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

RE: [802.3af] Your TR Comment



Chad, Goeff and all.
 
The scenario presented here is exactly what we have discussed over the reflector on the least two weeks and I have already commented on that.
I have suggested the remedy of mandating full diode bridge on data pairs and spare pairs (i.e. polarity insensitive in mode A and mode B) which is option 4
in the choices offered below.
 
Option 5 may work too however the bets is option 4 which gives polarity insensitive feature and solves other problems.
 
The difference in cost between option 4 and 5 is negligible even for low cost designs.
Power Dissipation is the same so option 4 is my favorite.
 
In addition, we also have a requirement for the PD that it will not source power to the MDI (or back power to the MDI) however we didn't put numbers behind this requirement and the best time to do this is now.
 
I am suggesting the following:
In page 60 line 19 paragraph 33.3.1 the draft requires "The PD shall not source power on its PI"
I suggest to change it to say:
 
"The PD shall not source power on its PI.
Compliance for this requirement shall be tested by applying a voltage of 57.0V across the PI at a polarity specified to mode A according to table 33-7 for MDI or MDI-x configuration and measuring a voltage less or equal to 2.8Vdc across the PI for Mode B.
 
The same requirement shall apply when a voltage of 57.0V across the PI at a polarity specified for mode B according to table 33-7 and measuring a voltage less or equal to 2.8Vdc across the PI for Mode A.
 
Yair.
 
 
 
 
-----Original Message-----
From: Geoff Thompson [mailto:gthompso@xxxxxxxxxxxxxxxxxx]
Sent: Tuesday, April 08, 2003 7:55 PM
To: cmjones@xxxxxxxxx
Cc: stds-802-3-pwrviamdi@xxxxxxxx
Subject: [802.3af] Your TR Comment

Chad-
Please post your diagram to the reflector ASAP
Geoff

==============================================

There exists a scenario whereby two or more PDs with different power mixing schemes,
when PSE-Midspan powered, may cause damage to a legacy 802.3 data switch to which
the PDs are attached which may be sharing unused pair termination.

Chad to provide a drawing if necessary.

SuggestedRemedy

There are several ways to prevent this damage to legacy 802.3 equipment  (listed in order
of least desireable):
1. Disallow the Midspan PSE
2. Force environment B for Midspans
3. Define an environment B- that can address this mixed scenario
4. Mandate a full diode bridge on each power pair in the PD
5. Mandate a half diode bridge on each power pair in the PD
Proposed Response
Comment Status X
Response Status O

Jones, Chad Cisco Systems


|=========================================|
| Geoffrey O. Thompson                    |
| Vice Chair,  IEEE 802                   |
| Nortel Networks, Inc.  M/S: P79/06/B04  |
| 4655 Great America Parkway              |
| P. O. Box 58185                         |
| Santa Clara, CA 95052-8185  USA         |
| Phone: +1 408 495 1339                  |
| Fax:   +1 408 495 5615                  |
| E-Mail: thompson@xxxxxxxx               |
| Please see the IEEE 802 web page at     |
| http://www.ieee802.org/