| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
|
All, the group has finalized a comment for
716.433.1.101. The comment has evolved considerably from what was first proposed a few weeks back. In that, a detailed description of PSE operation was added so that the TG could understand the who IEEE 802.3 could interpret their text as applying and
how the text was incompatible with PSE operation. The question now is, with the new first paragraph, do we still need this description? Is it too much detail? Can some detail be removed? Please take some time to review and reply to me with any thoughts. Comment: The sentence “The circuit shall not reset automatically.” seems to disallow products that have multiple levels of current limit
with different protections at each limit. The 802.3 PoE specification requires a (self-resetting) current foldback when the maximum rated supply current is exceeded. Outside the scope of the IEEE 802.3 specification, products usually have a fuse to meet the
stated overcurrent protection required by specifications such as IEC 62368. The text should be clarified so that it is more explicit that multiple protection levels are allowed as long as one of the protections does not reset automatically. Power Sourcing Equipment (PSE) operation as specified in IEEE P802.3bt is described in the following: PSEs have current limits Ilim and Icut. Icut is a ‘PSE may shut off power to the port’ current limit and Ilim is a ‘PSE shall shut off power to the port’ current limit. Icut is reached when a Power Device (PD)
draws more power than it advertised during classification for longer than 75ms max. Ilim is reached when the PD exceeds this power by a given percentage that varies according to the Type (this percentage decreases with increase power output/draw ~14% max).
The Ilim min timing is Type dependent but is 75ms max for any Type. Most PSEs will choose the lower Tlim numbers (6 or 10ms) as thermal dissipation can be problematic at these operating points. In either of these cases (Icut or Ilim), the PSE moves back to
detection after shutting off power to the port. If the PD presents valid detection and classification signatures and falls within the startup inrush current limits, power will be applied again automatically. This is by design. Customers do not want to manually
intervene in a misbehaving PD (misbehaving being a PD product that is poorly designed such that it exceeds the current limits occasionally but is not damaged or a hazard). The PSE has an error delay timing spec that requires a 750ms delay before subsequent
powering attempts. These timing requirements limit the duty cycle of any error event to less than 10% to ensure that there is no heating due to successive fault events. IEEE 802.3 requests that the language of this section be modified to allow the above described behavior as hundreds of millions of ports are already in existence that will not conform and likely cannot be upgraded
to comply. One solution would be to clarify it is just the overload protection that cannot reset automatically. This can be accomplished by making
a new section just for overload and adding a sentence to the overcurrent section: After the list of principle measures in 433.1.101 add: The measure used to protect against thermal effects and overcurrent may reset automatically. 716.433.1.101a Protection against overload Thanks! Chad Jones Tech Lead, Cisco Systems Chair, IEEE P802.3bt 4PPoE Task Force Principal, NFPA 70 CMP3 To unsubscribe from the STDS-802-3-SCC18 list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-SCC18&A=1 |