Difference between revisions of "TDIS DAQ, 7 August 2017"

From Hall A Wiki
Jump to: navigation, search
(→‎Agenda)
(→‎Minutes)
Line 23: Line 23:
  
 
== Minutes==
 
== Minutes==
 +
* Kondo:
 +
** Discussed GET system/AGET chip (see talk)
 +
*** similar to DREAM
 +
*** talked to F. Sabatie, thinks AGET better than DREAM for rTPC
 +
*** F. Sabatie is willing to put us in touch with Saclay developers, need to send our requirements so that they can help
 +
*** Not done yet
 +
**Also discussed VMM chip
 +
*** Signal peak ADC and tie information
 +
*** No waveform
 +
*** Might still work
 +
**Thinks best option is likely to work with Saclay to develop
 +
**PK comment: AGET would have same data rate constraint as DREAM, maybe VMM looks more promising
 +
*** but, peak time may be late (~100ns)
 +
*** need to find out if flexible/adjustable
 +
* Ed:
 +
** Discussed SAMPA chip (see talk) used in sPHENIX, STAR
 +
*** new ASIC for ALICE TPC
 +
*** all will use continuous readout
 +
*** ALICE 1 TByte/sec
 +
*** prototypes this year
 +
*** shaper permits sampling at reasonable rates, ~20 MHz possible
 +
*** 1.28 GByte/sec max off chip, need on-chip data reduction
 +
** NL comment: SAMPA may (not I think) be giving us an 80 ns peaking time, will go with 160 ns standard, a problem for us
 +
*** Sao Paulo group would be willing to go back for ~$100k
 +
** SAMPA continued:
 +
*** triggered - continuous mode possible, both time-stamped
 +
*** "heartbeat" triggers
 +
* Paul:
 +
** Discussed general constraints, ideas for data rate and choices (see talk)
 +
*** Trigger rate if we record all electrons in SBS ~6 kHZ
 +
*** What about time and amplitude only?
 +
*** What about only tracks that have good timing (would need at least some basic track finding)?
 +
*** How many time bins?
 +
*** Full reconstruction online (online processor approach)
 +
**** NL commented that ALICE is trying similar approaches
 +
*** Reduce 1-3 TByte/sec to 1-few GByte/sec via online cluster finding, etc.
 +
*** Larger experiments doing this kind of online processing (FPGAs even), factors of ~10-15 reduction
 +
* Discussion:
 +
** Nilanga/Bogdan, summarize detector changes discussed after SBS collaboration meeting
 +
*** 20 usec memory time higher than proposed (a problem)
 +
*** may not need as good momentum resolution
 +
*** how large does rTPC need to be?
 +
*** Rachel to estimate, study ~few to 10 cm gap
 +
*** '''talk for next meeting!'''
 +
** Design Option 2 - jet structure
 +
*** '''discuss with drawings at next meeting'''
 +
** Design Option 3 - multiple TPCs with E, B, z parallel
 +
*** 50 cm length divided into ~5 sections
 +
*** Also can use faster gas to get to 1 usec
 +
*** Kijun to simulate
 +
*** '''talk for next meeting!'''
 +
** '''Nilanga/Kondo/Bogdan to present sketches/ideas at next meeting!'''
 +
** Kondo also looking at readout of rate new ideas
 +
*** '''Discuss this as well at next meeting!'''
 +
** NL commented that hardware programming (data reduction, cluster programming,...) will also be important
 +
*** need to work with more groups doing this
 +
** JZ pointed to a set of simulation plots that he has done (see link) looking at resolution, for VIPs 60-400 MeV
 +
*** assumes 0.3 mm resolution on hit, 70/30 Ar/CO2
 +
*** 100 hits along track
 +
*** for current BONNUS12 tracking software
 +
*** ~3% momentum resolution, 1 mm in z
 +
** We need to know what kaon momentum resolution we will need
 +
*** '''k-TDIS will try to estimate for next meeting'''
 +
** PK had one other concern: background in physics region vs. worst? Will coordinate with Rachel.
 +
** KG asked if simulation covers ion flow-back also?
 +
*** could be important
 +
*** will need to simulate ions flowing back from the primary, also reduction of pressure/ionization density
 +
*** '''Put on the to do list!'''

