask. Type 0x0.

After a few (up to 60 sometimes...) seconds the message just below the RUN_CONTROL button should read "CONFIGURED". If instead it says "ERROR" look for error messages in the xyplex and bbr-odf06windows. If a ROM has failed to configure, contact an expert from the subsystem and notify them of the problem.

Otherwise, cleanup the mess (see instructions below) and try again.

After a successful configuration, click on RUN_CONTROL one more time, and select "BEGIN". You are now taking calibration data. The run will continue until you will decide to stop it (see following instructions).

During a data taking Run

• Look at the xyplex windows and make sure there is some activity going on.
  
• Look at the testOepFSimpleShmReader window. For each damaged event received, you'll see a summary giving (among other things) the size of the event, and the damage value. There are three common damages:

• 0x2000 = Some part (maybe all) of the detector was out of synchronization. This is normal at the beginning of the run after some of the ROMs were booted. You should see a message:

  • *** received SYNCH occurence

  in all xyplex windows.

  0x100 = Some part of the detector was late to deliver the data for this event and was ignored.

• Check that the data is being recorded. Look at the most recent file by typing:

• • ls -lt theData/*.xtc | head -n 1

  and make sure it's getting bigger. The file is named babar-timestamp.xtc where timestamp is some string which is not easily interpreted by a human viewer.

At the End a Calibration Run

• Run Control goes back to Configured. Then: 

1. Click on the same button and select "UNCONFIGURE".
2. Click again and select "UNMAP".
3. Click once more and select "CLOSE PARTITION"

NB: It is extremely important to go all the way through the above steps. Look at the RUN_CONTROL button now and make sure it reads "INITIAL_STATE". Failing to do this may result in a serious mess

Prepare the data for staging

There is one manual step needed to give the data a BaBar run number and prepare it for staging to tape. From the Production area, do:

1. pushd theData
   number_runs.sh
   popd
   

When you finish your calibration (very important!!!)

At the end of your shift you should clean up the DAQ processes so that the next shifter can start fresh. To do so:

• Hit Return in the window where OEP is running . You should see an exit message printed.
• In the same window, type:

• • cleanup
    killOrc
    

  You don't need the killOrc if you ran fcgui and didn't use Run Control.

  The Run Control GUI should vanish.

• Next, release the xyplex windows by typing Ctrl-] in each, followed by quit at the telnet prompt.

The system is now clean for the next user.

For comments/questions about anything in the above contact Gabriella Sciolla <sciolla@slac.stanford.edu>