[SI-LIST] Re: Traces don't cause EMI - really?

  • From: "Loyer, Jeff" <jeff.loyer@xxxxxxxxx>
  • To: <MikonCons@xxxxxxx>
  • Date: Thu, 16 Oct 2003 13:06:28 -0700

Thanks for sharing!

This ("a 100-Ohm trace radiated >6 dB more than a 50-Ohm trace") seems =
to agree with Dockey's paper - the farther away you are from the =
reference plane (and, consequently, the higher the impedance), the more =
EMI you radiate.

The question I'm confronted with at this point is: do "well-behaved" =
single-ended signals generate appreciable EMI?  Dockey's paper outlined =
the effect, but with geometries (90mil wide trace over a 62mil =
dielectric) that aren't likely to be used.  He then demonstrated the =
reduction in EMI when the trace was brought closer to the reference =
plane.  Given current geometries, is it likely for a bus of properly =
designed single-ended traces to cause failing EMI?  My =
experience/impression is no. =20

If that's incorrect, please let me know.

On the other hand, as I understand it, differential traces are much more =
"forgiving", EMI-wise, of non-idealities.  And these we often induce, =
knowingly or not.  So, tightly-coupled differential traces are =
desireable, EMI-wise, given our imperfect world and the issues we must =
deal with.

That's not to say there aren't other reasons, besides EMI (increased =
loss, and some yet to be discovered/published), why we may NOT want to =
route differential traces closely-coupled.

Unless I'm incorrect about current single-ended traces and their =
tendency (or lack of) to generate EMI, I'd stick with my original =
assertion, that closely coupling differential signals has benefits, but =
is not absolutely necessary for their entire length (but compensation =
for impedance change due to reducing or increasing coupling might be =
necessary).


Thanks for your time/energy.  Actually, I've been contracted by the =
fishees to distract you from pursuing them :-)

Jeff Loyer

-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx
[mailto:si-list-bounce@xxxxxxxxxxxxx]On Behalf Of MikonCons@xxxxxxx
Sent: Thursday, October 16, 2003 11:33 AM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Traces don't cause EMI - really?


In a message dated 10/16/2003 10:14:54 AM Pacific Standard Time,=20
jeff.loyer@xxxxxxxxx writes:
Just to be clear - are you saying that if I connected 2 properly =
designed=20
chips (driver and receiver) together with many properly designed =
single-ended=20
transmission lines, they would likely fail FCC standards?
Yes, but only with qualifications. Picture a single-ended, 50-Ohm =
microstrip,=20
not multiple closely spaced traces. If excited with a 1 ns risetime =
signal at=20
33 MHz clock frequency from a 50-Ohm generator, and terminated with a =
50-Ohm=20
chip resistor, expect failure. For example, the EMCAD1 software =
application=20
(circa 1992) is based on radiated emissions equations and predicts ~39 =
dBuV at=20
33 MHz at 30 meters antenna distance. The first few harmonics are also=20
predicted to be out of limits. Note that this software uses worst-case =
predictions.=20
The same software also indicates an increase in the single-trace =
emission level=20
by SQRT(N), where N =3D the number of traces; however, the assumption is =
that=20
all traces are carrying the same excitation (which is unrealistic).=20
Now, if you place grounded planar areas around the trace, or have a =
densely=20
routed board, the additional copper provides a substantial reduction in =
the=20
measured emissions from that single trace because many field lines that =
would=20
otherwise leave the board terminate on the added copper (even though =
they may be=20
additional 50-Ohm traces). And, in a normal board design, many of the =
signals=20
on these surface traces will generate fields that cancel each other.=20
Therefore, the bare (i.e., unshielded, and unenclosed) PCB MAY or MAY =
NOT fail an FCC=20
or CISPR radiated emissions test.=20

The IBM paper I mentioned recorded up to 20 dB variation in radiated=20
emissions from a single microstrip on a bare board as the trace was =
moved in-board=20
from the PCB edge. This measured reduction also illustrates the =
field-capturing=20
effect of surrounding copper, reference planes, or other conductors.

The PCB I created contained significant open surface space around the =
traces=20
(including the guarded trace structures) which tended to increase the =
radiated=20
emissions relative to a real board design, but this was on purpose to=20
illustrate the effect.

As an added note, I found that the radiated emissions from a single =
trace=20
increased more than linearly with the impedance of the line; i.e., a =
100-Ohm=20
trace radiated >6 dB more than a 50-Ohm trace. This result makes sense =
as a better=20
match to the 120*pi =3D 377 Ohms of free space is achieved with the =
100-Ohm=20
line.

Mike

Michael L. Conn
Owner/Principal Consultant
Mikon Consulting

*** Serving Your Needs with Technical Excellence ***


------------------------------------------------------------------
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 archives are viewable at:    =20
                //www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages=20
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu
 =20

------------------------------------------------------------------
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 archives are viewable at:     
                //www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages 
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu
  

Other related posts: