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

Re: [8023-POEP] Classification Concept Based on Diab/Schindler Motion



Clay and all,

See below few comments.
Yair
--------------------

Page 1: No comments
Page 2: No comments.

Page 3: 2nd RED bullet: I prefer the previous text that says "...it can
either ping twice or just once".
The rational is: If we say that "we don't care" then it will be open to
different interpretations in the future by test equipment vendors.
Adding the information is positively and explicitly says that you can do
either options and PD needs to take these both possibilities in account
(to prevent a case that after 2nd class PD circuitry behavior is
undefined.)

Page 3: There is no text explaining the requirements and behavior of an
AT PD that requires more then 12.95W and connected to AF PSE.
We may add something like: "Type 2 PD shall supply indication if
connected to Type 1 PSE or Type 2 PSE. The implementation details of
such indication are beyond the scope of this standard". Type 2 PD can
consume more then 12.95W if connected to Type 1 PSE after successful
Detection, classification, startup phase through Type 1 PD parameters
and reaching steady state operation. After this point in time if
overload occurs due to the fact the PSE is a Type 1 or The PSE is a Type
2 with not enough power (e.g. limited budget etc.), the behavior of the
PD is undefined. 
(I must pointing out that the above text that I am suggesting raise the
following question:
1. Is PD indication objective is really needed?
   I remember the historical discussions that says that we don't want
complains from user that PD is bad and not working when connected to un
known PoE port type (AF or AT).
In my opinion (and I would like Switch and PD vendors to comment on
this) such complains probably will be negligible due to the following
reasons:
When I as a user of IP Phone or other device connect it to the RJ45 and
it is not working, do I call the Switch or PD vendor to check what is
the problem? I call first to my IT manager and he tells me that the port
is af or at or not enough power or what ever? Now every thing is
controlled and known so why we have to pay for 2nd finger in PSE and
Indication circuitry in PD for something that personally I am not sure
is real and I may mistaken here therefore please comment on this.
2. If I am correct then PSE and PD became even simpler... (We started
with 1000 classes and ended with 4 classes isn't it funny...)
3. I am not trying to make revolutions here (I am still OK with the
current decisions and consensus) I just remembered how we started and I
recheck the rational behind what we did since we get more wisdom and
realistic about the real needs .. I hope..

Page 5: No comments. I guess in time after some of us will do some
hardware tests we may need to twick some numbers however I don't see big
issues here.
If somebody already did some tests and has data to show that it can be
easily implemented in PD and PSE it will be wonderful.

Page 6: The "ON" state in both at PSE branch and af PSE branch is 42V.
PD threshold voltage numbers are not going to be changed (this was at
least the consensus of the Vport ad hoc).

Page 6: Class event 3: It says I=40Ma. I suggest to consider that I will
not be defined since it is not important and can be implementation
specific as long as through classification range it is limited to
802.3af limits.

Page 7: Update state flow according to updates required in page 6.

Page 7: It says that PD should not present 25K during mark range but in
page 6 the mark range allows transition to signature detection which may
be impossible to due. We specify that at Vmark range signature should be
invalid numbers but we allow transition to detection which can not be
met.
We need to discuss it more.

Have fun.

Yair

      
 

-----Original Message-----
From: owner-stds-802-3-poep@xxxxxxxx
[mailto:owner-stds-802-3-poep@xxxxxxxx] On Behalf Of Clay Stanford (LTC)
Sent: Tuesday, December 12, 2006 6:06 AM
To: STDS-802-3-POEP@xxxxxxxxxxxxxxxxx
Subject: [8023-POEP] Classification Concept Based on Diab/Schindler
Motion

See attachment.

We will discuss on Wednesday 12/13.

thanks

clay


 
________________________________________________________________________
____________
Need a quick answer? Get one in minutes from people who know.
Ask your question on www.Answers.yahoo.com

ping pong mark low r19.pdf