If we ignore the specifics of the failed device as Bob suggests, I
think it is easy to address the liason's concerns.
802.3 is mostly a
feature specification, not a product specification - 802.3at makes it clear that the requirements are functional and
are not substitutes for EMI and safety standards. A
product specification might include 802.3 clauses (by reference) and many
additional feature specifications, as well as the appropriate
environmental requirements (MICE catagories), and so forth. As Geoff pointed out, there is little environmental
information in 802.3 - leading me to believe that the implementer (product
designer) must overlay the appropriate safety and environmental specifications
for that particular product.
Maybe the right question from the IEC should have been "To
your knowledge, is there anything in 802.3 clause 33, or 802.3 that
would preclude ethernet from operating in a noisy environment?" This
is similar to the intent that 802.3at does not preclude an implementation
from meeting safety standards, however following only 802.3 does not imply
an implementation will meet the safety
standards.
I am
sure there are over 50M PoE ports operating in the field, some of which are
in industrial environments and functional. Since nobody on the committee
itself raised the flag, and since this is the first external inquiry in 4 years,
we can assume there is not a raging misdetection problem in the industrial
environment out there. I take this as a proof that it is possible
to implement 802.3 clause 33 in a manner that operates in the
industrial environment. If someone
in the task force (a PSE vendor) has applicable experience this would
help. We can state this in a
liason - 802.3 clause 33 does not preclude implementations that operate in noisy
environments.
Evidence was
provided that problems can and have occured. We suffer from a lack of facts
here. Despite this, the consensus amongst the task force is that
perhaps the implementation that damaged devices was suspect, there was a
defect, or the hardware was used outside its environmental design
intent. If the IEC could show us how to break the
standard, I am sure we would jump at the opportunity to fix
802.3at.
The industrial community's concerns might be
addressed if we added a note or caution to the PSE detection section
that reminded the PSE designer to consider the appropriate electrical
enviroment standards when implementing detection (and maybe MPS per Peter). We have notes elsewhere
in 802.3at.
Shadi's
email shows that existing standards recognize the requirement that
devices be used in the manner intended by the maker, or alternatively this
means that measures be taken to effectively ameliorate the environment
to one the equipment was designed for.
So I am
hoping it is possible to answer the liason and address their root concerns
in a manner that is consistent with 802.3 practice by providing a little
guidance to the PSE designer in this last draft of 802.3at. Explaining
that 802.3 is a feature specification and not a product specification might
clarify our position.
Regards,
Martin
Martin
Patoka Systems Engineering Manager Texas Instruments O:
214-567-5487 mpatoka@xxxxxx
Shadi-
It is no secret to the Industrial Ethernet folks that
consideration of the extreme environmental conditions found on a manufacturing
floor have never been a consideration within IEEE 802.3.
Conversely, in
the earliest days of networking the indeterminate/asynchronous timing
characteristics of Ethernet were considered to by the manufacturing folk to be
completely unsuitable for use in a manufacturing environment. They then went off
on their own, first to 802.4 and when that did not pan out off to other arenas.
We have hardly heard from them since. In the meantime, we continued on our merry
way in pursuit of our original goal, that of high bandwidth in an environment
that was driven by that found in the office and light industrial market. (They
are by no means unique, on-board automotive networking is another fine
example.)
While there is very little text that addresses this in our
standard, our tradition harkens back to the the text in the introduction section
(1) of the original DEC/Intel/Xerox Ethernet Specification which says:
"Use in situations demanding resistance to hostile environments,
real-time response guarantees, and so on, while not specifically excluded, do
not constitute the primary environment for which the Ethernet is
designed."
My personal belief is that text is still operative
(despite the improvement in real time response made available through full
duplex and the even more stringent desires being put forth by 802.1 AV Bridging)
in the dominant philosophy of operation of 802.3.
Best
regards,
Geoff
Geoffrey O. Thompson GraCaSI 158 Paseo
Ct Mountain View, CA 94043
<thompson@xxxxxxxx>
On 7/8/09
6:13 AM, Shadi AbuGhazaleh wrote:
I am tending to think
along the same lines as Martin.
The industrial
environment has some pretty harsh conditions. Accordingly, the cabling and
vendor standards dealing with industrial facilities require that when
non-industrial hardened devices/applications are used that they should be
protected or isolated (through the use of higher performing components OR
appropriate shielding, conduit, enclosures, etc.. Ref: TIA-1005) to make sure
that the systems (cabling and active hardware) do not experience the
levels of high noise.
