dt muon sorter - istituto nazionale di fisica nuclearetridas_nov04).pdf · a.montanari-infn bologna...
TRANSCRIPT
Tridas week, 09-Nov-04A.Montanari-INFN Bologna 1
DT Muon Sorter
DT front endBarrel
DT Track Finder
DT Sorter
GLOBAL MUON TRIGGER
Detector in UXC
Countingroom in USC
LVDS copper link
Opto link
Towers in UXC
DT station (minicrate)
CCB
Trigger Boards(BTI,TRACO,TSS)
Sector Collector
Server Board
12xWedge Sorter
In: max 12 tracks from 6 f Track-Finder of a wedge
Out: 2 “best” tracks
1xBarrel Sorter
In: max 24 tracks from 12 Wedge Sorters
Out: 4 “best” tracks
Main news:
-Trigger chain from front-end upto Wedge Sorter tested at bunchedtestbeam at beginning of October
- All final boards, but Sector Collector..
Tridas week, 09-Nov-04F.Odorici-INFN Bologna 2
DT Sector Collector
LVDS link:
- 2 Ethernet cables/minicrate FTP cat.6
- TX rate @ 480 Mbps (< 50 m), with National
Semicond. chipset:
a) serializer 10-1 DS92LV1021 (8 IC/link)
b) cable equalizer CLC014 (8 IC/link)
c) deserial. 1-10 DS92LV1212A (8 IC/link)
Opto link:
- 6 multimodal fibers (Ericsson)/SC (< 100 m)
- TX rate @ 1.6 Gbps, with GOL serializer
(32 bits @ 40 MHz), and Honeywell opto-ICs:
a) VCSEL trasmitter HFE4190-541
b) Pin Diode receiver HFD3180-102
DT front endBarrel
DT Track Finder
DT Sorter
GLOBAL MUON TRIGGER
Detector in UXC
Countingroom in USC
LVDS copper link
Opto link
Towers in UXC
DT station (minicrate)
CCB
Trigger Boards(BTI,TRACO,TSS)
Sector Collector
Server Board
Main functionalities:
- a SC board serves a sector (4/5 stations)
- track sorting (1 out of 2 tcks/station)
- DT trigger path syncronization (compensate forTime-Of-Fligth, TTC fibers and LVDS cables) via clock delay and pipe lines.
Tridas week, 09-Nov-04F.Odorici-INFN Bologna 3
Sector Collector to Track Finder opto-link
ETTF(η view)
1 board/wedge
PHTF(φ view)
6 board/wedge
2 Opto Rx (5 ch) board / ETTF
1 Opto RX (4 ch) board / PHTF
Opto Tx (4 ch) piggy board
Sameboards
TF crate backplane
MB1
MB2
MB3
MB4
Clock
Vcc
Control
LVDS Rx piggy boards
SC
Virtex-II FPGA used as:
- Deserializer (Rocket IO)
- RX parity checker
- I/O Router
Tridas week, 09-Nov-04A.Montanari-INFN Bologna 4
Combined test with PHTF in preparation of testbeam (1)
Combined test with emulated track segments @ 40 MHz with new PHTF was ok!
