From: Denia Bouhired-Ferrag

Sent: 08 December 2017 16:33

To: Erik Van Der Bij; Tristan Gingold; Javier Serrano; Claude Dehavay; Emmanuel Said; Benjamin Ninet; Maciej Marek Lipinski; Grzegorz Daniluk

Cc: Dimitris Lampridis; Evangelia Gousiou; Ioan Kozsar

Subject: Outcome of meeting on optical repeater design

Dear all,

Thank you for you attendance today. It was important to get your opinions on some of the design choices, and I think indeed we managed to achieve that.

Here are some notes on the main conclusions and the actions to be implemented where applicable:

• Signal specification (for pulses)
• Overall, the specification covers most of the currently existing applications
• Action: CO-HT, contact Alessandro Masi about specific use case to see whether new spec is suitable
• Optical output of Tx, jumper or trimmer?
• Is pulse-width distortion (trailing 16ns extra) problematic?
• CTRV PLL only “cares” about rising edges (Gateware checked by Tom), so trailing falling edge should not pose a problem for GMT
• Decided that final design will include trimmer footprint (not mounted by default, board will have 0 Ohm resistor instead). If no pulse width distortion is tolerated, then user can mount and use trimmer for calibration.
• Cable losses/Signal attenuation
• For same length links, attenuation can vary greatly depending on connector cleanliness, number of patch connections..etc
• There is little value in measuring only a handful of links
• Action: CO-HT/CO-IN, Follow up with EN-EL for possibility of some measurements, still bearing the above in mind
• Could we add diagnostics on receiver power. FPGA can raise an alarm if signal power is below threshold
• Could add complexity to hardware and gateware (1 ADC/comparator per channel, more complex vhdl…)
• Action: CO-HT, study possibility and evaluate benefit Vs development overhead
• Connector type and patch cables
• There is an obvious case to be made for ST connectors:
• Mechanical robustness of the board (no optical cable adapters on the PCB)
• More ports (up to 3 channels, 1 Rx, 2 Tx)
• Decrease production costs
• However, ST connectors are not compatible with existing installations
• Incompatible in case of roll back
• Fibre cables have been pulled and are ready with E2000 connectors
• Use of patch panels should be seriously considered for ST to E2000 conversion
• Action: CO-IN will discuss and come up with final decision based on whether more channels are preferable or E2000 connectivity.
• Concrete plan, with milestones and delivery dates should be mentioned
• Will depend on whether on-board diagnostics are required. This would extend design and testing phase. Inclusion of diagnostics will be decided in a matter of a few days, as per one of the preceding actions.
• Action: CO-HT, evaluate realistic times and update relevant issue on Kanban board

This project can be followed up on task page in the HT board [1], there, you can find the slides from today’s presentation [2].

Have a good weekend,

Regards,

Denia

[1] https://issues.cern.ch/browse/HT-74

[2] https://issues.cern.ch/secure/attachment/51615/opt-repeater-design.pptx

Denia Bouhired

BE-CO-HT

774/ 1-042