It seems
that Bob is essentially asking 802.3 to create new requirements on noise
immunity and address the needs of the industrial environment
directly. Also seems to me that this request is
inclusive of ethernet, not only PoE, as the question of BER has been asked and
I would offer that Ethernet operation in noisy environments (up to
10V/m) is more susceptable than PoE. Is this something that 802.3
is looking to do? If not, then Hugh's response is appropriate, I would
just add some pointers to appropriate environmental/cabling standards
such as TIA-1005 in combination with ANSI/TIA-568-C for Bob to use in
configuring the industrial systems. Regards, Shadi
AbuGhazaleh Development & Technology Manager, Hubbell Inc. sabughaz@xxxxxxxxxxx
Hi Bob,
It sounds like you have additional information about the
incident not included in the liason letter.
What can you tell us about the electrical environment of the
PSE, channel, and (non-PD) end-point device discussed in the liason?
What can you tell us about the physical channel
implementation?
Can you provide the specific requirements that you allude to in
your posting?
Thanks and Regards,
Martin
Martin
Patoka Systems Engineering Manager Texas Instruments O:
214-567-5487 mpatoka@xxxxxx
----- Message from Robert E Lounsbury <relounsbury@xxxxxxxxxxxxxxx>
on Wed, 8 Jul 2009 06:38:06 -0400 -----
Hello Hugh,
Dave,
Please read the liaison letter. Please disregard the damaged
device. The damaged device only served to start the discussion in IEC.
The letter is asking two very basic questions that might precipitate
some considerations in the specification. First Dave mentions noise
coming from the device loads back onto the channel. This was one of the
questions in the letter. Dave alludes to violating the specification.
Please point to where in the specification where noise produced and
coupled back to the PoE channel is defined, limited or otherwise discussed.
Second the letter asks if the PoE TG has considered analyzed what might
happen if there is Radiated, conducted or other environmental noise in the
area where the PoE system is installed might cause problems during the probing
process. This has nothing to do with any design flaws of a product but merely
the environment. If you consider Industrial noise levels as defined by
IEC out of the scope for 802.3at, then please respond indicating so. Then we
might be in a position to disallow PoE in industrial. Thanks for
your time Regards
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,
I can see that the
liaison from IEC has caused some interesting discussion amongst PoE
experts (and myself), however I think that we are being asked to comment
on what may well be design flaws in specific products and I do not think
that is a healthy discussion within 802.3.
I suggest that we should
send a reply along the following
lines:
==========================================
The problems
described in your liaison spurred some vigorous discussion amongst our
task force members regarding possible causes for the damage that you
describe. However the members are not aware of any similar reports that
might indicate a systemic problem with 802.3 compliant equipment. It is
the opinion of the members of IEEE P802.3at Task Force that the standard
allows product manufacturers to build reliable and interoperable equipment
that will meet the requirements for supplying power over Ethernet in many
environments. However, the standard does not define how a manufacturer
must build the product to ensure reliability or how an installer should
ensure that the media is suitable for correct operation within the
standard. We suggest that you should work with the equipment manufacturers
involved to determine whether the failure is the result of a systemic
problem with the standard and whether a specific amendment may be
required.
With respect to the bit error rate performance of 802.3 links
when power is being supplied over the same link, the members of IEEE
P802.3at Task Force believe that a compliant system supplying power over
an 802.3 link will not perturb the channel sufficiently to degrade the
performance of the underlying link. However, it is the responsibility of
the product manufacturer to ensure that noise introduced by the load does
not couple to the link and violate the power over Ethernet specifications
or the channel specifications required for the link. Similarly it is the
responsibility of the system installer that the channel characteristics
are met in the presence of environmental
noise.
===========================================
Hugh.
David
Law wrote:
>All, > >The IEEE 802.3 Working Group has
received a liaison letter from IEC >TC65/SC65C/JWG10, Industrial
process measurement, control and >automation/Industrial networks with
respect to Power over Ethernet >performance in industrial environments.
> >I just wanted to inform you that I intend to delegate the
generation of a >draft response to the IEEE P802.3at DTE Power
Enhancements Task Force >during the plenary week in July. The draft
response will be consider and >then voted upon at the closing IEEE
802.3 Working Group plenary as part of >the IEEE P802.3at closing
report. You therefore may wish to review the >letter prior to the
meeting, the letter can be accessed at the URL [ >http://www.ieee802.org/3/minutes/jul09/0709_IEC_SC65C_JWG10_to_802_3.pdf
>]. > >Best regards, > David Law >
IEEE 802.3 Working Group Chair > >
>
*******************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom
they are addressed. If you have received this email in error please
notify the system manager. This footnote also confirms that this
email message has been swept for the presence of computer viruses.
www.Hubbell.com - Hubbell Incorporated**
Hotmail® has ever-growing storage! Don’t worry about storage limits. Check it out.
|