From grave@ipno.in2p3.fr Thu Jun  1 07:50:34 2000
Date: Thu, 25 May 2000 09:04:54 +0200 (CEST)
From: Xavier Grave <grave@ipno.in2p3.fr>
To: pate@nmsu.edu
Cc: marchand@ipno.in2p3.fr, bimbotl@ipno.in2p3.fr, arvieux@ipno.in2p3.fr,
     furget@in2p3.fr, beise <beise@physics.umd.edu>, sellem@ipno.in2p3.fr
Subject: Re: French electronics configuration

Hi Steve,

There was already a Data stream for the orsay electronics
(cf icd-det-elec-comp-230400.ps)

This is the same scheme except that I add a few 32 bits of information to
separate the MPS number and the general information (status + program
number : frontal or concentrator one)

----
coda header ?
----
32 bits : status + # board 0
32 bits : status + # concentrator program number
32 bits : inner concentrator counter (count the MPS)
32 bits : status + # frontal_0 program number
32 bits : inner frontal counter (count the MPS)
1024Bytes DSP 0 spectra
        2 spectra + 2 buddy spectra, 128 channel each ,1ch = 16 bits
32 bits : status + # frontal_1 program number
32 bits : inner frontal counter (count the MPS)
1024Bytes DSP 1 spectra
32 bits : status + # frontal_2 program number
32 bits : inner frontal counter (count the MPS)
1024Bytes DSP 2 spectra
32 bits : status + # frontal_3 program number
32 bits : inner frontal counter (count the MPS)
1024Bytes DSP 3 spectra
32 bits : status + # board 1
32 bits : status + # concentrator program number
32 bits : inner concentrator counter (count the MPS)
32 bits : status + # frontal_0 program number
32 bits : inner frontal counter (count the MPS)
1024Bytes DSP 0 spectra
32 bits : status + # frontal_1 program number
32 bits : inner frontal counter (count the MPS)
1024Bytes DSP 1 spectra
32 bits : status + # frontal_2 program number
32 bits : inner frontal counter (count the MPS)
1024Bytes DSP 2 spectra
32 bits : status + # frontal_3 program number
32 bits : inner frontal counter (count the MPS)
1024Bytes DSP 3 spectra
and so on...


This gives a 32KBytes spectra information + 44*8=352Bytes for each MPS
So we need a 993600.0Bytes/sec bandwitdh

And with DMA on the VME Bus I reach at least 6MBytes/sec.

With such a scheme we can check program consistency for all the board and
that we don't miss any MPS signal. In fact the counter have to give all
the same number for each board and each DSP.

             xavier

> Subject: French electronics configuration
> 
> Dear Jacques
>       Since the French electronics are programmable, we need to have a
> method to know which program is loaded into the DMCH16X.  Is there a
> simple way to do this?  For example, could the DMCH16X insert an
> identification word into the datastream?  The identification word would
be
> different for each version of the DMCH16X program.  In this way, the
> analysis program would know which histograms to expect (how long, what
> size time bins, etc.) in the French data buffer.
>       The "identification word" is just one idea, there are probably
> many ways to do this.
>                                               yours,
>                                                       Steve

grave@ipno.in2p3.fr
33 (0) 1 69 15 79 59

                  ,/{}
                ,/  {|
            ,,,/    {|,
      __--~~        {| ~-,
__--~~              {     `\
                        ,__ \
                       `,\{),\,
                      __-~  `_ ~-_
                   _-~        ~~-_`~-_
                  '               `~-_`~-__
                  `,                  `~-\_|
                   `,      _-----___    _,'
                   / /--__  ~~--__  `~,~
                    /     ~~--__  ~-',
                   /            ~~--'