Re: [8023-10GEPON] [Dual-Rate-PMD] New merged draft of presentation
Dear Frank and David,
I agree to Frank and I have two comments.
I don’t think PMD must discriminate data bit-rates at each burst signal
reception. Because the pattern matching of delimiters and/or preambles
(in addition to 64B/66B and 8B/10B rules) can discriminate valid data
from 1G and 10G signals even if CDR can discriminate both bit-rate
signals and PHY don’t know which signal comes.
I don’t think the electrical splitter loss degrades SNR directly because
major noise sources are optical front-end circuits (PD + TIA). So I
believe we only have to compensate the electrical splitter loss.
Best Regards,
Ken-Ichi
Frank Effenberger wrote:
> David,
>
> When the 10G signal is operating, the 1G channel will not be used, and vice
> versa. That is the whole idea of the parallel scheme. So, crosstalk is
> meaningless in this context.
>
> Regarding losses, these filters are in the electronic domain, and so I fully
> expect that sufficient amplification will be available to make the filter
> losses tolerable.
>
> Frank E.
>
>
> -----Original Message-----
> From: David Li [mailto:dli@ligentphotonics.com]
> Sent: Thursday, March 01, 2007 2:12 PM
> To: Frank Effenberger; STDS-802-3-10GEPON@listserv.ieee.org
> Subject: Re: [8023-10GEPON] [Dual-Rate-PMD] New merged draft of presentation
>
> Frank E,
>
> On page 9 and page 10 of the PMD101 slides, two parallel LPFs are showed
> after the TIA. One is for the 1Gbps Signal, the other is for the 10Gbps
> Signal. Do you need a Data Rate Detect Circuit (maybe hard to realize in the
> PMD modules) to switch these two LPFs?
>
> In the 10G mode, without the Data Rate Detect Circuit, the low frequency
> components of the 10Gbps Signal will leak through the 1GHz LPF and the ISI
> (Inter-Symbol Interference) penalty will be very high for the 10G mode,
> since the low frequency spectrum of the 10Gbps Signal can be lower than
> 10MHz.
>
> In the 1G mode, the 1Gbps Signal will all leak through the 8GHz LPF besides
> the signal path in the 1GHz LPF, then the two LPFs will act like a power
> splitter, not ideal filters. The power penalty will be at least 3dB (ideal
> case) in 1G mode.
>
> Best regards,
>
> David
>
> David Li
> Ligent Photonics, Inc.
> 2701 Dukane Dr., Suite 102
> St. Charles, IL 60174
> Phone 630-513-7226 ext 15
>
>
> ----- Original Message -----
> From: "Frank Effenberger" <feffenberger@HUAWEI.COM>
> To: <STDS-802-3-10GEPON@listserv.ieee.org>
> Sent: Wednesday, February 28, 2007 9:43 PM
> Subject: [8023-10GEPON] [Dual-Rate-PMD] New merged draft of presentation
>
>
>> All,
>>
>> I have merged the additional analysis and data from Mr. Suzuki and Mr.
>> Tasuta (that you both very much!) into the "PMD101" slides, producing the
>> attached file. I think this gives a good technical review of where we
> stand
>> regarding the ad-hoc's analysis of the dual-rate burst receiver system.
>>
>> Regards,
>> Frank E.
>>
>
>
>
--
Ken-Ichi Suzuki
NTT Access Network Service Systems Labs.
E-mail:kenyichi@ansl.ntt.co.jp
Tel:+81-43-211-3189/Fax:+81-43-211-8250