RE: Candidate Detection Tolerance Allocation and Test Limits
I disagree, see below
At 12:09 PM 11/21/00 -0800, Schwartz, Peter wrote:
My preference would be the 5% detection
tolerance, This eases the
requirements on the reference voltages (or reference currents) used by
PSE during its detection process, while still allowing for some degree of
power-draw "enumeration" by resistor value.
On the second issue, I believe that power-draw enumeration is an
better handled in software than in hardware. I don't believe that
economical, in the final analysis, to use a 100Mb/s Ethernet port
trickle-charge a 12V or 24V exit sign battery,
First, no one is saying that it will be a 100 Mb/s port. It is more
likely 10 Mb/s shared.
Second, we have seen some pretty compelling presentations that it is very
likely to be more economical over time.
and - just speculating - I
don't even know if that would comply with code.
Current code perhaps but it is fairly clear that it is not legal to
standardize implementations where performance can be used (the Hydrolevel
case). I would expect code to migrate when proven product is
More importantly, some
devices may need to be given priority access to power (if the CEO wants
plug his PDA onto the net, I don't want my .mp3 player telling him
That may be so but it is beyond the scope of the current project.