|Thread Links||Date Links|
|Thread Prev||Thread Next||Thread Index||Date Prev||Date Next||Date Index|
you bring up a lot of issues, all of which are important.
Yes, I agree that we should keep the specifics out of the specs as much as possible.
Still, the general method of implementation does have an impact on practicability
The maximum current we have been given is not an instantaneous value, rather a continuous value which is there
to protect the wire and the connections. These have a rather long thermal time constant.
Instantaneous current limit could and will be quite a bit higher.
After all, we will probably have some output capacitance from the sourcing power supply.
To my mind, the hipot test has to do with safety and fault propagation.
Whatever the requirements are, they should apply whether or not the DTE power is on the link.
It would be very unwise to violate the isolation requirement when the power is on.
The rules of environment A or B apply.
The current limit subject should be completely independent from hipot.
The PSE power supply is driving a load which is floating from ground, with the same isolation requirements.
To test hipot on the PSE, I would say to hook each pin of the RJ45 under test to the hipot tester through separate high value resistors.
When hipot is passing, there will be no voltage drop across these resistors anyway.
Some RJ45 pins can be hooked directly together and use the same high value resistor.
Then place the DTE minimum load, and discovery network (whatever it turns out to be) across the appropriate power sourcing pins.
Now DTE power can be on or off during the test.
To test hipot with DTE power on, it will first be necessary to complete discovery, then establish DTE power into
the minimum load, finally turn on the hipot tester.
From: R karam [SMTP:rkaram@xxxxxxxxx]
Sent: Thursday, August 10, 2000 9:05 AM
To: Brooks, Rick [SC5:321:EXCH]; stds-802-3-pwrviamdi@xxxxxxxx
Subject: More details...
I would have to revisit the PTC specs at this point, but on to the "active"
current limit, we will find that a hand full of analog houses have the process
and the expertise to deliver parts that can take the transients on and survive.
the good news is these parts do exist.
Since I am new to the standards, and the purpose I guess, is for us is to spec
a current limit in this case, we would have to guarantee ahead that an implementation
exist, yet not force any implementation. at this point it will be crowded around the
RJ45 to do the job properly. Until the volumes get higher and better alternatives are
Are we set on meeting the 2000v high potential spec for 802.3 when the connection is
live? (POWER IS ON) I heard that PBX's do not meet high-pot requirements- Arlan?
and may be someone from LUCENT can comment?
and if the active current limit is set to 350ma when the connection is live how does that factor into the high pot
breakdown?. I think we have been staying away from the isolation requirements, but if we
are talking current limits, then reality now sets in and we must face the details.
Come to think of it, how would the 2000v High pot be tested when the connection is live?
Currently we test high-pot with the board off NO POWER, and thank's to the magnetic the isolation
across it is NOT in question.