Frdric Lefvre Subatech

1
Alice Central Trigger Processor Status

• Followed by DAQ systems for test beams

2
Recent history

• User requirement document
• Draft 1st june 2000, 1st december 2000,1st
  february 2001 -gt rejected March 5th 2001
• Lack of trigger input
• Lack of trigger  classes 
• New version 24th october 2001,
• Accepted by the TB in february 2002

3
What s new ?

• Number of inputs increased
• Number of sub detectors and trigger dead time are
  fixed
• Introduction of the so called  detector
  cluster 
• Some others number (scalers, output to the daq )
  are fixed

4
Trigger input

• 24 L0 20 L1 and 6 L2 inputs
• Spare today 6 (7) L0 4 (5) L1 and 6 L2

5
Sub detector

• Definition independant TTC branch
• Number limited to 24.
• Spare number is small. Look for merging

6
TTC tree

• Each tree can deliver a trigger independantly
  from the others
• Each tree carries its own trigger messages
• Grouping possible for detectors that are ALWAYS
  fired together (eg all triggers)

7
Trigger classes

• 50 classes. Each is defined by
• 1 detector cluster, set of L0, L1and L2 pattern,
  Past Future protection and pre scaling

8
Detector clusters

• Group of detectors (in a TTC tree definition of
  it) that are read out together
• May receive different L1 messages but receive L1
  together.
• 6 clusters (no more)
• Clusters are defined at each run.

9
Example
10
Calibration software trigger

• Requested by software
• ASAP or at a specified bunch crossing
• Calibration trigger during the LHC long gap
• Prepulse L0 L1 L2
• Past future protected -gt a request may fail

11
Calibration Timing
12
Past Future protection

• Interaction defined by a subset of 4 detectors
• A-A 2 kinds peripheral and (semi) central
• Counters and threshold for each kind of event
• Our demand to have asymetric P/F was rejected.
• Counter and threshold on p p runs

13
CTP Output

• Cu cable L0
• TTC tree (one per detector)
• A channel L1
• B channel L2 (a r) L2 words L1 Words
• DDL -gt Daq
• Each event L1 word L2a words interactions
  records.
• On request scalers, debugging memories (to be
  defined in more details)

14
Trigger type Word(proposals)

• L1 8 bits
• Trigger type (physics or software/calibration)
• 4 bits Readout mode (detector specific)
• Enable segmented readout (for tpc only)
• 2 bits spare
• Alternative 5 bits readout mode, no trigger type
• Readout mode function of the 50 classes
  bitpattern. To be defined by sub-detector
• Please make it not too complicated, thank you.

15
L2a message
16
AOB

• L2a may change if the trigger classes bit pattern
  is requested.
• 60 bits of date instead of 48 -gt 3 more TTC
  transmissions
• Sub detector provides the cables and connection
  to the CTP
• Yes it does mean cost.

17
Worries

• 1.2 µs delivery time of the L0 to the Front end
  electronic is still not granted.
• Wont discuss here about consequences
• Implementation of the CTP is really a challenging
  task.
• See above
• TTCrx receiver boards for the dimuon.
• TTC channel B overcrowded

18
Daq for beam test _at_ CERN

• Current system
• Date V3.7 on VME Motorola boards running AIX 4.3
• Stay there and alive at T10 up to end 2002.
• Is able to read the Si tracker system
• Future system
• DATE V4 on
• Concurrent Technologies VME boards (Intel PIII)
• or PC running Red Hat 7.1 linux
• Should appear at T10 4th quarter 2002.

19
CCT VME Board

• Intel processor, Linux Red Hat 7.1
• Not yet available in the dimuon collaboration.
• We should invest some money in.

20
What kind of DAQ do we need in 2003 ?

• DDL, no tracker
• Use linux PC, PCI RORC and DATE 4
• No DDL and tracker
• Use CCT VME, and DATE 3
• DDL and tracker
• Use Linux PC, 1 PCI RORC and DATE 4
• CCT VME for the tracker
• One GDC.
• No DDL no tracker
• That should exist in our lab.
• Not a joke, see GIF test f.i.

21
What kind of support from CERN ?

• Piotr will reduce his care of daq test beam
  system.
• We are safe up to end 2002 but after
• What kind of manpower can we invest in there ?
• -gt ask CERN to continue support on the test beam
  line.