Hi Brian,
Thank you for double checking the spectrum and the impact on the equalizer and Echo canceller using PAM2. I did not see strong data correlation for PAM2 in my simulation. If there
is no downside to use PAM2, why not? It makes PMA training robust and efficient.
I am very glad to discuss on this on the reflector and during the plenary meeting.
Best wishes,
Tingting
发件人: Murray, Brian <Brian.Murray@xxxxxxxxxx>
发送时间: 2024年11月2日
1:09
收件人: STDS-802-3-SPEP2P@xxxxxxxxxxxxxxxxx
主题: Re: [802.3_SPEP2P]
答复: dg meeting today October 29
Tingting
Thank you for the details on the calculation of the power spectrum. It does appear that the randomization that you have applied
has helped to ensure there are no spurs in the spectrum, which was an issue identified during the 10BASE-T1L development.
I do agree that PAM-2 has advantages of more robustness during blind acquisition and start-up as PAM-2 will have more SNR
than PAM-3. I am not sure it is solving a problem that exists as we appear to be able to always open the eye with PAM-3. But I think it does warrant further investigation and if there are no downsides to including PAM-2 during training then it may be worth
including.
I will calculate a power spectrum for your proposed PAM-2 line coding table of NND PAM2 6-tuples with your running disparity
scheme to ensure randomization and control of running disparity during PAM-2 training. I am sure I will get the same result, but it is important to ensure we don't have any frequency spurs here, so good to have a cross check. And I will investigate using your
PAM-2 line coding table during start-up. I would like to look at the difference between the convergence of the DFE and Echo cancellers in PAM-2 and when we switch to PAM-3. There must be significant data correlation in the smaller PAM-2 table, though your
results appear to show that the adjustment moving from PAM-2 to PAM-3 is small. I would like to confirm that.
I suggest we discuss further on the reflector and during the plenary meeting.
Thanks
Brian
From: zhangtingting (O) <00001e92abc91102-dmarc-request@xxxxxxxxxxxxxxxxx>
Sent: Wednesday, October 30, 2024 12:56
To: STDS-802-3-SPEP2P@xxxxxxxxxxxxxxxxx <STDS-802-3-SPEP2P@xxxxxxxxxxxxxxxxx>
Subject: [802.3_SPEP2P] 答复: dg meeting today October 29
|
[External] |
Regarding the question on spectrum from Brian,
The spectrum on Slide 6 and 7 (https://www.ieee802.org/3/dg/public/May_2024/Tingting_3dg_01_29_10_2024.pdf)
is achieved by averaging 4000 spectrum and each spectrum corresponds to 393216 PAM2 symbols generated from random bits. As shown following, introducing the delimiters to the 4B6B PAM2 training sequence does not change the spectrum. It is a good solution for
fast synchronization, which of course depends on the application requirement.
Without considering delimiters, PMA training using PAM2 instead of 8B6T only requires 4B6B encoding and decoding, the hardware complexity of
which I believe is negligible.
PAM2 training has many advantages. PAM2 is more robust than PAM3. Rx DSP converges much easier and faster especially in low SNRs, indicating
easier for longer distance transmission. Besides, PAM2 training also allows the equalizer to work stably when the transmitted signal is switched from PAM2 to PAM3, regardless of the operation mode (traditional or PR). Therefore, the risk of training failure
is minimized while the training efficiency is increased and the coding gain brought by the PR equalization is ensured.
Bests,
Tingting
-----邮件原件-----
发件人: George Zimmerman <george@xxxxxxxxxxxxxxxxxxxx>
发送时间: 2024年10月30日 1:46
收件人:
STDS-802-3-SPEP2P@xxxxxxxxxxxxxxxxx
主题: Re: [802.3_SPEP2P] dg meeting today October 29
Dieter - I believe Tingting was referring to PAM-2 at the same line signalling rate (baud) as the 100 Mbps PAM-3 signal would be (i.e., 80 MBaud). This way the attenuation is the
same as it will be for run-time PAM-3... (otherwise I don't see how the training would work usefully...)
George Zimmerman, Ph.D.
President & Principal
CME Consulting, Inc.
Experts in Advanced PHYsical Communications
george@xxxxxxxxxxxxxxxxxxxx
310-920-3860
-----Original Message-----
From:
stds-802-3-spep2p@xxxxxxxxxxxxxxxxx <stds-802-3-spep2p@xxxxxxxxxxxxxxxxx>
On Behalf Of Dieter Schicketanz
Sent: Tuesday, October 29, 2024 10:41 AM
To:
STDS-802-3-SPEP2P@xxxxxxxxxxxxxxxxx
Subject: dg meeting today October 29
Regarding last presentation today (Tingting Zhang), it is my understanding that of coarse PAM2 needs 4 dB less S/N but at 500m the signal attenuation is much higher than for PAM3.
Is that considered ?
Because of time constraints I withdrew this question today.
--
Best regards
Dieter Schicketanz
________________________________________________________________________
To unsubscribe from the STDS-802-3-SPEP2P list, click the following link:
https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-SPEP2P&A=1
To unsubscribe from the STDS-802-3-SPEP2P list, click the following link:
https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-SPEP2P&A=1
To unsubscribe from the STDS-802-3-SPEP2P list, click the following link:
https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-SPEP2P&A=1
To unsubscribe from the STDS-802-3-SPEP2P list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-SPEP2P&A=1