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

Re: [802.3_PDCC] Changing resistance due to temperature rise

Peter - please be advised that the 802.3 working group does not provide official interpretations.
The behavior of the PSE and PD are found in the state diagrams in clause 104.
Below are my personal interpretations.
There is no specification that I know of that requires (or allows) the PSE to change the power class unilaterally, or for the PSE to provide a class other than the one requested via SCCP.  I also know of no requirements on cycling for failed startups.  These are all things that an implementer might consider (the PD indicates the requested class, and the PSE determines if the request is a valid class & there is power available - it doesn't respond with a different class - I suppose if the PSE declined the request, the PD might request a different class...)
Therefore, the behavior you are asking about is technically outside the standard.
Now, could it happen within the standard...

Classification is performed prior to power_up, not when the PSE is powering the PD, so cycling or change of class requires the PD to shut down or the PSE to shut down due to overload or insufficient power available.
I don't think the PD or PSE would be required to shut down per the standard. 

The effect of increasing link segment resistance results in increased voltage drop and therefore a demand for increased current, so shutoff by the PSE or the PD needs to be considered.   I don't believe the PSE would shut off due to an overload (104.4.7.2.1), but could remove power for sourcing more than P_class (79W) per 104.4.7.7.  One would need to examine in detail the situation to determine whether the powering attempts to restart - but there is margin in the system. 
Looking at this, if the PSE keeps to specification and maintains voltage, V_pse min is 50V, which, I calculate, provides 52W to a PD over a 10 ohm loop by sourcing 1.475A (which means the PSE shouldn't shut off for sourcing more than P_class), resulting in a V_PD over the 10 ohm loop of 35.25V, which is greater than V_PD min, so the PD shouldn't shut off either.

Anyhow, you can see that the trades on current draw and power drawn are more complex than just a fixed current.   Current will change (increase) with increasing heating, but there is margin in the system.




From: Peter Fischer <000042efa5d19355-dmarc-request@xxxxxxxxxxxxxxxxx>
Sent: Tuesday, April 28, 2026 12:56 AM
To: STDS-802-3-PDCC@xxxxxxxxxxxxxxxxx <STDS-802-3-PDCC@xxxxxxxxxxxxxxxxx>
Subject: [802.3_PDCC] Changing resistance due to temperature rise

Dear all
 
I have a question I would like to discuss during one of the next PDCC meetings:
 
Let us assume a procedure as follows:
 
Initiale situation: SPE-Link has 9 Ohm DC loop resistance
SPE device, which requires power class 15 is connected over the cabling to the switch and therefore to the power source.
Due to relativ resistance (ohm per meter) and installation environment conditions the cabling is getting hotter and therefore the DC loop resistance is rising to 10 Ohm.
 
What will happen?
  • Will the power source automatically switch to the next lower power class?
  • Will the device send information to the switch that the power class is not as expected?
  • After a while, the cabling has cooled down again and power class 15 would be possible again. Will the power source switch on? For how many times this cycle will be repeated? When will the electronics give up and report an error?
These points are important as it might impact the way cabling has to be defined to avoid certain unwanted behaviour.
 
Best regards
 
Peter

To unsubscribe from the STDS-802-3-PDCC list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-PDCC&A=1

To unsubscribe from the STDS-802-3-PDCC list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-PDCC&A=1