Revision as of 10:07, 29 August 2017

Back to Tagged DIS run group >> DAQ development Meetings

previous meeting << >> following meeting


Organizational details

11:00 AM, CC A110

We will be using the BlueJeans system. To join via Phone:

  1. Dial: 1-888-240-2560 (US or Canada only); other access numbers are available from bluejeans.com/numbers
  2. Enter Conference ID: 491884270, and press #

Meeting URL https://bluejeans.com/491884270

Agenda

Minutes

  • Kondo:
    • Discussed GET system/AGET chip (see talk)
      • similar to DREAM
      • talked to F. Sabatie, thinks AGET better than DREAM for rTPC
      • F. Sabatie is willing to put us in touch with Saclay developers, need to send our requirements so that they can help
      • Not done yet
    • Also discussed VMM chip
      • Signal peak ADC and tie information
      • No waveform
      • Might still work
    • Thinks best option is likely to work with Saclay to develop
    • PK comment: AGET would have same data rate constraint as DREAM, maybe VMM looks more promising
      • but, peak time may be late (~100ns)
      • need to find out if flexible/adjustable
  • Ed:
    • Discussed SAMPA chip (see talk) used in sPHENIX, STAR
      • new ASIC for ALICE TPC
      • all will use continuous readout
      • ALICE 1 TByte/sec
      • prototypes this year
      • shaper permits sampling at reasonable rates, ~20 MHz possible
      • 1.28 GByte/sec max off chip, need on-chip data reduction
    • NL comment: SAMPA may (not I think) be giving us an 80 ns peaking time, will go with 160 ns standard, a problem for us
      • Sao Paulo group would be willing to go back for ~$100k
    • SAMPA continued:
      • triggered - continuous mode possible, both time-stamped
      • "heartbeat" triggers
  • Paul:
    • Discussed general constraints, ideas for data rate and choices (see talk)
      • Trigger rate if we record all electrons in SBS ~6 kHZ
      • What about time and amplitude only?
      • What about only tracks that have good timing (would need at least some basic track finding)?
      • How many time bins?
      • Full reconstruction online (online processor approach)
        • NL commented that ALICE is trying similar approaches
      • Reduce 1-3 TByte/sec to 1-few GByte/sec via online cluster finding, etc.
      • Larger experiments doing this kind of online processing (FPGAs even), factors of ~10-15 reduction
  • Discussion:
    • Nilanga/Bogdan, summarize detector changes discussed after SBS collaboration meeting
      • 20 usec memory time higher than proposed (a problem)
      • may not need as good momentum resolution
      • how large does rTPC need to be?
      • Rachel to estimate, study ~few to 10 cm gap
      • talk for next meeting!
    • Design Option 2 - jet structure
      • discuss with drawings at next meeting
    • Design Option 3 - multiple TPCs with E, B, z parallel
      • 50 cm length divided into ~5 sections
      • Also can use faster gas to get to 1 usec
      • Kijun to simulate
      • talk for next meeting!
    • Nilanga/Kondo/Bogdan to present sketches/ideas at next meeting!
    • Kondo also looking at readout of rate new ideas
      • Discuss this as well at next meeting!
    • NL commented that hardware programming (data reduction, cluster programming,...) will also be important
      • need to work with more groups doing this
    • JZ pointed to a set of simulation plots that he has done (see link) looking at resolution, for VIPs 60-400 MeV
      • assumes 0.3 mm resolution on hit, 70/30 Ar/CO2
      • 100 hits along track
      • for current BONNUS12 tracking software
      • ~3% momentum resolution, 1 mm in z
    • We need to know what kaon momentum resolution we will need
      • k-TDIS will try to estimate for next meeting
    • PK had one other concern: background in physics region vs. worst? Will coordinate with Rachel.
    • KG asked if simulation covers ion flow-back also?
      • could be important
      • will need to simulate ions flowing back from the primary, also reduction of pressure/ionization density
      • Put on the to do list!
Retrieved from "https://hallaweb.jlab.org/wiki/index.php?title=TDIS_DAQ,_7_August_2017&oldid=33607"