[SI-LIST] Re: Guard traces for differetial pairs
- From: =?big5?q?sghsu55?= <sghsu55@xxxxxxxxxxxx>
- To: scott@xxxxxxxxxxxxx, leeritchey@xxxxxxxxxxxxx
- Date: Mon, 29 Sep 2003 09:14:50 +0800 (CST)
Scott,
I don't totally agree with you. I have watched
this problems for several years. My conclusion is
"guard trace can certainly improve coupling
coefficient, however, mostly, guard trace may cause
serious crosstalk".
---------------------------------- (victim)
|-------------|----------|---------| (guard trace with
stitching vias)
---------------------------------- (aggressor)
Consider the above case in multi-layer PCB, the
guard trace with stitching vias can certainly reduce
the coupling coefficient between victim and aggressor
in frequency domain. It is well known that the
guard trace behaves a quarter-wavelength band pass
filter. However, in time domain, the situation is
not so easy. The coupling coefficient between guard
trace and aggressor is siginicantly higher than
the one between victim and aggressor. The impedance
mismatch of stitching vias will result in more
serious corsstalk from guard trace to victim. Totally,
the case with graud trace induced more nosie
in victim. The guard trace will help in reducing total
noise in 2-layer PCB. In my opinion, the reason is
the ratio of the distance between signal and guard
trace and thickness between signal and reference
plan.
Sogo Hsu, Ph. D.
Simulation Center, Foxconn
--- Scott McMorrow <scott@xxxxxxxxxxxxx> ���T���G>
Lee is right and wrong.
> Some traces guard or shield neighboring traces ...
> some don't.
>
> There are no absolutes about what we might call a
> guard or shield trace.
> Lee is correct in advising due caution in their use.
> Many designs have
> failed because of mis-applied guard traces, causing
> untold crosstalk and
> emi problems. If a design can be made to work with
> just spacing, then
> just do-it. Don't mess with guard traces unless you
> are prepared to
> understand a bit of microwave resonance theory.
>
> However, in skilled hands, a shield trace can
> provide additional
> crosstalk isolation without adding space to a
> design. It is possible to
> achieve an additional 3 to 10 dB of isolation when a
> correctly designed
> shield is used. Mike Conn and I have discussesd
> this in the past on the
> reflector. The key is to use enough ground vias to
> "nail the shield
> down." We've previously discussed the control of
> resonance (Lee's band
> pass filter effect.) by the judicious use of these
> vias. We've even
> discussed the necessity for irregular spacing of the
> "stitch" vias, in
> order to inhibit the formation of periodic resonance
> patterns.
>
> Like any tool, guard traces or shields can be used
> correctly or they can
> be abused. If you don't understand how they work
> and the potential
> pitfalls, just don't even think about using them.
> But if you need to
> achieve extremely high isolation and significant
> reductions in EMI in
> your design, then a guard trace might be a possible
> solution.
>
> Regards,
>
> scott
>
> --
> Scott McMorrow
> Electromagnetic Field Wrangler
> Teraspeed Consulting Group LLC
> 2926 SE Yamhill St.
> Portland, OR 97214
> (503) 239-5536
> http://www.teraspeed.com
>
>
>
>
> Lee Ritchey wrote:
>
> >Tonglong,
> >
> >Guard traces don't "guard". If you have the bad
> luck to get just the right
> >length, they become a band pass filter. They
> should not be used. Handle
> >cross talk by setting proper spacings.
> >
> >Lee
-----------------------------------------------------------------
�C�ѳ� Yahoo!�_��
�����C��B�����𮧡B�R�A���ūסA�ɦb�H�ȩ���
http://tw.promo.yahoo.com/mail_premium/stationery.html
------------------------------------------------------------------
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:
