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

[802.3_4PPOE] PSE State Diagram - Dual Sig PD Conditions leading to ==> DISABLED



Hi All,

I'm working on (what I hope to be) a complete Type 3 and Type 4 PSE State Diagram to bring in to the next revision of the draft.

In the last meeting, it seemed like an overall satisfaction with the dual-state-machine approach and I plan to pursue that further.

One area that I am soliciting input on... What are the conditions that would cause a single channel of the PSE to communicate to the PSE_Controller that it wants to power down both pair-sets?

With the current approach, both pair-sets could be powered, and if one experiences an over-current, would move into the FAULT state for that pair set, ted_timer set then the pair-set would go back to IDLE. The other pair-set would continue to operate as normal unless it also experienced a condition that led it to move out of the POWERING state.

So, are there any conditions that would occur on a pair-set that would tell the PSE its time to return both Pair-Set controllers to the DISABLE state and begin the Connection_Check again?

It occurs to me that having a variable (BeamMeUpA/B) for each pair set that indicates a desire to move back to DISABLE can be used and if both variables (A and B) are set, that would communicate the need to return to the top of the state diagram. For instance, a state inserted into T10 that communicates "No Detection" on the pair-set. If pair-setA opened or shorted, went through FAULT, then returned to IDLE and attempted to perform DETECTION but could not, it would take T10, set the BeamMeUpA variable, then return to IDLE where it would continue to loop on attempting DETECTION until the other pair-set had the same condition, or the PSE management decided to reset the port.

Example:



Of course, these names are just to get your attention, and the BeamMeUP variable would have to be set to False up on the top level of the state machine.

There may be other cases where both pair-sets need a means to communicate to the main PSE Controller (Hey, Get US out of here!) Please let me know if you think of some.

Your feedback is appreciated. You can email me directly to reduce reflector traffic, or open a discussion if you feel it will be helpful.

Regards,

Dan Dove
Chief Consultant
Dove Networking Solutions
530-906-3683 - Mobile