Re: [8023-POEP] Liaison letter from IEC TC65/SC65C/JWG10 - Power over Ethernet performance in industrial environments
David,
If this is a "standard compliant 802.3 100BASE-TX device", it apparently has a termination network that will lead a PSE to believe it's a PD.
When 802.3af originally chose the PD detection criteria, it was based upon broad industry experience and evaluation of many devices, but its possible that somebody "less experienced in Ethernet design" might incorporate termination networks that could falsely indicate a PD signature.
Perhaps we need to include language in the copper PHY MDI specs that indicates the impedance between the pairs of a non-PD shall not fall within a valid PD range of impedances?
Dan Dove
----
Principal Engineer, LAN PHY Technology
HP ProCurve Networking
-----Original Message-----
From: owner-stds-802-3-poep@xxxxxxxx [mailto:owner-stds-802-3-poep@xxxxxxxx] On Behalf Of David Law
Sent: Thursday, June 18, 2009 3:46 AM
To: STDS-802-3-POEP@xxxxxxxxxxxxxxxxx
Subject: Re: [8023-POEP] Liaison letter from IEC TC65/SC65C/JWG10 - Power over Ethernet performance in industrial environments
All,
I have been asked to forward this information by Bob Lounsbury. It addresses a number of the questions and comments that there have been on the reflector in respect to this liaison letter.
Best regards,
David
----- Forwarded by David Law/GB/3Com on 18/06/2009 11:43 -----
Hello David,
I am sending this email to you as I am currently not part of the IEEE reflector site. Please distribute this email to your committee and any other IEEE 802.3 sub committee you feel is necessary.
I have gone through the IEEE thread in attempts to extract out the questions and then attempted to answer them. First attached is a picture of the PCB that is mentioned in the letter.
The damage was due to a misapplied POE power. Reasons unknown, but as you can see when this happens the result is very destructive (see Enet_PHY.jpg). If this would have been a POE device it would have been a PD, however this device is a standard implementation of 802.3 100BaseTx device.
Answers to comments and questions;
1) The Liaison letter is not asking for application support, the words "application support" does not exist in the letter. There is a question regarding noise isolation back on the cabling caused by POE devices and how this may have an effect on BER of the channel. The paragraph is trying to determine if in the development of POE(+), transient noise caused by the PD's changing load has been considered and if attempts have been made to limit the noise. We want to know if there is high dv/dt noise on the power in a POE channel if this will impact the BER performance of that specific channel.
I am sorry if we assumed that IEEE members had knowledge of the MICE environmental concept. This concept has been and become part of the following standards. TIA 568C.0, TIA 1005, ISO/IEC 24702, ISO/IEC TR 29106, IEC61918 and some of the IEC fiber specifications. The "E" of the MICE details possible noise types and levels in environments ranging from commercial (office) to harsh industrial environments.
I do not know what standard contains 33.4.1 (I assume IEEE 802.3), the device pictured below that was mentioned in the liaison letter is a compliant, standard 802.3 100BaseTX device that does not make use of POE(+) portion of the standard. There for it is not a PSE or PD device and never should have received power from the PSE. Why it may have has had power applied prompted the concerns in IEC SC65C/JWG10.
Regarding the Ideas for response;
I have gone through the email chains and extracted out the questions and comments. I have attempted to address each one below. Also I have included a picture of at least one device that was damaged by miss-application of POE power so you can see the results when a PSE applies power when it should not.
1) It is understood that the primary concern of IEEE 802.3 standard committees (aside of the protocol) is the technical aspects of the network performance as a whole. However in the development of the physical layer
(PMD) portion of the standard, grounding is discussed, shield RJ45 connectors are required and most of all 1500 volt isolation transformers are specified. In addition there is a mention of BER performance in 1V pk to pk noise. Intrinsic to the component specifications defined by the IEEE
802.3 standard and the standards referenced by IEEE 802.3 are safety requirements/noise performance requirements. All of this is irrelevant to the liaison letter as it does not discuss any safety aspects. The primary question is "if electrical noise was considered during the development of the POE(+) standard" specifically during the power probing process.
2) Not sure what this statement means with regards to the questions in the letter, however, if environments E2 and E3 cannot be supported by the IEEE
802.3 standard please let us know ASAP as Ethernet is moving into the industrial areas where E2 and E3 environments are more common at an extremely fast pace. The industrial community needs to know if we are at risk with any part of the network including POE(+). We have asked and received higher balanced cabling specifications and higher common mode attenuation specifications to help mitigate some of the noise issues.
3) Yes standards such as IEC 61000 require a certain level of performance be met in the presence of noise defined in these standards and test fixturing. We are asking if IEEE 802.3 has considered these requirements (or any other) in the design, specification from everything beginning at the protocol to signaling, adaptive filters and component definitions.
With regards to "hardening against these kinds of environments being outside the project scope" again the as I read the letter, it is not asking for any change to the specification or requirements, it is asking if noise was considered, and if it was and POE(+) is not compatible with these noise types and levels, please indicate that, so we can take the appropriate action in our standard. If it was not, someone needs to determine the suitability of POE(+) in high noise environments.
