Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

[802.3_OMEGA] Comments to contributions of April 6th

Dear colleagues,

In order to make a more efficient in time the following meeting of April 6th, I decided to use the reflector for commenting contributions.

  • Comments to slide 2:
    • Single PMD spec interoperable in the whole wavelength range 850nm to 1310nm implies:
      • Specifications for the whole range
      • Design / Implementation for the whole range
      • Qualification for the whole range
      • Production test for the whole range
      • Even with just 3 lambdas you are proposing to increase the cost! What is the benefit for the OEM?
    • "Reliability is not “a nice to have”. Reliability cannot be compromised"
      • Are you meaning that by specification of a 850 ~ 1310 nm RX, the reliability is improved?
      • Mission profile consistent with OEMs and international standards have been proposed in the TF.
      • This statement should be elaborated.
  • Comments to slide 3:
    • The devil is in the details.
    • Only InGaAs PD responsivity is shown. My experience is that responsivity decay in PIN diodes always implies slow discussion currents that penalize the device speed. The PD speed is wavelength dependent.
    • According to reported responsivity curve:
      • R = 0.65 A/W @ 1310nm
      • R = 0.4 A/W @ 980 nm
      • R = 0.2 A/W @ 850 nm.
      • This means a variation in RX sensitivity of 5.1 dB. Which link budget is going to support these extra 5 dB?
    • Responsivity depends of the ARC on top the PD:
      • There are InGaAs PDs with R > 0.8 A/W @ 1310 nm
      • There are InGaAs PDs with R > 0.6 A/W @ 980 nm 
      • PD ARCs are optimum for a shorter wavelength ranges. 
    • At 850nm it is much more efficient to use GaAs PINs with R ~ 0.6 A/W @ 850nm
    • Materials used in the optical coupling elements (e.g. lenses) should support the full wavelength range. This means extra cost.
  • Comments to slides 4 and 5: 
    • What is the value proposition for the OEM? Benefit?
    • What are the criteria of the OEM to select 850nm, 980nm, 1310nm for a given use case?
    • On top of expensive PHY, more expensive process to take decision.
    • The OEM does not have requirements related with wavelength.
  • Comments to slide 6:
    • In the CSD you have as rules for Distinct Identity:
      • Each proposed IEEE 802 LMSC standard shall provide evidence of a distinct identity. Identify standards and standards projects with similar scopes and for each one describe why the proposed project is substantially different.
      • Substantially different from other IEEE 802.3 specifications / solutions.
    • One of our answers is:
      • The project may define multiple PHYs, but will define only a single PHY for each rate, media, and link reach combination.
  • Comments to slide 7 and general:
    • This contribution is not a baseline proposal (also stated in slide 2). 
    • However, it introduces new topics for discussion in the task-force, without clear focus.
    • It is not shown link budget, economical feasibility, technical feasibility, reliability data, etc for multi-lambda receiver and BiDi, however, they are presented as they would be valid options to meet the project’s objectives.  

Rubén Pérez-Aranda 

Knowledge Development for POF, S.L.
A: Ronda de Poniente 14 2º CD, 28760, Tres Cantos (Madrid), Spain 
P: +34 91 804 33 87 Ext:110 
M: +34 689 319 866

To unsubscribe from the STDS-802-3-OMEGA list, click the following link: