DAQ Testing/20190123

Back to Main Page >> DAQ Documentation Portal >> DAQ Testing >> DAQ Commissioning Notes

Previous Day of Testing << >> Next Day of Testing

January 23rd, 2019 Testers: Cameron Clarke, Paul King, Ye Tian

Goals

• Verify JAPAN can connect to 64 bit ET system
• Understand FLEXIO output writing in CH crate for BMW test

JAPAN Online ET Connection

JAPAN is capable of running in live data collection mode as a ET client attached to a station. It needs to attach to the 64 bit ET, and should be run in non-blocking mode to ensure that data flow is unobstructed.

• Does the system connect to the ET?
• Does it write a ROOT file?
• Memory mapped root file writing can be turned on or off, and the files are written to /dev/shm
  • prminput/online_apar.conf contains the parameters needed to achieve this
    • 64 bit ET is "par2" located in /tmp/et_sys_par2 and "realtime" is the station name
    • The "enable_mapfile" flag turns on memory mapping
    • See ./qwparity --help for more flag options in .conf files
    • Run with: ./build/qwparity --config prex.conf --add-config online_apar.conf
  • Memory Mapped root file usage is non-trivial and requires using a TMapFile instead of a TFile, and SetMapAddress - so don't worry about this for now
• ET Connection with Online JAPAN
  • Run (after CODA activated and runs started) with: ./build/qwparity --config prex.conf --online --ET.hostname adaq3.jlab.org --ET.session par2 --ET.station realtime
  • Monitor the ET system with: et_monitor_64 -f /tmp/et_sys_par2 and look for the root file in root output directory
    • This "worked" in that it definitely connected and that there was a resultant root file, but the resultant root file was empty (see HALOG 3645013 for more information)
    • Run 1102

CH FLEXIO output bit control for BMW settings

The beam modulation (BMW) setup writes a FLEXIO output bit from within the CRL in order to send a trigger pulse to controllers in the Beam Switch Yard (BSY) service building. So, to test that this bit writing works Ye Tian and Cameron placed the FLEXIO output through a series of modules and input it into the SIS3801 scaler and read that out with JAPAN and with the scread scaler server Bob wrote.

• Run 1100 - no scaler data is picked up
• Run 1101 - put the scaler channel into the O-scope to look for it manually instead
• Run 1102 - worked with the online Japan ET system, no data written to disk
  • 30 Hz, no missed events, non-blocking, safe DAQ
• Run 1103 - Start with executing "flexioWriteChan(1,2), which has 0 scaler rate, then mess with this to see what makes the scaler signal true or false
  • Then check during CRL execution to see if the scaler rate returns to 200k level that was seen earlier
  • It did! This means that the flexio is being turned to flecioWriteChan(0,2) during CRL execution
  • Concern: 200k == always on scaler level, w/ or w/out the CRL being active
    • This means that the CRL is sending a "constant" pulse (or the brief scaler rate = 0 pulse is too short to see??)
    • I thought it was supposed to be the opposite (or I don't understand what constitutes a true or false scaler signal)
    • Line 442 of bmwClient.c appears to turn this signal off by default
    • Definitely the scaler is 0 volts applied = 200k counts during "off" and during CRL running, if I understand it correctly
• Run 1104 - return to injector mode

Conclusions

• JAPAN can connect to the 64 bit ET, but it needs more work in order to properly analyze and write the ROOT file
• The BMW system can control the FLEXIO word, so there is hope that it makes sense, but understanding the timing and pulse width of the FLEXIO signal sent to BSY building will take more careful attention (or not using a V2F since its timing resolution is probably too small to see the brief time the FLEXIO is sending its signal)