[10GMMF] Notes from Aug 10th meeting on TP3 definition
Here are the notes from today's meeting of the TP3 Definition Group:
If I have missed anyone please let me know and I'll add him to the list.
Jan Peeters Weem
Andre Van Schyndel
2. Normative tests, what to test for, philosophy
Mike has sent a slide to the reflector with the agenda, but it did not arrive in time. Tom's email has clarified the situation. There was some discussion on how we decide what items attract normative tests and (not the same thing) what needs 100% testing. We can revisit the subject next week if necessary.
Agreed that the major stressors to consider are jitter, ISI (static and dynamic), modal noise and RIN, and the optical spatial conditioning.
Petre volunteered to address the ISI piece, with help from Piers. I said that if we found channels at similar percentiles on several metrics, such as PIE-D, PIE-L and 3 dB bandwidth, that would guide us to reasonable simulated channels. The 81 (65) fibre model is a good starting point: it probably has most of the variety of types of responses we are likely to see. Petre thought we would end up with a set of ISI cases, not a single one. We should consider if a Bessel-Thomson filter is a suitable candidate. Don't want to choose simulated channels that give receivers an untypically easy task.
On modal noise and RIN, we seek to know how to relate a Gaussian noise to a sinusoidal emulation. If this is feasible, we can combine these noises by one emulation, without knowing their statistics very precisely. Tom's expertise may help us here.
Other items to consider are the electrical-to-optical conversion, and its linearity (Abhijit with Lew), and the optical spatial conditioning - the offset launch patchcord may be all we need.
Need to make sure that calibration is feasible.
Please let me know if you have any comments or corrections.