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

Re: [10GBT] Updated Tables



Hiroshi,

Thanks for the input. I'm attaching the matlab source code I'm using so that
we can understand where the discrepancies are. This code is a slight
modification of Carlos and George's code that is on the task force web site
- please diff the two files and check what the differences are.

With respect to the integration of the noise to Fs/2, this assumes that the
optimum matched filter uses none of the signal above Fs/2. Of course, that
is an approximation. In any case, I would be fine if we agree to just assign
an additional

10log10(1000/825) = 0.836dB

noise immunity gain to PAM8 over PAM12. In other words, the starting "Total
EMI penalty" for PAM12 will then be 1.64dB over PAM8.

With respect to the (0.75+0.25D) transmit filter, this should be considered
independent of the choice of PAM levels. This is why I'm only assigning a
0.8dB emissions penalty to PAM12 over PAM8 in the table - if you check the
transmit PSD of PAM8 with the (0.75+0.25D) filter and PAM12 without the
filter, you will see that the emissions penalty is really on the order of
7dB.

However, I'm surprised if Albert is seeing a big difference in performance
with and without the transmit filter. Please ask him to send me his code and
I'll take a look.

Regards,
Sailesh Rao.
srao@phyten.com


>From: hiroshi takatori <hiroshi.takatori@KEYEYE.NET>
>Reply-To: "IEEE P802.3an" <STDS-802-3-10GBT@listserv.ieee.org>
>To: STDS-802-3-10GBT@listserv.ieee.org
>Subject: Re: [10GBT] Updated Tables
>Date: Mon, 19 Jul 2004 21:31:15 -0700
>
>Sailesh,
>
>Following is my results of your requested simulation. Note that the
>purpose of the simulation is to repeat what you did so that we can talk
>with the same language. Although I am not sure why PAM12 has penalty in
>coding gain, I used 10.2dB coding gain for PAM8 and 9.8dB for PMA12. I
>think these are what you used (Correct me if I am wrong). I also added
>your numbers in red for "BGN for 0dB margin" for the comparison.
>
>
>
>Model 1, CAT7 100m
>
>PAM Level
>
>TX Digital Shape
>
>Assumed Coding Gain
>
>TX Power
>
>Assuming 2Vpp with simple LPF at Fs/2
>
>SNR
>
>Noise margin with BGN =
>
>- 150dBm/Hz
>
>BGN for 0dB margin
>
>Difference between PAM8 and PAM12
>
>PAM8
>
>0.75+0.25/z
>
>10.2dB
>
>3.3dBm
>
>25.4dB
>
>5.5dB
>
>-137.4dBm/Hz
>
>
>
>2dB worse
>
>NONE
>
>4.8
>
>25.6
>
>25.5
>
>5.7
>
>5.6
>
>-135.4
>
>-133.2(Sailesh)
>
>Normalized to 0dB
>
>PAM12
>
>NONE
>
>9.8
>
>4.6
>
>29.7
>
>29.6
>
>5.9
>
>5.8
>
>-136.1
>
>-134.2(Sailesh)
>
>0.7dB worse
>
>1.0dB worse
>
>Note: Red is from Sailesh results
>
>
>
>Model 2, CAT6 55m
>
>PAM Level
>
>TX Digital Shape
>
>Assumed Coding Gain
>
>TX Power
>
>Assuming 2Vpp with simple LPF at Fs/2
>
>SNR
>
>Noise margin with BGN =
>
>- 150dBm/Hz
>
>BGN for 0dB margin
>
>Difference between PAM8 and PAM12
>
>PAM8
>
>0.75+0.25/z
>
>10.2dB
>
>3.3dBm
>
>26.0dB
>
>6.1dB
>
>-124.9dBm/Hz
>
>
>
>2.2dB worse
>
>NONE
>
>4.8
>
>26.0
>
>26.0
>
>6.1
>
>6.1
>
>-122.7
>
>-120.7(Sailesh)
>
>Normalized to 0dB
>
>PAM12
>
>NONE
>
>9.8
>
>4.6
>
>29.1
>
>29.1
>
>5.3
>
>5.3
>
>-124.7
>
>-123.1(Sailesh)
>
>2dB worse
>
>2.4dB worse
>
>
>
>
>
>I, intentionally remove your final columns because it is not CORRECT. We
>will discuss later on the subject.
>
>
>
>First of all, SNR and noise margin matched very closely, within 0.1dB
>error, however, BGN for 0dB margin are off by about 2dB. We have to find
>out why but "Difference between PMA8 and 12" are about same. Yours are a
>little more than what I got but they are off only 0.3 to 0.4dB. At this
>stage, I agree with you that tolerable BGN is bigger for PAM8. However,
>note that PAM8 is the simple one without 0.75+0.25/z that you are
