Minutes of E391a DAQ meeting

        December 13, 2001

Attendance
 KEK: M.Yamaga, G.Y.Lim
 Osaka: T.Yamanaka, Y.Sugaya, K.Sakashita, T.Ooba, Y.Ikemoto
 JINR: N.Khomoutov, G.Macharashvili

1. Specification 

  Trigger rate will be ~200Hz with 10MeV threshold for veto counter 
  at 4x10^12 proton intensity (from Sakashita san's calculation).
  Conversion-time of ADC 1885 will limit the possible acquisition rate.

  The data size will be 1MB/sec in average.
  -> 80GB/day -> 7TB for 300 shifts of beam time.

2. FASTBUS system

  1) Interface 
       SIS4100(NGF) was chosen to be an interface of VME-FASTBUS interface.

  2) One crate, or two crates?
       Two crates for now, due to power-supply module's limitation.
         - Need one more NGF and CPU. 
         - Readout scheme rather complicated.
       One-crate system is simple and will be better.
         - Need high-power-supply module. 
           Cost? Can we borrow one from somewhere?
     Not finally decided.

3. TKO system

  Two crate system is working fine. No problem.

4. Readout CPUs

  Both MIDAS and UNIDAQ can handle multiple-CPU scheme.

  Onboard SPARC54 with Solaris8 operating system is being used for now.
  Other choice ... BIT3 (or SIS) VMEtoPCI interface with Linux PC.
  Which will be better ? ... not finally decided.

5. DAQ Software

  MIDAS was chosen for further extension.
  Yamaga will be in charge of MIDAS with the help of N.Khomoutov san.
  UNIDAQ system might be prepared by Sugaya san if the MIDAS fails 
  to work by some reason.

6. Computer and network

  Local address will be assigned to all DAQ computers 
  so that the DAQ system would be separated from KEK network.

7. Data storage

  Data will be saved on our storage disk first, then will be transferred 
  to the KEKCC storage-server.
  Need a local DLT for safety as an emergency back-up system.
  The data will be distributed to other institutes from KEKCC via network. 

8. Trigger 

  Multiplicity logic will be enough for the first stage of the experiment.
  FPGA will be prepared for further extension.

9. High Voltage control

  MIDAS can do what we need about HV system.

10. Cable and grounding

  Grounding and noise reduction will be an important issue for our experiment.
  We will construct a cable-housing shielded with copper plate between 
  the detector-hall and counting-hut in order to shut the bothering 
  magnetic field off.

  In order to reduce the noise from the HV-supply module, we might locate 
  the HV crates inside the detector hall.

  For the trigger signal, we can use fast-transmission line such as '3DFB',
  which will be faster by ~10% than 'RG58'. 
  For the charge signals from veto counters, we can use cables with 
  better performance in noise reduction and attenuation, such as 'RG223'.
  Performance of them should be checked.  

11. Shopping list

  1) One high-power supply for FASTBUS (1~2 million yen), or 
     one NGF and SPARC54 (~1.5 million yen) 
  2) DLT storage system (~1 million yen)
  3) Cables : coaxial (~15 million yen) + twisted pair (~3 million yen)

12. Call for DAQ members

  We will make a mailing list for the DAQ discussion.
  Those who are going to join the DAQ group, or those who have an interest 
  in DAQ activity,  please let us know.