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

[802.3af] Re: AW: Liaison report from ISO/IEC JTC 1 SC 25/WG 3 on mid-span pow er insertion


Your most recent question is a little different than your personal question that was covered on the cover page of the liaison letter.

The position of the current draft is that.
        If power is being inserted on the cabling from the switch/repeater towards the DTE (i.e. not mid-span insertion) then either the pairs on pins 1-2, 3-6 (Alternative A) or the pairs on pins 4-5, 7-8 (Alternative B), but not both.
        If the power is inserted on the cabling with a mid-span then the proposed standard mandates Alternative B.

The normal expectation is that the pairs used for power in a mid-span would not provide continuity through the mid-span. Those pairs on the unpowered side of the mid-span are expected to be unused but properly terminated. This is not mandated, however. The draft states in 33.2.2 that:
"Midspan PSE must use alternative B as described in section 33.2.1.Midspan PSE are limited to operation with 10BASE-T and 100BASE-TX systems,1000BASE-T is beyond the scope of this standard."

So it may be possible to build a Midspan PSE that would serve 1000BASE-T. Many people think this is possible. We have not investigated it during this project.

I hope that provides a definitive answer to your question.

Best regards,


At 08:10 AM 3/16/02 +0100, Pattay Walter wrote:
Dear Geoff,
thanks for this quick answer that - if complete and future proof - makes
life much easier.
But to be on the safe side, let me confirm explicitly: there is no need for
us to consider mid-span-power insertion for applications using all four
pairs for information transfer.
Kind regards

-----Ursprüngliche Nachricht-----
Von: Geoff Thompson [mailto:gthompso@xxxxxxxxxxxxxxxxxx]
Gesendet am: Sonntag, 10. März 2002 06:31
An: Pattay Walter
Cc: Alan Flatman; scarlson@xxxxxxxxxxxxx; thompson@xxxxxxxx;
Betreff: Re: Liaison report from ISO/IECJTC 1 SC 25/WG 3 on mid-span power
Dear Walter-

We will review your Liaison at our meeting in St. Louis during the upcoming

Re: your text on the cover of the Liaison Letter:
Due to the discussion in Kyoto it would be important for WG 3 to
know whether mid-span power insertion will always use separate
pairs, that is pairs which are not used for information transfer or
whether there is also a need to provide this function for pairs used
for information, e. g. when only two pairs are terminated at the TO.
The next meeting of WG 3 will take place near Washington,
D.C., USA, 2002-09-23/26.
The answer to the above question is as follows:
Mid-span will always use separate pairs.
End-to-end may use either the data pairs or the "separate pairs" but not

Therefore, mid-span may not be used in a situation where only 2 pairs are
terminated at the TO

I hope to be able to attend the upcoming meeting in the Washington, D.C.

Best regards,


At 11:36 AM 3/6/02 +0100, you wrote:

Dear Geoff,
please receive the liaison report as it was approved at our last meeting.
Unluckily we were short of time - as usual - and not all the output already
created on this subject was found complete or  ripe enough to be forwarded
to you in this report.
We definitely will do some more work on this subject at our next meeting. As
the outcome may be an interface to generic cabling in addition to those
defined in the present model, it promises to be a challenging subject. 
Hoping to see you there again I remain with best wishes
yours sincerely

| Geoffrey O. Thompson                    |
| Nortel Networks, Inc.  M/S SC5-02       |
| 4401 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: gthompso@xxxxxxxxxxxxxxxxxx     |