[Halld-level1] [New Logentry] FADC125 (250) BUSY Distribution and High-Rate performance

[somov at jlab.org]

Logentry Text:
--
We made some studies of the BUSY distribution and high-rate performance of 
the CDC and FDC  (BCAL was studied earlier) with Ben several  weeks ago; 
I forgot to make a log entry.

Configurations:
--------------------

     - 4 CDC crates with FADC125  and two trigger crates

     - 10 FDC crates with FADC125 and two trigger crates 
        (the FDC test was performed later, after Cody updated 
         firmware)

     - Detector occupancy: 4 hits in each fadc  in each 
       crate ( 5.6 % occupancy)  simulated by lowing threshold
       below  the baseline.  Readout mode - short.

     - Trigger: TS front panel external waveform generator
        (random trigger used for high-rate performance) 

     - Standard DAQ settings for CDC / FDC, blocklevel = 40, 
       bufferlevel = 4. Busy was enabled from fadc modules.

Summary:
-------------

We were able to run CDC/FDC at high rates without 
crashing the system; at 100 kHz random trigger rate 
(5 %  detector occupancy), the live time constitutes 95 %.
Note, the expected trigger rate for high-lumi runs
is about 70-80 kHz
 
The data rate was about 50 Mb/sec from each crate. We ran  
DAQ with CDC for about 9 hours to check  stability. The  
performance was testes/optimized for various trigger rules 
and busy thresholds. We measured the BUSY distribution 
time and  processing times of fadc250/fadc125 as a function 
of the  number of hits. Below are some performance results


Performance table  (4 CDC crates)
--------------------------------------------

Rate (kHz)        Live Time  (%)      Data Size (Mb/sec from crate)

70                     98                    35
84                     97                    42 
100                    95                   46 
117                    90                   58
137                    89                   65 
160                    85                   72



Some plots:
---------------

Fig. 1: Check busy signal: busy is asserted after the second  
unprocessed trigger (BUSY THR = 2). Yellow - FP input triggers, 
blue - triggers seens by fadc, magenta - busy propagated to TD.

Fig.  2:  Busy distribution for runs with random trigger (BUSY THR = 3).
 
Fig.  3: DAQ screenshot for the CDC run at 100 kHz: rate

Fig.  4: DAQ screenshot (CDC):  live time

Fig.  5: DAQ screenshot (CDC):  data rate


---

This is a plain text email for clients that cannot display HTML.  The full logentry can be found online at https://logbooks.jlab.org/entry/3430281
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/halld-level1/attachments/20161018/226fc601/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: tek00006.png
Type: image/png
Size: 29061 bytes
Desc: not available
URL: <https://mailman.jlab.org/pipermail/halld-level1/attachments/20161018/226fc601/attachment-0005.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: tek0000.png
Type: image/png
Size: 34991 bytes
Desc: not available
URL: <https://mailman.jlab.org/pipermail/halld-level1/attachments/20161018/226fc601/attachment-0006.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cdc_busy1_2.png
Type: image/png
Size: 327519 bytes
Desc: not available
URL: <https://mailman.jlab.org/pipermail/halld-level1/attachments/20161018/226fc601/attachment-0007.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cdc_busy2_2.png
Type: image/png
Size: 312853 bytes
Desc: not available
URL: <https://mailman.jlab.org/pipermail/halld-level1/attachments/20161018/226fc601/attachment-0008.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cdc_busy3_2.png
Type: image/png
Size: 330921 bytes
Desc: not available
URL: <https://mailman.jlab.org/pipermail/halld-level1/attachments/20161018/226fc601/attachment-0009.png>