The table below lists feedback to the hardware specification documents:
These documents were presented during the New WRS-4 Workshop and the feedback mostly comes from its participants.
The resolved items constitute part of the hardware specification of the WRS-v4.
Comment | Issue | Resolution |
---|---|---|
Ports - WRS-v4 should have at least 20 ports, ideally 24 ports (HDL support) | #1 (closed), #40 (closed) | Accepted - WRS-v4 prototype with have 24 ports, final product TBD (minimum 20 ports) |
Ports - Use compact SFP (not for DWDM) | #2 (closed) | Rejected - Not needed, won't be used |
Ports - It should be possible to have 1Gb on 18 ports | #3 (closed) | Accepted - At minimum 18 x 1Gbps will be supported be default |
Clocking circuit - Architectural suggestions | #4 | Open - to be evaluated with a prototype |
Clocking circuit - Use PPS rising edge as reference instead of 10MHz | #6 (closed) | HW should allow this, HDL/SW to be discussed |
Clocking circuit - What should be the new WRS timing/noise performance? | #35 (closed) | MDEV limited at 5E-13/s due to flicker noise. |
Interfaces - PPS in should be 5V TTL compatible, PPS out also TTL, always 50 ohms | #7 (closed) | Accepted |
Interfaces - 10MHz output should be AC coupled instead of LVCMOS 3.3 | #8 (closed) | Accepted - to be tested later with measurement equipment |
Interfaces - To gain space in the front panel, squeeze the Aux/Abscal connectors | #9 (closed) | Remove Aux |
Interfaces - Do not use CP2105 dual USB, use FTDI FT2232HL | #10 (closed) | Accept |
Interfaces - Use USB-C instead of mini USB type B. | #11 (closed), #37 (closed) | Accept |
Interfaces - Locator LED on the back of the switch | #12 (closed) | Accept |
Interfaces - LEDs - color-blind friendly | #13 (closed) | Accept - we will try... |
Interfaces - Bad idea to have the button to flash directly from USB | #14 (closed) | Accepted - do not flash from USB, remove flash button |
Interfaces - Have buttons next to LCD to change the content | #15 (closed) | Accept |
Interfaces - LCD - overkill and takes space | #16 (closed) | Rejected - LCD will be in the prototype, TBD for the final product |
Interfaces - SMA connector prefered | #17 (closed) | Accept - this is our choice already |
Interfaces - UFL connector - very bad choice | #18 (closed) | In the end,we will not use internal connectors, see #9 (closed) |
Interfaces - External temperature sensor to know environmental conditions | #28 (closed) | The front panel USB host can be used to connect temperature sensor |
Interfaces - Have USB connector inside to connect some additional sensors | #34 (closed) | Accepted - USB host will be mounted |
Interfaces - Put management ports to the right | #36 (closed) | Rejected - Management I/F stays on the left side |
Expansion board - Mimic the pinout of FMC in the expansion slot | #19 (closed) | Accepted |
Expansion board - Provide proper airflow/cooling for the expansion board | #20 | Accepted |
Expansion board - High speed interface for data (link aggregation) | #21 (closed) | Accepted - route to expansion board GTH/GTY links |
Expansion board - SAMTEC connector to the expansion board | #42 (closed) | Accepted - after evaluating different options we will use SAMTEC, see |
Memories - Add high bandwidth external packet storage | #22 (closed) | Rejected - we are OK with the internal memory |
Memories - A selection criteria for storage could also be TBW (TeraByte Written) | #24 (closed) #39 (closed) | Accepted - add mSATA connector |
Memories - To speed up configuration of the FPGA one should use Dual QSPI. Is 256 MB QSPI enough? | #24 (closed) | Not an issue Dual QSPI -> for recovery, not programming FPGA |
Power supply and fans - Power consumption of SFP+ can be 2W | #32 | Accept - In a default HW configuration, PSU will support up to 20 ports with 2W |
Power supply and fans - Emphasize that hot-swappable PSU/fan is critical/desirable | #30 (closed) | Accepted |
Power supply and fans - Consider RF-Noise and EMC when selecting PSU | #5 (closed) #31 (closed) | Accepted - the selected y YH5151-1EBR looks good |
Power supply and fans - Can we make the PCB/switch fanless-ready? | #41 | Ongoing - we will try ... |
Other - Continued supply of WR switches… - drop-in replacement | #25 (closed) | Accepted |
Other - Filter on the inlet of airflow | #26 | investigating |
Other - Dust cover for unused connectors | #27 (closed) | Accepted |
Other - Have simple version of the switch - no features & reliable - oriented for data-center | #29 (closed) | This is to be considered later by companies |
Other - N4000-13SI substrate 10 GBps? | #32 | In collaboration with manufacturer selected substrate suitable for up to 25Gbps |
Other - Feedback from GSI | #38 (closed) | Accepted - all aligned, important: non-redundant option of the WRS v4.0 |