Thanks for your time and I look forward to any additional questions or comments.
Regards
(Embedded image moved to file: pic27067.gif) Rockwell Automation Bob Lounsbury - Principal Engineer
Control and Visualization Business
1 Allen-Bradley Drive, Mayfield Heights, OH. 44124
Tel: +1.440.646.4297
Fax: +1.440.646.3076
Cell: +1.440.610.4485
email: relounsbury@xxxxxxxxxxxxxxx
David_Law@
3com.com
To
06/04/2009 Robert E Lounsbury <relounsbury@xxxxxxxxxxxxxxx>
08:34 AM cc
gwood@xxxxxxx, russof76@xxxxxxxxx,
shariff07724@xxxxxxxxx, Terry Cobb
<trcobb@xxxxxxxxxxxx>, "McCormack, Michael"
<mike_mccormack@xxxxxx>, "Alan Flatman"
<a_flatman@xxxxxxxxxxxxx>
Subject
Re: Fw: [8023-POEP] Liaison letter from IEC
TC65/SC65C/JWG10 - Power over Ethernet performance in
industrial environments
Hi Bob,
Thank you very much for this information.
While the formal consideration of the letter will occur during the IEEE
802.3 plenary in July, to allow maximum time for understanding and consideration of its contents, I have provided the IEEE P802.3at DTE Power via MDI Enhancements Task Force email reflector a preview as I intend to delegate drafting a response to that Task Force. I have also decided to provide a preview to the IEEE P802.3az Energy-efficient Ethernet email reflector. While it doesn't directly relate to that project I know that a number of PHY experts participating in that project so may be individually interested in the letter.
Based on the above I wonder if you could have a look at the email archive for the IEEE P802.3at Task Force - it is a public web site. The thread in response to the preview can be found at the URL [ http://www.ieee802.org/3/poep_study/email/thrd1.html ] and I note there were a few of questions of clarification from some of the participants. I therefore wondered if you would like to either add to information below I supply it to the reflector - or respond to them directly yourself on the reflector - if you wish to join the reflector please see the URL [ http://www.ieee802.org/3/at/reflector.html ].
While I understand that IEC TC65/SC65C/JWG10 can't provide anything formal until they next meet - assistance that you - or anybody else - can provide as an individual is appreciated.
Thanks and best regards,
David
Robert E Lounsbury <relounsbury@xxxxxxxxxxxxxxx> wrote on 04/06/2009
12:28:48:
>
>
>
> Hello David,
>
> As a member of SC65C/JWG10 and co-author of the letter, I wish to
> clarify the intent and concerns that precipitated the letter. This
> concern has been validated by at least one event where a PSE
> incorrectly applied power to a non-powered device. The result was a
> burned up board and melted plastics. The PSE in this system was
> manufactured by a large switch vendor who is active on the IEEE 802.
> 3 standards. The intent of the letter is to solicit a response from
> IEEE 802.3at committee whether or not during the development of the
> POE(+) standard, if noise was considered specifically during the Power
> Discovery process. Additionally has anybody analyzed and tested the
> power discovery process and algorithms in the presence of noise?
>
> In our industrial environments our expected noise levels are generally
> a magnitude greater than that of the commercial environment. Below I
> have provided an extracted partial table from the MICE table printed
> in the ISO/IEC24702 standard. E1 is a typical description of the
> noise levels in a commercial building environment, additionally E3 is
> a typical description of the noise levels in a harsh industrial
> environment. Since the release of this standard the MICE table and
> concept has been adopted by many other standards in IEC and TIA as a
> description of environments ranging from commercial office to harsh
> industrial environments. Many of the test procedures and requirements
> for the electrical noise
> (immunity) tests can be found in a series of IEC 61000 standards.
>
> My personal feeling is that this needs to be addressed before the
> standard is released this year. This is not a cabling issue and cannot
> be solved by the selection of specific types of cabling such as UTP or
> STP/FTP cabling. Each cable type has it's own noise ingress
> mechanisms that are dependant on cable balance, coupling attenuation
> performance and grounding conditions. I hope you consider the request
> from SC65C/JWG10 seriously and respond in a appropriate time so we can
> continue to expand the coverage of POE(+) systems in to industrial
> applications with minimal risk to the end users.
>
> Feel free to distribute this email to your committee We anxiously
> await your response.
>
>
> [image removed]
>
>
>
>
> Regards
> ___________________________________________________
> Rockwell Automation
> Bob Lounsbury - Principal Engineer, Logix/NetLinx
> Automation Control & Information Group (ACIG)
> 1 Allen-Bradley Drive, Mayfield Heights, OH. 44124
> Tel: +1.440.646.4297
> Fax: +1.440.646.3076
> Cel: +1.440.610.4485
> email: relounsbury@xxxxxxxxxxxxxxx
>
>
(See attached file: Enet_PHY.jpg)