Re: [802.3_4PPOE] Backfeed presentation + baseline

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

Chad,

I have good arguments for why 20.5V.

Between 20.5V and 30V I am still checking. I wonder what is PSE is connected to the PD and PD is not requesting power which means we will see 30V. Now, if backfeed is not met in the ideal diode bridge, the FETs are ON and locked to this position (due to the problems in these briges).

Now when PD will require power and Rsig is connected, backfeed will be violated also in the 2.8-10.1 range which is a problem.

Please check this if you see error in my anaylsis.

Regarding what you have said“You asked my opinion. I gave it to you. How can you argue with my opinion? I mean, you can argue but it’s my opinion so I will ignore.”

Yair: So, you said that I can argue so we are in agreement. Moreover, it is my job to argue with you .. 😊 more seriously, I am looking for facts to see if we are good or not.

Yair

From: Chad Jones (cmjones) [mailto:cmjones@xxxxxxxxx]
Sent: Monday, May 14, 2018 5:07 PM
To: Yair Darshan <YDarshan@xxxxxxxxxxxxx>; STDS-802-3-4PPOE@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_4PPOE] Backfeed presentation + baseline

EXTERNAL EMAIL

You asked my opinion. I gave it to you. How can you argue with my opinion? I mean, you can argue but it’s my opinion so I will ignore.

Regards,

Chad Jones

Tech Lead, Cisco Systems

Chair, IEEE P802.3bt 4PPoE Task Force

Principal, NFPA 70 CMP3

From: Yair Darshan <YDarshan@xxxxxxxxxxxxx>
Date: Monday, May 14, 2018 at 9:57 AM
To: Chad Jones <cmjones@xxxxxxxxx>, 4PPOE Reflector <STDS-802-3-4PPOE@xxxxxxxxxxxxxxxxx>
Subject: RE: [802.3_4PPOE] Backfeed presentation + baseline

I Chad,

Never say never. We made a good progress since the time some said, “we have no issues at all” until now we have address more than 10 topics and found 3 at least that corrected by additional text.

And I don’t believe it is FUD. This is in my opinion the difference between a good and clear spec to something else.

How would you know without checking if you have a problem?

Yair

From: Chad Jones (cmjones) [mailto:cmjones@xxxxxxxxx]
Sent: Monday, May 14, 2018 4:38 PM
To: Yair Darshan <YDarshan@xxxxxxxxxxxxx>; STDS-802-3-4PPOE@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_4PPOE] Backfeed presentation + baseline

EXTERNAL EMAIL

In my opinion, backfeed is purely about detection which means 10.5V. There is slop in the classification range that allows us to ignore there, but I could be convinced 20.5V. after class, none of it matters. You could never convince me that 30V is required.

But, I don’t think any of this needs to happen. This theoretical PD that backfeeds under 3P power enough to spoil detection fails other PD requirements (as this is now a 4P standard). Backfeed is a 2P spec, written for a purely 2P problem. This additional restriction is FUD that is making us spin around the drain and threatening to make us miss the July submission date.

Chad Jones

Tech Lead, Cisco Systems

Chair, IEEE P802.3bt 4PPoE Task Force

Principal, NFPA 70 CMP3

From: Yair Darshan <YDarshan@xxxxxxxxxxxxx>
Date: Monday, May 14, 2018 at 8:35 AM
To: Chad Jones <cmjones@xxxxxxxxx>, 4PPOE Reflector <STDS-802-3-4PPOE@xxxxxxxxxxxxxxxxx>
Subject: RE: [802.3_4PPOE] Backfeed presentation + baseline

Chad,

In your opinion, from which voltage in 3-pair mode, the PD doesn’t need to meet backfeed requirement (2.8V, 28uA)? 10.5V and up? 20.5V and up ? 30V and up?

Yair

From: Chad Jones (cmjones) [mailto:cmjones@xxxxxxxxx]
Sent: Friday, May 11, 2018 4:50 PM
To: STDS-802-3-4PPOE@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_4PPOE] Backfeed presentation + baseline

EXTERNAL EMAIL

Heath, your review includes this statement: Moving the PD requirement to the PSE is onerous and moves requirements from the out-of-spec PD to the compliant PSE.

