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

Re: [802.3af] My quest for low power PSEs continues




Mike-

It is easy to write down a "shall".
It is quite a bit less easy to actually conform.
You presume:
         * That there is already sufficient space for the legally mandated
labels
             (from each and every country that has to get its licks in)
         * That there is sufficient space left over from that (on a per
port basis) for all of the connectivity information called for by the
existing standard.
         * And now you want to mandate more labels which would, of course,
be required on a per port basis 'cause we don't spec anything else?

The PRIME directive that we get from OUR customers is more ports per box
AND shrink the vertical rack space requirement.

I think this problem is already over constrained.

On the other hand, I'm willing to look at (with a critical but somewhat
open mind) any text that you might propose.

Geoff


At 05:22 PM 1/22/2003 -0500, Mike_S_McCormack@xxxxxxxx wrote:


>Geoff:
>
>As recently as 1000BASE stuff we felt it was within our jurisdiction to
>require labeling:
>
>   40.8.2 Crossover function
>     ... Additionally, the MDI connector for a PHY that
>     implements the crossover function shall be marked
>     with the graphical symbol X.  ...
>
>I think we can still do it, 1000BASE is not ancient history.
>
>Personally, I would not be adverse to leaving the specification of the
>format of the power rating (amps & volts, watts, horses, etc.) open ended
>so that vendors who need to label for regulatory requirements could just
>say it once.  If you live in a location where no regulatory agency requires
>power labelling, well then we would add the burden, but then your label
>should be pretty much a wide open space anyways.
>
>Mike
>
>
>
>
>
>Geoff Thompson <gthompso@xxxxxxxxxxxxxxxxxx> on 01/22/2003 04:33:53 PM
>
>Sent by:  Geoff Thompson <gthompso@xxxxxxxxxxxxxxxxxx>
>
>
>To:   Mike S McCormack/US/3Com
>cc:   stds-802-3-pwrviamdi@xxxxxxxxxxxxx
>Subject:  Re: [802.3af] My quest for low power PSEs continues
>
>
>Mike-
>
>I believe that old hands would argue that the scope of the requirements
>that you uncovered was restricted to the equipment spec'd in their
>respective clauses.
>
>Ergo, 7.3.2 would be limited to 10 Mb/s and might well be overridden by
>some more permissive text in other clauses, e.g.
>14.8 MAU labeling
>It is recommended that each MAU (and supporting documentation) be labeled
>in a manner visible to the user
>with at least these parameters:
>a) Data rate capability in Mb/s,
>b) Power level in terms of maximum current drain (for external MAUs),
>c) Any applicable safety warnings, and
>d) Duplex capabilities.
>
>It was felt by then that mandatory labeling was onerous in terms of space
>available on newer equipment.
>
>geoff
>
>At 03:28 PM 1/22/2003 -0500, Mike_S_McCormack@xxxxxxxx wrote:
>
> >Ladies/Gentlemen:
> >
> >One of the persistent complaints about my request to allow small single
> >port PSEs (see comments 169 and 170) is that the IEEE can not mandate
> >labeling.  Well, I was casually reading the great opus and as I was
>closing
> >in on the first 150 pages of the spellbinding text, I notice the following
> >little item:
> >
> >   7.3.2 Signaling rate
> >     . . . To facilitate the configuration of operational
> >     systems, DTE and MAU devices shall be labeled with
> >     the actual signaling rate used with that device . . .
> >
> >My initial reaction was that this was an obvious mistake, the 802.3 group
> >would never has required labeling.  While it was obviously for a good
> >reason, someone must have messed up.
> >
> >I read on, then to my wonderment I came across:
> >
> >   8.4.2.2 Jacket marking
> >     The cable jacket shall be marked in a color
> >     contrasting with the background color of the jacket.
> >     The markings shall be spaced at 2.5 m  5 cm
> >     regularly along the entire length of the cable. . . .
> >
> >Now I was confused, our forbearers not only had committed the sin of
> >labeling in good faith, but apparently they had done it to the most
>mundane
> >level imaginable.  With two "shall"s in adjoining sentences no less!
>(Read
> >after the text I quoted and you find they even request a specific color
>for
> >the cable, but that is just a request, not a requirement.)
> >
> >Well, now I was curious, what else could have been so important that the
> >giants who walked the standard before us could have deemed it necessary.
> >Then I found the most interesting text:
> >
> >   9.9.3.2 Power consumption
> >     . . .
> >     The FOMAU shall be labeled externally to identify
> >     the maximum value of power supply current required
> >     by the device when the AUI mechanical connection
> >     is implemented.
> >     . . .
> >
> >So the founding fathers thought power was so important that they required
>a
> >label for it.  As a matter fact, as I slogged through the rest of the
> >document I discovered that when power was specified on an external
> >interface, there was just as likely a specification for labeling.
> >
> >I therefore put forward that there is sufficient precedence within 802.3
>to
> >allow us to require labeling of PSE output capabilities.  If you feel that
> >a label on the device is not sufficient to warn people about how much
>power
> >they can expect, we could also require that vendors documentation contain
> >the power rating, the precedent for that is:
> >
> >   27.6 Repeater labeling
> >     It is required that each repeater (and supporting
> >     documentation) shall be labeled in a manner visible
> >     to the user with these parameters:
> >
> >I think, if informing the user of the limitations of low wattage PSEs is a
> >barrier to legitimizing them, we can force the disclosure.  Once the user
> >can look on the device and see if it is suitable for their needs, what is
> >to keep us from filling that need?
> >
> >Mike
>(See attached file: C.htm)
>
>
>






----- End Included Message -----