[10GMMF] TP3 call Meeting Notes Nov 30th
Dear TP3ers
Here are my notes from Tuesday's call ...
1. List attendees
Aronson, Lew
Bhoja, Sudeep
George, Matt
Ghiasi, Ali
Hanberg, Jesper
Latchman, Ryan
Lawton, Mike
Lindsay, Tom
Lingle, Rob
Popescu, Petre
Rausch, Dan
Thompson, Joey
Van Schyndel, Andre
Weiner, Nick
Witt, Kevin
2. Review notes from last meeting
No Comments
3. Accept/Reject TP3 timeline, goals and mode of operandi
Accepted
4. Remove TP3 Jitter TBD's Tom L.
Tom presented his thinking that the jitter arising form both the additional noise (added to emulated Modal noise and RIN) and the ISI impairment (from the representation of the channel) are sufficient to adequately stress the receiver from jitter tolerance standpoint. In essence he and Piers (whilst not fully in alignment) are both converging on a view that any additional jitter required, above and beyond that introduced by the noise and ISI, would be sufficiently neglible that it is not required in the test.
An informative low frequency jitter test is still required.
The group did not have unaminous support behind Tom proposals. However it was agreed that we would proceed as per his suggestion unless a complete counter proposal is presented. Those wishing to seek further understanding should contact Tom who has more analysis which he is happy to share.
5. Agree approach for ISI generator
PIE-D
-----
We discussed MMSE vs ZF PIE-D. It was agreed that for the purposes of TP3 the distinction is not that relevant. Our goal is to use a metric to rank channels within a fiber set and then pick appropriate channels for Rx testing. The group felt that MMSE and ZF PIE-D's may have different values but would likely rank IPRs in a similar way.
Given that we decided to stay with MMSE PIE-D as this has the strongest historical context with previous work by the IEEE committee.
Sudeep kindly agreed to go through the process of picking a fiber set, ranking the fibers with the MMSE PIE-D metric, and then attempting to pick 3 channels (pre, post cursor and quasi symmetric) close to a given PIE-D. The group will then review this work, consider what is needed to emulate the channels (number of taps, spacing etc) and attempt to define a processs for selection/synthesis of appropriate channels. This work is due to be reviewed on the Dec 14 call.
For next week's call Sudeep is planning to present a view on how we close out our TBDs for the informative sensitivity test. This essentially involves picking the appropriate filter (Bessel Thompson with a given BT) to simulate the ISI in the system.
Any comments/corrections please get back with me.
Best Regards
Mike