|Dear Peter, dear Colleagues,|
thank you very much for your feedback on the presentation. We have performed new measurements
and analysis that have been included in a new version of the presentation already posted at:
I hope all the gaps identified by you are covered in the new version. Note: you will see that the value
of table in pg. 18, line 14, column 6 has been changed from -19,73 to -18,73. It was a typo error when I copied
the data from the lab notebook.
In the addition, the new version also includes in the last slide links to experimental results about the robustness
of the specification against variations of the channel response (your last point). Please, also see the links:
KDPOF - Knowledge Development for POFhttp://www.kdpof.com
First I would like to thank you very much for submitting this presentation with test results to the September meeting. I discussed these results with Pete Anslow. We both feel that this is a very good start of an activity that we feel should have been started earlier in the process. We note that you conclude that these results indicate that the specification probably will need to be modified. · Currently tested transmitters, working perfectly in the link, fail the specification. None of them actually pass the test. We are happy to read that you will consider a refinement of the specification. However, even with the changes proposed on page 22, there are still many of the results columns that fail the ER specification but appear to work in this link. · Currently tested transmitters, failing the test and not working in the link, have characteristics that fail the test by a large margin. They are not even close to the spec. We would like to encourage using transmitters that fail the test by a smaller margin. · Regarding your observations on page 20 of your presentation, o We can see that all the implementations produce very similar sensitivity in the receiver. o #2, #3, and #4 does not meet HD4 at 110ºC. The HD4 parameter was selected to be far enough of the SNR needed in the detector for sensitivity (25 dB), avoiding the necessity of HD4 compensation in the receiver. Also no correlation with TP3 sensitivity differences · We observe that there is limited correlation between transmitter measurements and the receiver performance. We would be expecting a stronger correlation to ensure that the specification is adequate to ensure multi-vendor interoperability. We would like to encourage the Task Force to further investigate this. · In your test you are using 15m of POF in line with the requirements for the automotive applications. Our biggest concerns were (and still are) related to the home application with 50m of POF. We would like to see testing with several lengths of POF around 50m, so 45m, 50m, 55m, to create an understanding that the specification is sufficiently robust, and that the performance doesn’t “collapse” at just above 50m. We also would like to see testing where the POF conditions are made more stressful, simulating home environments, where POF is intended to be deployed, so not just a reel of POF, but maybe even bent, squeezed, etcetera. So in conclusion we are very happy with this activity and we strongly feel that it should be continued and extended towards making the specification sufficiently robust, especially for home applications. Because of the current conclusions we feel that the P802.3bv draft is not ready to be submitted to RevCom and therefore we need to maintain our disapprove of latest draft of the BV standard. We hope to continue our exchange of analysis of (further) test results. Kind regards, also on behalf of Pete Anslow, Huawei Technologies Ltd, 华为技术有限公司 European Research Center, 欧洲研究所 Karspeldreef 4, 1101CJ Amsterdam
As we agreed in the 802.3 Plenary Meeting of July, a presentation has been posted to
where I provide measurement results that validate the transmission distortion parameters
as requested by the comment #118 against P802.3bv/D2.0.
The presentation also provides an small tutorial on the script specified in the draft to compute the
distortion parameters based on the signal captured in TP2, when the PMA is configured by MDIO
to generate test mode 6. I hope that this small tutorial helps you to understand the rational
of the signal processing that the script implements to calculate the transmitter distortion parameters.
I will present these measurement results in Sept Interim Meeting in the 802.3bv TF. Any comment or
question that may help to improve the presentation and get the comment solved is very welcome.
Thank you and best regards,
P802.3bv Comment Editor, Editor
thank you to open the thread on this topic. Regarding to your two key points, I would like to put the attention
of the group on two presentations:
- This presentation shows real measurement results of PMD TX for 4 different PMD TX integrating 4 different LED part numbers.
- Based on real measurements, the harmonic distortion (HD) of the PMD TX is characterized and modeled, and simulations on sensitivity of the receiver as a function of the PMD TX HD are reported
- A method to qualify the PMD TX as valid for GEPOF transmission is developed and provided.
- It is worth to mention that the GEPOF specification is an evolution of real products in the market of Gigabit over POF (VDE norm). Therefore, the TF has used the previous experience of this real products to build the 802.3bv specification. Though we have made a number of different choices from that VDE draft, both, VDE and 3bv, are based on PAM16 plus THP and the same type of photonics. During SG, the technical feasibility was demonstrated by theoretical analysis that supported the baseline specification, and by real experiments using VDE based existing implementations. Following presentations show VDE based devices operating satisfactorily in the field on a standard version of POF (A4a.2).
- This presentation contains the measurements of transfer functions carried out for several different SI-POF that meet IEC 60793-2-40 sub-category A4a.2.
- Based on those measurements, the lower bound limit for the spec have been developed.
- As it is stated in sub-clause 114-7: "1000BASE-RHx operation requires fiber optic cable meeting the requirements of IEC 60793-2-40 subcategory A4a.2 multimode plastic optical fibers with appropriate augmentation as specified in this sub-clause”. This is because 1000BASE-RHx specifies the modal power distribution (MPD) of PMD TX in spec point TP2 and there is not international Std. that specifies the insertion loss and the transfer function for that MPD.
- The specification/control of the MPD is necessary in TP2 to guarantee a minimum bandwidth of the channel and a maximum insertion loss by modal dispersion.
- The P802.3bv TF have measured the most representative POF fibers of the market (>98% of market share). Is it enough to create an standard?
- Subclauses 114.7.4 and 114.7.5 provide references to international standards on methodologies to measure the insertion loss and transfer function. The P802.3bv TF thinks that response and insertion loss specs have to be guaranteed by the POF manufacturer and the MPD of TP2 per EAF by the PHY implementor. Is the end user of the technology to measure the transfer function of the POF in his home? Do really that for Cat 5 cables?
The 802.3bv TF is very open to work with you (Peter and Pete) to improve the draft receiving from you specific feedback for doing that.
KDPOF - Knowledge Development for POF
I initially sent it to the wrong reflector. Kind regards, also on behalf of Pete Anslow This email is intended to provide some further exposure to several unresolved comments on optical subclauses in IEEE 802.3bv, especially because of the intent of bv to request the 802.3WG for approval at today’s Closing Plenary to go for Sponsor Ballot from the San Diego meeting. This email is written on my behalf and of Pete Anslow, with whom I share similar concerns. There are unresolved comments to bv regarding what we feel is an insufficiently validated optical specification to ensure multi-vendor compatibility. This is especially important because of the claimed application for home users. We strongly feel that home users should be protected by an extremely robust specification. Our comments specifically addressed the need to validate the optical interface specification by testing, showing that devices meeting the specification would work in the application and that devices not meeting the specification would not work in the application. As far as we are aware this testing has not been done and only simulation results have been presented which we feel are an inadequate validation of the specification. Additionally, while the draft standard calls out a particular optical fiber specification not all fibers compliant to this specification will operate satisfactorily so the draft standard contains additional requirements, which are not realistically measurable by consumer end users. Kind regards, also on behalf of Pete Anslow, Huawei Technologies Ltd, 华为技术有限公司 European Research Center, 欧洲研究所 Karspeldreef 4, 1101CJ Amsterdam