Advices for working in ECR Mode

From Elsys Wiki - Help and Recources for TranAX and TraNET DAQ
Revision as of 15:54, 26 September 2017 by Rbertschi (talk | contribs)
Jump to: navigation, search

Tricks and Advices for working in ECR Mode with large number of events

Single versus Multi Channel ECR

ECR single channel mode has the power-full functionality that every channel can produce his independent trigger event. On high channel count systems this functionality can bring the system to their knees when trigger condition occurs on many channels almost simultaneously and in short intervals. In such condition multi-channel ECR gives better system stability and supports higher trigger intervals.

Use the Hold-Off parameter reasonable

The Hold-Off parameter defines the time after the trigger condition until a new trigger condition will be detected again. This parameter is especially useful in measurement on which the same impulse is seen on several channels but time shifted (e.g. AE measurements). For preventing the system to trigger several times on the same impulse on different channel, the hold-off parameter must be used and set to a time accordantly that all "echos" from the same impulse are inside this time interval. Since TpcServer Version 1.3.2 from 09/2011 the hold-off parameter is implemented fully in hardware and produces now much lower system load on high trigger rates.

ECR Retrigger

ECR Retrigger is a powerful option if you expect block length with different sizes. Each trigger event in a specific period after the initial trigger will make the measurement block longer. When the maximum block length is reached a new block will be generated. In high channel count systems, it is recommended to activate the "Enable Hardware ECR Retrigger" on the configuration website for having high system performance.

Working with BDF Files

During ECR or continuous measurements the data are stored in so called heap files. They have the extension .bdf for binary data file. BDF files are fully supported in TranAX and allow the same treatment in TranAX as TPC5 files. The advantage from TPC5 is his open format (can be opened in Matlab) and some performance gain during visualization with larger files.

At measurements with a large number of events, it became apparent that TPC5 has a much bigger overhead due his internal structure than BDF files which makes BDF the more power-full format. The overhead of TPC5 files is also a reason why saving measurements with a large number of events takes a while.

BDF files are available for post processing just after the measurement is stopped. No save procedure is needed. The BDF files are located in the path specified with the TraNETConfig tool or over the configuration website.

The line "Nr. of BDF Files" specifies how many files are kept in the BDF file folder. Once this number of files is reached, the oldest file gets deleted at the next measurement start.

Splitting up the measurement in several BDF files

For limiting the maximum file size and so the maximum number of events per file the autosequence in TranAX can be used. The condition is that a short interruption of the measurement is acceptable. The procedure is as follows:

  • Configure the measurement that it stops after a number of blocks, e.g. 10000
  • Configure “Nr. of BDF Files” to a high number e.g. 10000.
  • Setup an autosequence as follow:
    • Repeat FOREVER
    • Start Recording
    • Wait on EOR
    • Next

Do not visualize all channels during the measurement

The measurement is running independently from the visualization in TranAX, it even continuous when TranAX is closed.

Visualizing many channels in TranAX during the measurement puts significant load to the system, especially when there are many recorded blocks. Information for each block must be read from the hard drive while the drive is fully occupied with writing new blocks on it. As the writing process is favored, the visualization and navigation in TranAX can stock or even hangs.