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

RE: [bp] March Meeting Goals and Request for Presentations




Petre,

As John points out, there was discussion on this topic at the January interim and the following are the points of information that I recall from that discussion:

1.  ATCA defines a 21" backplane with 5" per line/fabric card for total of 31".
2.  Not every chassis that would wish to take advantage of Backplane Ethernet fits the ATCA model.  As John points out, the break down of interconnect distance between line card/backplane/fabric card can vary widely amongst these hypothetical chassis.  As an example, while you proposed 4"/32"/4", one could also envision a system with a breakdown of 6"/22"/12".
3.  It was the concensus of the group that 40" and two connectors covered all the hypothetical cases of interest.

For the purpose of your model, you will need to use your judgement as to what constitutes a reasonable breakdown of the interconnect distance (you may even wish to consider multiple scenarios).  If anyone has information that they can share on this subject, please post it to the reflector.

Thank you,
-Adam



-----Original Message-----
From: owner-stds-802-3-blade@majordomo.ieee.org
[mailto:owner-stds-802-3-blade@majordomo.ieee.org]On Behalf Of Petre
Popescu
Sent: Monday, February 16, 2004 2:25 PM
To: DAmbrosia, John F
Cc: stds-802-3-blade@ieee.org
Subject: Re: [bp] March Meeting Goals and Request for Presentations



John,
Thanks for your comments. Assuming that we will define a channel, we 
will need a compliance test
defined which will include some copper traces from a test connector and 
the backplane via  on the
board side. I will assume the 40" backplane copper traces.
Petre

DAmbrosia, John F wrote:

>Petre,
>This issue was discussed during the meeting, and the group as a whole
>decided not to specify card length, because it can vary so much in a real
>system environment.  
>
>In addition, the card length will have a substantial impact on return loss.
>I am not sure you will get to where you want to be.
>
>john
>
>-----Original Message-----
>From: owner-stds-802-3-blade@majordomo.ieee.org
>[mailto:owner-stds-802-3-blade@majordomo.ieee.org] On Behalf Of Petre
>Popescu
>Sent: Monday, February 16, 2004 11:09 AM
>To: Healey, Adam B (Adam)
>Cc: stds-802-3-blade@ieee.org
>Subject: Re: [bp] March Meeting Goals and Request for Presentations
>
>
>Hi Adam,
>We are working on the channel model for 40" copper traces on FR4 and two 
>connectors.
>For clarification, we propose to define the model as two sets of 4" 
>copper traces (daughter
>cards) and one 32" copper trace on the backplane. Please let me know if  
>this definition is
>what you have in mind.
>Thanks, Petre
>Petre Popescu
>Quake Technologies
>