Strategy for unified data link

Zhen-An LIU

Institute of High Energy Physics

New Ideas on Belle II DAQ• Consideration

Base– HS data

transmission– Optical link– Unified Readout– Unified Data link– Also trigger?– Also timing?

• My talk on PMC/COPPER link

Trigger path suggestion to Iwasaki-san

• If HS data link used– RocketIO (SEDES)– Synchronization needed (per channel)– Channel Alignment needed (per system)– Pipelined processing

• Trigger latency should reserve big enough– 5 us maybe enough but conservative– 7 us in BESIII

• Trigger distribution– More time needed

Scheme1 for Data link (to Copper)• Based on discussions two schemes in consideration • Scheme 1

– Detector specific readout( Gary’s board or other?)• 16 ch/optical line• Data preprocessing• Partial packaging• Data frame/protocol• Length/ch/event?• Trigger L1 driven

– Data merger• Further packaging • Necessary to save COPPER boards• Output to COPPER PMC• Data frame2

– PMC • Data mover to CPU on COPPER

VMEBUS

Clock(40MHz)

Test output

Reserved

95Con

95Con

5V/3.3VBus

SwitchIDT

QS316211

OpticalFiber 9

HFBR5921L 0

HFBR5921L 9

HFBR5921L 1OpticalFiber 1

OpticalFiber 8 HFBR5921L 8

34 PinLVDS

34 PinLVDS

SN75LVDS387LVTTL-LVDS

SN75LVDS387LVTTL-LVDS

16*12 Test Pins

16 Test Pin

16 Test Pin

16 Test Pin

16 Test Pin

Power SupplyLT1764EQ-DD1.5V,2.5V,3.3V,5V1.6~2.5V Adjustable

LED Display

D(0:31)SN74

ALVTH162245

D(0:31)

SN74ALVTH162244MC100EP16FD

CDCF580180MHz

XC2VP40FF1152-6

Download circuit

1.6 GBps

/ch

Merger test module

From UnifiedReadout

To COPPERPMC Control

SignalsVia USERpins

Optical Fiber0

To be decided with Scheme 1

• Clock for HS transmission• System clock• How many input ch / board?

– Data rate– Data size

• VirtexII Pro / VirtexV ?• Synch/Asych?

– Synchronized data transmission.

• Data frame

Status• Hardware available

– Data source – Merger board– Test system available

• Protocol– Need FRST, Tsync sign

als from timing sys.– Need more discussion w

ith timing/readout/COPPER people

• PMC to be designed

Scheme2 for Data link

• In scheme 1:– Optical transceiver inside detector( say CDC)– Question:

• Reliability not so good in BESIII– Radiation? Experiment did before design– ESD effect? Experiments under taken– Unclear yet!

• Stop machine for replacement?• Open the Detector for replacement?

– Possible solution• Optical transceiver outside detector• Merger before HS data link

Scheme2 for Data link

• Scheme 2– Merger near detector – Electrical to FEE– Optical to

PMC/COPPER– Easy maintenance– SFP transceiver

More discussion scheduled

• A discussion is scheduled tomorrow

• Idea from detector and FEE welcome