In Sep04 the SC-to-TF interface and Wedge Sorter, were tested togheterthe new PHTF in Vienna setup at CERN
I/O was tested at 40 MHz by using 2 Pattern Unit board in a portable VME crate
The WS processor was programmed in a simplified way, in order to pass the tracksfound by only one PHTF
Read back the outputthrough WS
Emulation of inputsfrom 2 MB stations
Pattern Unit Pattern Unit
Tridas week, 09-Nov-04A.Montanari-INFN Bologna 5
Combined test with PHTF in preparation of testbeam (2)
Portable VMEw/ 2 Pattern Units
Track Finder crate
LVDS-to-TTL
PHTF
WS
SC-to-TF
Vienna lab at CERN
Front
Back
Tridas week, 09-Nov-04A.Montanari-INFN Bologna 6
Setup at bunched testbeam
Copper link, 40 m(2 cables for redundancy) Data reduction
(Actel ProAsic)
LVDS deserializer(1:10) + equalizer
Record data at IN (Track segments from chambers) andOUT (Reconstructed muons) of PHTF
(max rate determined by VME bandwidth)Pattern Unit 1 Pattern Unit 2
CENTRAL_DAQ VME crate
2.5 m
Chambers syncronization, SC functionalities, PHTF, WS tested in real environment!First look at data in Jorge’s talk
SPY_DAQ through JTAG
Tridas week, 09-Nov-04A.Montanari-INFN Bologna 7
Wedge Sorter production- Start production submitted mid Oct 04- First evaluation board expected on 22 Nov 04- If the evaluation board is ok, confirm full production (18 boards)- Production start expected by Jan 05 (as scheduled)
- Each board will be tested with our test jig
TTLGTL+
TTLGTL+
LVDSTTL
Shielded flat cables (VME)VME
TTL 128 I/Oup to 100 MHz
Pattern Unit
Tridas week, 09-Nov-04A.Montanari-INFN Bologna 8
Barrel Sorter requirements and design approach
TASKS: BS algorithm is similar to WS one BUT– 24 input tracks à much bigger I/O ( ~ 870 )
à complex ghost bustingà sorting 4 out of 24 (heavy task)
Design approach similar to WS:– All functionalities on one big FPGA:
Altera StratixII EP2S130, 1508 fBGA (new device!!)– 9U, 400 mm VME board– Latency: 3 BX for ghost-busting + sorting&multiplexing
with some main different features:– Inputs from cable (LVDS) through connectors on board top– Main FPGA on mezzanine board
Tridas week, 09-Nov-04A.Montanari-INFN Bologna 9
Barrel Sorter board layout
Mother board prototype expected by mid of December 04
Tridas week, 09-Nov-04A.Montanari-INFN Bologna 10
BS main chip design
Sorter 1 Sorter 2 Sorter 3 Sorter 4
MUXs
Ghost Busting
75 ns
• Notice that (sorting 4 out of 24) >> 2 x (sorting 2 out of 12) !!!– In the fully-parallel sorting algorithm used in WS, logic usage and routing
congestion increase as very steep functions of the number of input words
• New FPGAs work better with fine segmented pipelines:– BS uses sequentially 4 x (1-out-of-24) sorters in pipeline
StratixII floorplan
Tridas week, 09-Nov-04A.Montanari-INFN Bologna 11
Milestones update
DT Sector Collector Prod start Jun-03 Delay: Sep-04DT Barrel Sorter Proto done Nov-03 DoneDT Sector Collector Prod done Dec-03 Delay: Mar-05DT Wedge Sorter Prod done Jul-04DT Barrel Sorter Prod done Jul-04
DT Sector Collector Proto done Jan-05DT Sector Collector Prod start Jun-03 Delay: May-05DT Barrel Sorter Proto done Nov-03 Delay: Jan-05DT Sector Collector Prod done Dec-03 Delay: Jul-05DT Wedge Sorter Prod done Jul-04 Delay: Jan-05DT Barrel Sorter Prod done Jul-04 Delay: Mar-05
Changes into:
Tridas week, 09-Nov-04A.Montanari-INFN Bologna 12
Boards status file
Aug-05May-050Jul-050May-05Jul-05May-05PPP-Jan-05
70Bologna
60SectorCollector
SCDT
Apr-05Mar-050Mar-050Feb-05Mar-05Feb-05PPP-
Jan-053
Bologna1
DTBarrelSorter
BSDT
Mar-05Feb-050Feb-050Dec-04Jan-05Dec-04PPP-
Nov-0415
Bologna12
DTWedgeSorter
WSDT
Sep-05May-050Aug-050May-05Jul-05May-05PPP-
Jan-05110
Bologna96
DTTFGigabitOptical
Rx
OpticalInputDT
Readyfor
USC55
Readyfor904
ProdTest %Compl
ProdTestFinish
Prod %Compl
Prodtest start
ProdFinish
ProdStart
StatusTotal-respNeededExplanItemSubsys
In red, modified/new items: