CWDM SM-VTRx project update meeting

Wednesday 25 Mar 2020, 10:00 → 11:00 Europe/Zurich

CERN

Subsequent discussions on NanoFIP with BE-CO

• NanoFIP is a WorldFIP Agent (slave) in the form of a rad-tol FPGA-based FMC 
  • https://www.ohwr.org/project/nanofip
  • https://forums.ohwr.org/c/fmc-nanofip
  • Developed and supported by BE-CO
  • Compatible with the current WorldFIP infrastructure in the LHC (e.g. masters, cables, 31.25kbps)
    • It will be able to operate at higher data rates with updated masters (up to and 2.5Mbps)
  • It features a subset of the WorldFIP services
  • It requires a carrier board
    • CO offers an FPGA-based FMC carrier in compact PCI form-factor
    • Custom carrier boards may be used
  • It features 16 Inputs and 16 outputs for slow control
  • New revision of the NanoFIP FMC foreseen
    • Ruggedized version following BI-BL specifications
    • Added I2C/SPI slave mode capability (NanoFIP will be the slave)
    • New hardware features may be added by CO if required (and possible)
  • Its FPGA is used at 50%, so It is possible to add application specific features
    • Most of other teams at CERN using NanoFIP add application specific features
      • CO only supports the base functionality
  • The NanoFIP FPGA gateware may be used as IP Core on other FPGAs
  • Low cost: < 200CHF

• NanoFIP for slow control of BPM LHC consolidation:
  • WoldFIP infrastructure already available
  • For BPM we may fabricate our own simplified rad-tol FMC carrier with no FPGA (we already have one on the NanoFIP)
  • A single NanoFIP per crate would be sufficient
  • We can add an I2C master for programming the CWDM-VTRx
    • It would use the already existing I/Os for slow control

• Further discussion required between BP and BL since both BPM LHC Consolidation and BLM LHC may use NanoFIP and share the crate in LHC

10:00 → 10:20 Latest CWDM SM-VTRx Characterisation Status

Speakers: Carmelo Scarcella (CERN), Manoel Barros Marin (CERN)

CWDM SM-VTRx Characterization.pdf

CWDM SM-VTRx Characterization.pptx

Latest CWDM SM-VTRx Characterisation Status (Carmello)

16 CWDM Single-Mode Versatile Link modules in new SFP enclosure now available (4 per laser wavelength)

All 16 successfully tested for operation as:

• Transmitter @ 5.12Gb/s
  • Using default bias, modulation and rise/fall emphasis
  • Average laser power >2.5mA
• Transmitter @ 5.12Gb/s with lpGBT pre-emphasis for input signals
  • Using optimised bias, modulation and rise/fall emphasis
  • Better eye-diagram with reduced power ~1.5mA
• Transmitter @ 10.24Gb/s with lpGBT pre-emphasis for input signals
  • Nice eye diagram maintained with average power ~1.5mA
• Receiver @ 5.12Gb/s
  • Bit error rate shows good uniformity for all 16 modules

 Discussion:

• Long-term reliability
  • Prediction important
  • Expected to be similar to existing VTRx as only component which is different is the laser die
  • However working in different configuration of laser driver
  • Estimate based on component information (MIL)  (ACTION: EP)
• Pre-emphasis of input when not using lpGBTx
  • Need to ensure this is still OK by systems not intending to use lpGBTx
  • (ACTION: BI-BP to check this for direct ADC solution)
• SFP compatibility
  • Standard pin-out used BUT
    • Commercial SFPs use 3.3V while VTRx uses 2.5V
      • Need to ensure this is correct on power-up
    • Rx & Tx pins correctly routed but P/N inversion compared to SFP standard
  • Component tolerances
    • What is expected? In particular for output power.
    • Can this cause saturation in standard SFP receivers?
    • Should be able to see if this is an issue from data sheets of commonly used receivers
      • Should not forget attenuation in fibre & fibre connections for real installations
    • ACTION: BI to provide EP with such data sheets to see if this is an issue
• Request for documentation (ACTION: EP to provide latest layout, schematics as well as general VTRx documentation & application notes)

 Other follow-up ACTIONS:

• Possible optimisation for BPM project where receiver is not required
  • 2 options
    • Not to populate receiver
    • 2 in one design for transmitter
  • Needs careful evaluation for Cost saving v overheads
    • Overheads
      • 2 designs
      • 2 qualifications
      • No common building blocks for future designs
      • No general standard pool of spares
      • No way back for BPM system if receiver later required
  •  ACTION: EP to provide BI with estimated cost saving for final decision
• Setting-up default VTRx values
  • Request to modify defaults by BI-BP
    • Requires ASIC re-design
      • EP comments that manpower not available
      • Also requires production of new masks with significant additional cost
    • May not be optimal for users not requiring 10Gb/s transmission
    • DECISION: Not to pursue this approach
  • Can be set by lpGBTx
    • LpGBT order deadline by Summer 2020 (ACTION: Manoel to contact EP for deadline extension or future batch productions)
    • If lpGBTx not used then need to foresee I2C bus
      • Can this be provided by nanoFIP?
      • Investigate possible input to new nanoFIP design (ACTION: Eva & Manoel to contact BE-CO)
• Number of units
  • Currently estimating a total of ~8000
    • ~5000 for BPM
    • ~3000 for BLM

 BI specific follow-up:

• Need to coordinate future developments with this component
  • Set-up design kit for all future designs
  • Should we coordinate this through a new BI rad-hard link forum (include Leonardo - EP - in any discussions)?
• BI-BP to study best layout for the front-end board to set-up VTRx
  • Use of additional receiver & fibre v NanoFIP or FPGA solution
• ASIC development
  • Look at ultimate limit of VTRx in final radiation environment
    • Likely to be a fluence not TID issue
    • BI to provide input to EP on composition & energy of expected fluence at representative locations to get a better estimation on the final radiation tolerance

10:20 → 10:40 Future projects for CWDM SM-VTRx in the BI-BP Section

Speaker: Manoel Barros Marin (CERN)

200325_bpm_projects_featuring_vtrx.pdf

10:40 → 11:00 Future projects for CWDM SM-VTRx in the BI-BL Section

Speakers: Francesco Martina (University of Liverpool (GB)), Manoel Barros Marin (CERN)

VTRxMeeting.pdf