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

48V bus topics

Title: RE: power delivery question from Liaison report
For phone infrastructure they use a -48V bus, supported primarily from an AC input converted to a -48V output that is isolated from the AC line. The secondary means of supporting the -48V bus is from a battery bank. The 2 sources are diode or-ed together.  In actuality the system sometimes runs off the batteries all the time and the AC source just keeps the batteries charged. Typical -48V bus design criteria call for the voltage range to be from 36 to 72V. Converters operating off this bus are designed to handle the input swing.
There are lots of folks that make modules that operate from a 48V (-48V) to produce an isolated output voltage (12v, 5v, 3.3v, etc). The main difference with the 802.3 requirement is going from -48V to a 55V output.  This is easily done with todays power technology but not commonly available, YET. Waiting on a standard I guess.
Regulation tolerance can easily be controlled to within +-5% at the source.  55 x 1.05 = 57.75, below the 60V SELV limit.
What the load will see depends solely on its current draw and the resistance of the power path. The PowerDsine team did an excellent job of describing the path resistive drop effects.
So what you will have in this scenario is a Power supply (module):
Vin=-36 to -72V
Vout=55V +-5% isolated
Iout=350mA current limited per port.
For multiple port powering there are folks who isolate each port and some who group several ports together. If each port needs isolation then to maintain the 5% regulation some sort of regulator module will be needed for each port.  If each port does not need isolation, say groups of 5 to 10, then the power supply design get easier, all you have to do is limit the current to each (probably using the circuit that actually applies power to the line after discovery).

Ed Walker
Technical Staff

Analog Field Applications 
Texas Instruments Incorporated
7001 Weston Parkway #100
Cary NC 27513

INTERNET: ed_walker@xxxxxx
Office = 919-677-7061
Fax = 919-677-7030

-----Original Message-----
From: Rick Brooks [mailto:ribrooks@xxxxxxxxxxxxxxxxxx]
Sent: Wednesday, July 19, 2000 10:06 AM
To: Nick Tullius
Cc: 'stds-802-3-pwrviamdi@xxxxxxxx'
Subject: RE: power delivery question from Liaison report

Just to be clear, we are almost certainly not putting a battery voltage directly on the RJ-45.
At the very least each ethernet UTP port to frame ground needs 2250 VDC of isolation, this is an existing spec.
When the working group talks about 48 VDC, we are talking about what the PSE (power sourcing equipment)
is putting out on the RJ-45, and it will need to be at least somewhat regulated.
We are now discussing what the voltage range needs to be.

If someone wants to regulate a 48V battery voltage to obtain the PSE output voltage, then fine.
If someone wants to regulate a 120V/240V, or 208, 3 phase line voltage to obtain the PSE output voltage, then that's also fine.

The working group will not be involved in that decision; that it outside the scope of IEEE802.3af.
IEEE802.3af will only spec the output V, I  of the PSE, not the input power.

Also, the Liaison report was talking about restrictions beyond SELV.
Some (many?) of us would desire the absolute maximum voltage to be above 48 VDC, like maybe 52 to 56 VDC,
otherwise we will have to limit the available power.

- Rick