[SI-LIST] Re: Ethernet emissions.
- From: Todd Hubing <HUBING@xxxxxxxxxxx>
- To: "<si-list@xxxxxxxxxxxxx>" <si-list@xxxxxxxxxxxxx>
- Date: Tue, 9 Apr 2013 21:00:54 +0000
Curt raises some good points. I don't think anyone was advocating that the
ground plane needed to extend beneath the magnetics. There are sometimes valid
reasons (all low-frequency concerns) for keeping the plane away from the
secondary. Also, as you mention, it is important to keep high-frequency signals
and the VCC plane away from the secondary side of the connector magnetics. I
believe most designers do a good job of following these guidelines.
Even if Bryan's ground plane did extend completely under the connector, it was
not the source of his radiated emissions problem. The ground plane doesn't
bring the high-frequency noise to the cable. The radiated emissions are the
result of the cable being driven at a different high-frequency potential
relative to the ground plane.
I'd like to comment on Joel's observation about the vendor application notes
though. Many products with a metal chassis require low-frequency isolation
between their circuit ground plane and their chassis ground. Low-frequency
transient currents need to be directed to the chassis ground and kept off the
circuit ground. Anyone that is designing a product in a metal chassis with
high-speed Ethernet on a two-layer board, may want to consider putting their
chassis ground and digital ground on the same board layer and stitching them
together with a large number of capacitors. This was apparently the
state-of-the-art when many of these app notes were written. :)
However if your board has 4 or more layers, and if you need to isolate your
chassis ground and digital ground; then you should put the chassis ground on a
different layer than the digital ground plane and they should overlap in the
region near the connectors and be stitched together with capacitors. Most
boards I've seen fall into this category.
At high frequencies, we can only have one ground. Any two pieces of metal of
significant electrical size that are not at the same potential can present a
radiated emissions problem. Bryan's product only had one ground, so the main
concern was not to split it and not to add another ground at a different
potential.
Todd
Todd Hubing
LearnEMC.com<http://www.learnemc.com/>
-----Original Message-----
From: Curt McNamara [mailto:curt.mcnamara@xxxxxxxxxxx]
Sent: Tuesday, April 09, 2013 1:21 PM
To: joel@xxxxxxxxxx; bryan@xxxxxxxxxxxxxx
Cc: leeritchey@xxxxxxxxxxxxx; Todd Hubing; <si-list@xxxxxxxxxxxxx>
Subject: RE: [SI-LIST] Re: Ethernet emissions.
One way to view Ethernet transformers is that they separate the grounds between
the systems at each end of the Ethernet cable.
There are often common mode choke components built into the transformer (or
RJ-45) to keep common mode signals from propagating from the product internals
to the output side of the transformer.
Keeping all signals away from the output side of the transformer will reduce
the possibility of noise coupling to the output pair. This includes any grounds
associated with the product internals. So this is one reason folks eliminate
the ground under the output side of the transformer.
The other reason ground is eliminated is due to safety considerations. The
transformer gives a few thousand volts of isolation. Having ground under the
output traces will typically violate creepage and clearance requirements set in
the safety standard.
So far I have just been talking about eliminating ground and all signal traces
under the output side of the Ethernet circuit. You asked about having earth or
chassis ground under the output side. The answer to that is "it depends".
Curt
Curt McNamara, M. Eng. P.E. // principal electrical engineer | electrical
engineering
Logic PD
T // 612.436-5178
NOTICE: Important disclaimers and limitations apply to this email.
Please see this web page for our disclaimers and limitations:
http://logicpd.com/email-disclaimer/
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx<mailto:si-list-bounce@xxxxxxxxxxxxx>
[mailto:si-list-bounce@xxxxxxxxxxxxx] On Behalf Of Joel Brown
Sent: Tuesday, April 09, 2013 11:56 AM
To: bryan@xxxxxxxxxxxxxx<mailto:bryan@xxxxxxxxxxxxxx>
Cc: leeritchey@xxxxxxxxxxxxx<mailto:leeritchey@xxxxxxxxxxxxx>;
<HUBING@xxxxxxxxxxx<mailto:HUBING@xxxxxxxxxxx>>;
<si-list@xxxxxxxxxxxxx<mailto:si-list@xxxxxxxxxxxxx>>
Subject: [SI-LIST] Re: Ethernet emissions.
Almost every company that makes Ethernet controllers and Phys (Including
Intel) recommends to put a cut in the ground plane around the RJ45 connector
with built in magnetics and to ground the connector to chassis ground and then
strap some capacitors across the plane split. Is it better to just use a solid
plane all the way to the connector or does the plane split sometimes solve a
problem? In general I avoid plane splits. One case where we did use a ground
plane split was with an SDI phy and the manufacturer (Genum) said it wouldn't
work without it.
On Sun, Apr 7, 2013 at 7:15 PM, Bryan Ackerly
<bryan@xxxxxxxxxxxxxx<mailto:bryan@xxxxxxxxxxxxxx>> wrote:
> Lee, I try to never cut up ground planes. This is the one "golden
> rule" i took away from one your own training courses I attended many years
> ago....
>
> Regards,
> Bryan Ackerly
>
>
> On 07/04/2013, at 4:19 AM, "Lee "
> <leeritchey@xxxxxxxxxxxxx<mailto:leeritchey@xxxxxxxxxxxxx>> wrote:
>
> > Todd,
> >
> > Thanks for reiterating the need to avoid cutting up ground planes.
> > I am puzzled as to why anyone would recommend such a tactic. But,
> > then, there
> are
> > a host of such "off the cuff" things that get circulated year after
> > year without any proof they are valid. Even more puzzling is why
> > engineers accept such claims without supporting evidence that they are
> > valid.
> >
> > They keep me very busy most of the time fixing problems that they cause!
> >
> > -----Original Message-----
> > From: Todd Hubing
> > Sent: Saturday, April 06, 2013 10:54 AM
> > To: <si-list@xxxxxxxxxxxxx<mailto:si-list@xxxxxxxxxxxxx>>
> > Subject: [SI-LIST] Re: Ethernet emissions.
> >
> > Bryan,
> >
> >
> > Please don't cut up your ground. I wouldn't want anyone on this list
> > to
> get
> > the idea that isolating grounds is going to solve a radiated
> > emissions problem above 100 MHz. From the symptoms and the
> > measurements you
> describe,
> > it appears that you are driving the cable relative to your circuit
> > board ground with a CM voltage produced directly by the PHY and
> > insufficiently attenuated by the connector magnetics. If you've
> > maintained balance in
> your
> > layout between the PHY and the connector, you may need to consider
> > using
> a
> > different connector with better CM rejection.
> >
> >
> >
> > At these frequencies, nothing you do with your ground is going to
> > improve upon the solid plane that you currently have. Trying to
> > establish
> separate
> > grounds that are at different potentials is likely to make things worse.
> >
> >
> >
> > Todd
> >
> > -----
> >
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx<mailto:si-list-request@xxxxxxxxxxxxx> with
> 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> //www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx<mailto:si-list-request@xxxxxxxxxxxxx> with
> 'help' in the Subject field
>
>
> List forum is accessible at:
> http://tech.groups.yahoo.com/group/si-list
>
> List archives are viewable at:
> //www.freelists.org/archives/si-list
>
> Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
>
>
>
--
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx<mailto:si-list-request@xxxxxxxxxxxxx> with
'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx<mailto:si-list-request@xxxxxxxxxxxxx> with 'help'
in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
