| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
Lennart: Finally spent a few more minutes thinking about Ilim/Tlim. A “short circuit” condition is basically any condition that causes the PSE output to meet or exceed Ilim (Ilim, a PSE-specific variable, of course). I agree, and pointed out in some rejected comments, that there is a fundamental inconsistency in text in the following 3 sentences: From 145.2.10.8: If a short circuit condition is detected on a pairset, power removal from that pairset shall begin within TLIM as defined in Table 145–16. (I’m taking Tlim to be a PSE-specific variable here that has a well-defined range.) If IPort-2P exceeds the PSE lowerbound template, the PSE output voltage may drop below VPort_PSE-2P min. A PSE in a power on state may remove power from that pairset without regard to TLIM when the pairset voltage no longer meets the VPort_PSE-2P specification. The final of these 3 sentences washes out any behavior implied by the first 2 sentences. The practical truth is, a PSE limiting output current will drop outside the Vport_PSE-2P specification in the vast majority of cases. Sentence #2 addresses this. No doubt, this then puts strain on the PSE controller as it dissipates power. That’s why sentence #3 exists. A further practical truth is that many PSE’s that hit their Ilim threshold will sustain power, perhaps with current foldback to preserve the PSE controller, until at least Tlim_min. Other PSE’s will implement policy that when in current limit and voltage drops, power goes off immediately without regard to Tlim_min. 802.3at essentially enabled this duality and 802.3bt is continuing it. I do see that sentence #3 here is completely redundant to the sentence in 145.2.10.1: A PSE in a power on state may remove power from a pairset when the pairset voltage no longer meets the VPort_PSE-2P specification. I would propose that we don’t need this sentence given the overlap. Finally, I would propose setting Tlim_max in Table 145-16 to hard number 75 msec. I don’t think this would break anything and it would eliminate the absurd case of a PSE operating permanently in a current limited state. Regards, Pete J Peter Johnson Sifos Technologies 1061 East St. Tewksbury, MA 01876 Sifos is relocating on June 1, 2018 to: 1 Technology Drive Suite 100 Andover, MA 01810 To unsubscribe from the STDS-802-3-4PPOE list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-4PPOE&A=1 |