I disagree with this statement. This is not an out of spec PD. The PD 100% complies to the backfeed spec as written in Clause 33. We carried forward that requirement simply because we copied 33. As I’ve said repeatedly, the backfeed requirement is a true 2P requirement. When you apply 3P power, you are not performing this test and the requirement no longer applies. The real first question we should have asked is does a true 2P test belong in a 4P standard.

This is no longer simply a ‘FET bridge’ problem. It’s been reported on this reflector that schottky diodes also fail the misapplied 3P backfeed test. The PD is only “misbehaving” because the PSE is improperly applying 3P power. If you want to outlaw this behavior, then you are mandating silicon diodes. This standard fails if that is a requirement.

This is a PSE misbehavior that we decided to make compliant. The PSE holds the burden to fix.

Chad Jones

Tech Lead, Cisco Systems

Chair, IEEE P802.3bt 4PPoE Task Force

Principal, NFPA 70 CMP3

From: "Stewart, Heath" <Heath.Stewart@xxxxxxxxxx>
Reply-To: "Stewart, Heath" <Heath.Stewart@xxxxxxxxxx>
Date: Thursday, May 10, 2018 at 8:24 PM
To: 4PPOE Reflector <STDS-802-3-4PPOE@xxxxxxxxxxxxxxxxx>
Subject: Re: [802.3_4PPOE] Backfeed presentation + baseline

Lennart,

As always we appreciate your ongoing work. We continue to have some concerns regarding the PSE requirements. As written they do are in effect during 4-pair power operation (and at all other times, without limit.)

We consider a PD which presents a reflected voltage (above Vrefl) during classification to be undesirable. As such we have marked up your presentation to suggest some changes.

Cheers,

-Heath

From: Lennart Yseboodt [mailto:lennart.yseboodt@xxxxxxxxxxx]
Sent: Thursday, May 10, 2018 10:15 AM
To: lennart.yseboodt@xxxxxxxxxxx; Stewart, Heath <Heath.Stewart@xxxxxxxxxx>; STDS-802-3-4PPOE@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_4PPOE] Backfeed presentation + baseline

Hi Heath,

v220 and v230 correspond to both approaches we discussed during the adhoc call.

At the end of the call, I believe the consensus was to split Irev into two values (for >21V and <21V), which corresponds with v230.

Overview:

v220

- PSE requirement Irev is 1.3mA and applies when applying 2-pair (=3-pair) power on the unpowered negative

- PD has to safeguard classification, hence has a 3P-backfeed requirement from 0-21V

v230

- PSE requirement is split into 500uA when highest pairset voltage <21V and 1.3mA when highest pairset voltage >21V

- Because the PSE has a current limit that does not corrupt classifcation in the classification range, the PD is only required to limit 3P-backfeed in the detection range 0-10.1V

The advantage of v230 is that it allows for wider implementation of PD bridges.

Kind regards,

Lennart

On Thu, 2018-05-10 at 16:05 +0000, Stewart, Heath wrote:

Lennart,

Version 220 showed 21V. This version shows 10.1V. Can you comment?

Heath Stewart
Design Center Manager

PoE/PoDL Products

Cell      (805) 895-0499
Office   (805) 560-7658
Websites   analog.com, linear.com

Linear Technology is now part of Analog Devices.  Learn more.

From: Lennart Yseboodt [mailto:lennart.yseboodt@xxxxxxxxxxx]
Sent: Wednesday, May 9, 2018 12:49 PM
To: STDS-802-3-4PPOE@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_4PPOE] Backfeed presentation + baseline

Updated as discussed during the meeting, this is version 230.

On Mon, 2018-05-07 at 20:40 +0000, Lennart Yseboodt wrote:

Hi all,

As presented today. I’m currently updating the text based on the feedback I got.

Kind regards,

Lennart

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

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

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