Discovery process
Geoff,
I'm a concern raised by the two mails here below.
On one side, it is necessary to take into account a possible
low current induced by an indicator and in the other side
it is necessary the power sink draw a minimum current
to hold the power sent via MDI.
Question : how will it be possible to distinguish these two
types of current? and how will it be possible to detect the
disconnection of the terminal?
Regards.
Gérard.
Objet: Re: Discovery process
Date: Wed, 19 Apr 2000 11:07:56 -0700
De: "Geoff Thompson" <gthompso@xxxxxxxxxxxxxxxxxx>
A: verger <verger@xxxxxxxxxxxxxxxxxx>
Copies à: stds-802-3-pwrviamdi@xxxxxxxx, Raymond Gass
<raymond.gass@xxxxxxxxxxxxxxxxxx>
Gérard-
This issue was discussed in committee. The decision was that:
1) If the "discovery characteristic" was found then power would
be sent
2) For a device to KEEP the ability to draw power without
restarting discovery there would be a requirement for the
load to draw a minimum amount of power (5 ma was
mentioned) over the link.
Geoff
>Hi guys,
>
>I am thinking about the different proposals done, concerning
>the discovery process, that in case where the terminal is
>locally powered then the power via MDI is not sent.
>Unfortunately, this case will be unavoidable since very often
>the terminal is installed by the end user who is not awared
>about the properties of the network infrastructure.
>Moreover the network is able to provide a safe power
>distribution (power with battery-back-up) and therefore
>it will be better to provide such a power to any terminals.
>
>What are you thinking about this problem?
>
>Our solution is to analyse the impedance under alternative
>voltage and this analyse stays efficient even the terminal
>is locally powered and therefore this allows providing a
>safe power to terminal whatever its state and in case of mains
>breaking holding the current communication with this terminal.
>
>Gérard
------------------------------------------------------------------
Objet: Re: Continuous Current Capacity of Panels, Jacks, and Plugs
Date: Thu, 10 Feb 2000 16:52:24 -0800
De: "Geoff Thompson" <gthompso@xxxxxxxxxxxxxxxxxx>
A: Raymond Gass <Raymond.Gass@xxxxxxxxxxxxxxxxxx>
Copies à: stds-802-3-pwrviamdi@xxxxxxxx, jhkee@xxxxxxxxxx
Raymond-
This proposal sounds like it is in line with the discovery proposals
that the group to be homing in on. The one item that I would take
direct issue with would be the null value. I believe that we will end
up defining a minimum, non-zero current that we will decide
represents the non-connected condition. This will be to make
provision for a small amount of current to be leaked into the
TO/socket itself so that an indicator can be lit.
Geoff
>Geoff,
>
>we have been looking at the various proposals for
>remote powering / discovery process, and we would
>like to stimulate discussion about a simplified
>solution.
>
>the benefits of this proposal are as follows:
>- it does not introduce extra hardware in the power
>sink, and thus no extra cost in the terminal part
>- no impact on the data transmission
>- it can be used whatever choice for powering
>(phantom or direct mode, on data or on spare wires)
>
>the method is based on a measurement of the
>value of the current, performed by the "check
>box" associated to the power source:
>
>- during discovery process, a low frequency (some kHz),
>low voltage (2-3 V) is sent, that makes it possible to
>measure the line impedance (whatever line), and to
>decide if this particular line needs to be fed or not,
>based on the peak value of the AC current
>
>- during feeding phase, DC current is being monitored.
>Power feeding is maintained as long as DC current
>is not null. Discovery process is re-initiated when
>DC current is null
>
>we have prepared some slides to describe further
>this solution, and we are ready to make a presentation at the
>next 802.3 meeting
>
>thanks for your comments
>
>Raymond