>proposing. Your proposal with 0.75+0.25/z filter is the worst. Albert is
>now getting similar results witn mine in time domain simulation and he
>is going to release it soon. In anyway, at this stage, I am not claiming
>simple PAM or with digital filter and let us assume simple PAM for both
>cases to compare. The important issue is whether that is 1~2dB or 2~4dB.
>Let's discuss the issues of intentionally removed columns in which you
>integrated to fs/2. You assumed that "ERXTERNAL NOISE" is fitting just
>fs/2 for both systems. That is not correct. The "EXTERNAL NOISE" could
>be at the higher frequency up to 1GHz or higher. Let's assume that we
>test performance with single tone at 450MHz in which PAM12 can tolerate
>a lot better than PAM8. What I am saying that in order to argue EMI
>tolerance, we need to define the spectral content of the "EXTERNAL
>NOISE".
>
>
>
>
>
>We have had four models for the system comparison back in March meeting.
>Those models do not have any "EXTERNAL NOISE" other than A-crosstalk. We
>assumed 10GBaseT is the self-crosstalk limited system. We believed that
>is reasonable and we did intensive study with the good IC
>considerations. As everyone agreed, PAM12 is the best noise margin
>solution in those criteria. Now, Sailesh raised the issue "Immunity for
>the EXTERNAL NOISE" at the last meeting.
>
>
>
>I am very positive to improve IEEE standard within a reasonable time
>frame. If A-crosstalk is not the biggest impairment or EMI noise is as
>big as that, let's redefine that and create 5th model on the top of the
>existing four models. Dan made a good point in his mail today. Will you
>be able to get a data of the frequency content of the EMI noise that
>everyone is comfortable with? I myself want to take a quick look at that
>even if you can not claim that is the general universal conditions. That
>should tell how many dB PAM8 is better in EMI tolerance (or non issue)
>and people can judge both A-crosstalk immunity and EMI tolerance at the
>same time.
>
>
>
>Hiroshi Takatori
>
>Keyeye Communications, Inc.
>
>
>
>
>
>
>
>-----Original Message-----
>From: stds-802-3-10gbt@IEEE.ORG [mailto:stds-802-3-10gbt@IEEE.ORG] On
>Behalf Of sailesh rao
>Sent: Monday, July 19, 2004 11:16 AM
>To: STDS-802-3-10GBT@listserv.ieee.org
>Subject: [10GBT] Updated Tables
>
>
>
>10GBT'ers:
>
>
>
>In the attached, I've updated the 3 tables in our July presentation
>based on
>
>the following:
>
>
>
>1. Change PAM12 symbol rate to 825Ms/s from 820Ms/s.
>
>2. Delete PAM10 entry.
>
>3. As Luc pointed out, add a 1.2dB emissions penalty for PAM12 due to
>its
>
>higher transmit PSD.
>
>4. As Jose pointed out, subtract 0.4dB from the PAM12 emissions penalty
>due
>
>to the THP peak voltage adjustment.
>
>
>
>Next, I integrated the WGN for 1E-12 BER over the Nyquist frequency
>range to
>
>get a "wideband noise tolerance" measure for the two proposals. Finally,
>I
>
>summed the noise immunity penalty and the emissions penalty for the
>PAM12
>
>proposal to form a "Total EMI Penalty" metric over the PAM8 approach.
>
>
>
>In Models 1 and 3, the penalty works out to be 2.6dB and 2.2dB
>respectively
>
>for PAM12 over PAM8. However, in Model 2, which represents the existing
>
>cabling infrastructure, the penalty for PAM12 over PAM8 works out to a
>
>whopping 4.0dB!!
>
>
>
>Regards,
>
>Sailesh Rao.
>
>srao@phyten.com
>
>
>
>_________________________________________________________________
>
>MSN Toolbar provides one-click access to Hotmail from any Web page -
>FREE
>
>download! http://toolbar.msn.click-url.com/go/onm00200413ave/direct/01/
>
>
>

_________________________________________________________________
Donít just search. Find. Check out the new MSN Search!
http://search.msn.click-url.com/go/onm00200636ave/direct/01/

solarsep_varlen7a_sr.m