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

Re: [BP] ATTN: Magesh - Broadcom simulation results


Yes, I'm afraid we are using different values.  I'm simulating with 5% eye shutdown, which is consistent with the OIF spec.  I agree with your interpretation of our spec, but it appears we got it out of line with the intent.  The intent was to have it equivalent to the OIF spec (see page 4 of the same presentation from Steve Anderson you referenced).

The OIF uses the following definition:
The absolute value of the difference in the average width of a '1'
symbol or a '0' symbol and the ideal periodic time in a clock-like
repeating 0,1,0,1 sequence. Duty Cycle Distortion is part of the CBHPJ
distribution and is measured at the time-averaged signal level.

Both spec have the same value of 0.05UIpp, but the different definition results in a 2x difference in how it's applied.  


Thanks,        Joe


Joe Abler                                                             abler@us.ibm.com
IBM Systems & Technology Group            919-254-0573
High Speed Serial Link Solutions               919-254-9616 (fax)
3039 Cornwallis Road                                                                
Research Triangle Park, NC  27709


Magesh Valliappan <mageshv@BROADCOM.COM>

05/08/2006 07:18 PM
Please respond to Magesh Valliappan

       
        To:        STDS-802-3-BLADE@listserv.ieee.org
        cc:        
        Subject:        Re: [BP] ATTN: Magesh - Broadcom simulation results



Joe,

We may have interpreted the draft differently or we differ in
translating DCD's contribution to DJ.

Based on discussion at the Jan 2006 meeting and the presentation
http://www.ieee802.org/3/ap/public/feb06/anderson_01_0206.pdf,
we want to limit the clock period variation in alternate cycles to
within +/-2.5% or 0.05UIpp

In my simulations, the transmitter is modeled with "odd" clock periods
2.5% longer than
ideal and "even" clock cycles 2.5% shorter. When you plot these
waveforms with respect
to an ideal clock reference, the inner eye opening is 0.975UI. The
transition width
is 0.025UI. The eye closure or DJ component is 0.025UIpp.

From the definition of DCD in the draft, this transmitter has DCD =
0.05UIpp.  
"The Duty Cycle Distortion is defined as the unsigned ratio of the
difference in the mean pulse
width of a "1" pulse compared to the mean pulse width of a "0" pulse (as
measured at the mean of the high
and low voltage levels in a clock-like repeating 0,1,0,1 bit sequence)
and the nominal pulse width."

Regards,
Magesh


________________________________

                From: Joe M Abler [mailto:abler@us.ibm.com]
                Sent: Monday, May 08, 2006 5:16 PM
                To: Magesh Valliappan
                Cc: STDS-802-3-BLADE@listserv.ieee.org
                Subject: Re: [BP] ATTN: Magesh - Broadcom simulation results
               
               

                Thanks Magesh.  Could you clarify what amount of DCD you're
applying for the xtlk BER results?  Your table indications "DCD =
0.05UIpp (eye closure = 0.025UIpp)".  Which is it?  The spec calls for
up to 0.05UIpp closure.  Based on the following line where you state "DJ
at package output, excluding DCD = 0.105UIpp" and that your TJ =
0.28UIpp and RJ = 0.15UIpp it would appear that the amount of DCD you're
applying is 0.025UIpp.  That by the way is half what I apply, which
would essentially explain why you're not seeing the same level of impact
I am.
               
               
                Thanks,        Joe
               
               
                Joe Abler
abler@us.ibm.com
                IBM Systems & Technology Group            919-254-0573
                High Speed Serial Link Solutions               919-254-9616
(fax)
                3039 Cornwallis Road

                Research Triangle Park, NC  27709
               
               
               
                                 Magesh Valliappan <mageshv@BROADCOM.COM>

                05/08/2006 01:48 PM
                Please respond to Magesh Valliappan
               

                                         
                        To:        STDS-802-3-BLADE@listserv.ieee.org
                        cc:        
                        Subject:        Re: [BP] ATTN: Magesh - Broadcom
simulation results
               





                I have  updated the presentations with clarifications on
simulation conditions and  attached them to this email.
                The  transmitter conditions in both are based on the latest
version of the 802.3ap  draft and Charles' email
                regarding EIT simulations. The receiver  conditions are similar
to what I have presented at the last  meeting.
               
                 
                Magesh
                 
               
               
________________________________


                From: Joe M Abler [mailto:abler@US.IBM.COM]  
                Sent: Friday, May 05, 2006 5:52 PM
                To:  STDS-802-3-BLADE@listserv.ieee.org
                Subject: [BP] ATTN: Magesh -  Broadcom simulation results
               
               
                Hi Magesh,
                I'm trying to compare some of your simulation results  with mine
but can't get too far because I don't have sufficient information on
your simulation conditions.  You put some very basic information in your
presentations such as launch voltage and rise time, but I'd appreciate
it if  you could provide additional information on your assumptions for
jitter (Tx  & Rx breakdown including DCD), termination values, clock
offset, data  pattern type, random noise level, any accounting for CDR
algorithm, etc.   I'm currently looking at your last presentation
villiappan_c2_0506 and  would appreciate if you could send me the
conditions for those results.   More generally though, I'd like to
recommend that you update your other  presentations with such
information and perhaps ask Adam to repost them.   Anyone trying to
reference these in the future really wouldn't get full  benefit fro! m
them without knowing the conditions behind them.nt>  
               
               
                Thanks,         Joe
               
               
                Joe Abler
abler@us.ibm.com
                IBM Systems & Technology Group             919-254-0573
                High Speed Serial Link  Solutions               919-254-9616
(fax)
                3039 Cornwallis Road

                Research Triangle Park, NC   27709
               
               
               
                #### kr_xtlk_ber.pdf has been removed from this note on May 08,
2006 by Joe M Abler
                #### eit_sim_results.pdf has been removed from this note on May
08, 2006 by Joe M Abler