On the development of the final optical multiplexer

  • Slides: 21
Download presentation
On the development of the final optical multiplexer board prototype for the Tile. Cal

On the development of the final optical multiplexer board prototype for the Tile. Cal experiment V. González Dep. of Electronic Engineering University of Valencia, Spain

Outline Introduction Previous work OMB 6 U Final prototype OMB 9 U Present status

Outline Introduction Previous work OMB 6 U Final prototype OMB 9 U Present status

Introduction RODs interface LVL 1 and LVL 2 at each detector – Some preprocessing

Introduction RODs interface LVL 1 and LVL 2 at each detector – Some preprocessing available if needed For Tile. Cal, radiation issues changed front end connection to RODs – Two fibers with same data output each Front. End board – Need to select the fiber with good data in case of errors (OMB aka Pre. ROD)

Previous work OMB 6 U Specifications – VME 6 U Board format – 2

Previous work OMB 6 U Specifications – VME 6 U Board format – 2 input channels (4 input fibers) @ 640 Mbit/s – 2 output channels (2 output fibers) – Digital data analysis (CRC computation) – Trigger and Busy inputs

Previous work OMB 6 U Functionality – Error detection (multiplexing) – Data injection Preconfigured

Previous work OMB 6 U Functionality – Error detection (multiplexing) – Data injection Preconfigured event (counter) Configurable event loadable through VME – Different trigger options External (NIM or LVTTL) Internal. Variable frequency – – – Single Loop Non-stop

Previous work OMB 6 U Designed with Cadence PSD 15. 0 and Altera Quartus

Previous work OMB 6 U Designed with Cadence PSD 15. 0 and Altera Quartus II for FPGA programs Technical details – – – 2 Cyclone EP 1 C 12 and 1 ACEX EP 1 K 100 FPGAs 12 copper layers, 35 μm thick > 2000 routes 0. 15 mm width > 2000 vias 0. 2 mm SMD components widely used

Previous work OMB 6 U TEST at Valencia LAB – Firmware adjustment – Integration

Previous work OMB 6 U TEST at Valencia LAB – Firmware adjustment – Integration with ROD motherboard – Development of control software (XTest. ROD) TEST at CERN (USA 15) – During commissioning – Integration with 8 Tile. Cal ROD – Used now for ROD production tasks

Final prototype OMB 9 U Same basic functionality – CRC checking – Data injection

Final prototype OMB 9 U Same basic functionality – CRC checking – Data injection – 40 MHz clock Differences – 8 channels 16 input fibers 8 output fibers – TTCrx – Mezzanine connectors: more processing power?

Final prototype OMB 9 U Technical details – 9 U VME slave board –

Final prototype OMB 9 U Technical details – 9 U VME slave board – Dual optoelectronic transmitter/receiver (Stratos) – GLINK chips 8 HDMP-1032: transmitter 16 HDMP-1034: receiver – 8 Cyclone EP 1 C 12 (CRC FPGA) One for each channel (2 input fibers and 1 output fiber) – 1 Cyclone EP 1 C 20 (VME FPGA) Changed from ACEX in 6 U version for more logic resources Includes TTCrx control firmware

Final prototype OMB 9 U

Final prototype OMB 9 U

Final prototype OMB 9 U PCB details – Input/output signals designed for operation at

Final prototype OMB 9 U PCB details – Input/output signals designed for operation at 80 MHz – 10 copper layers Signal layers between power/gnd planes Adjacent layers routed orthogonally for minimum coupling

Final prototype OMB 9 U Signal integrity issues – Big area, not too much

Final prototype OMB 9 U Signal integrity issues – Big area, not too much populated Long distance traces (high coupling) Clock distribution JTAG chain distribution VME-CRC FPGA serial bus – Different voltage supplies +3. 3 Volt for FPGA I/O and some logic +1. 5 Volt for FPGA cores (island) +5 Volt for logic +12 Volt for NIM/TTL conversion (ext. trigger input)

Final prototype OMB 9 U Example of SI analysis: VME to CRC serial BUS

Final prototype OMB 9 U Example of SI analysis: VME to CRC serial BUS – Connected between the VME FPGA and all CRC FPGAs – For control, communication, configuration, etc. – 4 lines: CLK, DATA 0, DATA 1, DATA 2, DATA 3 – Frequency of operation 40 MHz

Final prototype OMB 9 U First attempt – Manual placement and routing of bus

Final prototype OMB 9 U First attempt – Manual placement and routing of bus lines – Keep stubs short – Keep same routing layer as much as possible – R, RC termination Problem – Characteristic impedance change at T junction – Reflections at first FPGA creates Inter. Symbol Interference (ISI)

Final prototype OMB 9 U Post. Layout simulation

Final prototype OMB 9 U Post. Layout simulation

Final prototype OMB 9 U Second attempt – Manual placement and routing of bus

Final prototype OMB 9 U Second attempt – Manual placement and routing of bus lines – Keep stubs short – Keep same routing layer as much as possible Problem – Reflections with different termination schemes

Final prototype OMB 9 U Post. Layout simulation

Final prototype OMB 9 U Post. Layout simulation

Final prototype OMB 9 U Final attempt – Manual placement and routing of bus

Final prototype OMB 9 U Final attempt – Manual placement and routing of bus lines – Keep stubs short – Keep same routing layer as much as possible Solution – Resistive termination at both ends

Final prototype OMB 9 U Post. Layout simulation

Final prototype OMB 9 U Post. Layout simulation

Present status Stratos change connectors to be Ro. HS compliant – Need to test

Present status Stratos change connectors to be Ro. HS compliant – Need to test new connectors Board routing on the way – SI analysis still undergoing: JTAG chain, clock distribution and power distribution FPGA firmware adaptation

THANK YOU FOR YOUR ATTENTION

THANK YOU FOR YOUR ATTENTION