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

Re: [802.3_4PPOE] End to End Channel Pair to Pair Unbalance Ad Hoc Presentation


Hi Jeff,

Nice work. Thanks for the presentation you send.

Due to the fact that for your presentation we allocate the last 20minutes and we must finish the business of the first 40minutes, I encourage the team to comment on Jeff presentation over the reflector so we can save time during the meeting for the issues that can't be resolved over the reflector and allow Jeff maximum possible time.

Please find my comments below.





-Title slide: The ad hoc name is End to End Channel Pair to Pair Resistance Unbalance. Please update accordingly.

-Slide 3: The PSE PI implies implementation. If you replace Vps, Vps1, Vps 2 with 4 voltages V1,V2, V3 and V4 and define what you want to spec on these 4 voltages, you will get truly implementation independent model.

-Slide 4: It looks like that you are suggesting what we have now at the adhoc. If this is correct please say so and we can continue with consensus on this slide if it is different please say clearly in what it is different. 

-Slide 5:

a) You have error in the column " if by %" for 1m length. It can't be 0.004% nor 0.004.

b) The 7% number which is a result of my work (see adhoc presentation which we will address this week) assumes some minimum requirements. Not clear if you have address it. (To guarantee 7% regardless of cordage and cable length and regardless of number of connectors we recommend 7% with the requirement that PSE PI and PD PI will have minimum resistance of 0.1 ohm(TBD) each. Otherwise, numbers can be as high as 19%.)

c) Also check with CAT6A AWG23 cable. You need both to verify were you will get maximum current over pair. We hope that it will happen at the 12.5 ohm channel but it is not sure. You have two equations to test prior answering this questions. The reason is that maximum pair current is (It/2)+P2PRUNB*It/2.  Now It_max is function of channel resistance. It is maximum at 100m, 12.5 ohm. The P2PRUNB in this case is X. In short channels, The P2PRUNB is Y and Y>X but It is lower. So you need to check if It(short_cable)*Y<>It(long_cable)*X. Please see adhoc material annex B. We expect worst case current will be at 100m, 12.5ohm but it is not sure until worst case end to end channel resistance unbalance (and voltage etc.) numbers will be finalized.

-Slide 6:

(a) If you require delta resistance maximum, you do force implementation. We need to check if under your definitions, the adhoc database table is included. If not, we have a problem since the adhoc table represents some existing implementation. Did you check it?

(b)  You said "Delta resistance values and delta voltage values are the best way to keep the PI definition implementation independent " this would be wonderful if it was scientifically true but due to work done on the channel only with different use cases, you have to specify Rmin also as a requirement. Otherwise, only the channel will get to 19% which correct if the channel was tested as standalone and not correct if the channel was part of the system were you have actually more resistance in the due to the presence of PSE PI and PD PI.

(c) You said in 3rd bullet: "We are NOT attempting to telling the OEMs how much resistance they can have in a PSE or PD". I fully agree that it is what we want and it is a good _expression_  implementation independent approach. The problem is that if we want to set the channel P2PRUNB to 7%, you need to add resistance to the channel over each pair externally from the PSE and PD PI. Which is part of the previous item (a). We need to discuss it maybe we have other solution so what you are proposing can be valid.

(d)  You said: "(A single point PI measurement cannot fully characterize a PSE or PI.)". Not clear how this bullet relevant to the rest of this slide. We need only single maximum number for each parameter that define the unbalance property. As a result it will be "worst case" in the entire operating range of interest. How many measurements to do in order to test for compliance is different question.



Slide 7:

(a)    I see AC MPS on positive and negative power rails, please explain?

(b)   Why AC MPS is part of 4P operation? Do you mean to AC disconnect circuitry? Are you addressing possible case of having AC disconnect circuitry on 4P systems?


Slide 8: What is the difference between slide 8 to 7? They look the same?

Slide 9:  Did you use the adhoc data base number or other? Please explain.

Slide 11:

-The title of the right column is not clear.

-What is long channel? Please add numbers for the parameters in the slide so we can see what you have used and compare to other work.

-What is short channel? Please add numbers for the parameters in the slide so we can see what you have used and compare to other work.

-What is the total PD current

-How different are the results from adhoc results?

-If different database was used? How we can compare?









From: Jeff Heath [mailto:jheath@xxxxxxxxxx]
Sent: Saturday, June 28, 2014 1:51 AM
To: STDS-802-3-4PPOE@xxxxxxxxxxxxxxxxx
Subject: [802.3_4PPOE] End to End Channel Pair to Pair Unbalance Ad Hoc Presentation




Here is the material I will be presenting in next Tuesday’s ad hoc.  Clearly there will not be time to go over all of the material in the 20 minutes allocated.  Therefore I will be covering the methodology, the PSE and PD PIs,  and Link Segment, or Channel, briefly and sharing some preliminary simulation results.




Jeff Heath
Design Center Manager

Description: Linear Technology Corporation




402 East Carrillo Street, Suite D


Santa Barbara, California 93101