[SI-LIST] : width of the return path

About this list Date view Thread view Subject view Author view

From: Eric Bogatin ([email protected])
Date: Sun Jan 16 2000 - 08:16:54 PST


Teddy-

You are asking about the width of the return path for microstrip and
stripline structures. The quick answer is to reach a characteristic
impedance value within 1% of an infinitely wide return path, use a return
path width that extends at least 3xh on either side of the signal trace
width, with h the thickness of the dielectric. For example, an 8 mil wide
microstrip with 4 mil thick dielectric should have a return path width at
least 3 x 4 = 12 mil wide on either side of the trace, for a total width of
12 + 8 + 12 = 32 mils. Making it narrower than 32 will typically increase
the characteristic impedance.

For more details on this, please see the paper I presented at DesignCon99.
It is available for download from my Sept 99 downloads on my web site, if
you check in the publications section.

As a side note, if you are trying to do transmission line design for PCB
stack up for example, I would recommend using a 2D field solver, rather than
a 3D solver. It will run faster, and the Ansoft Maxwell 2D Extractor, part
of the SI2D suite, will be more accurate.

--eric

Eric Bogatin
BOGATIN ENTERPRISES
Training for Signal Integrity and Interconnect Design
26235 W. 110th Terr.
Olathe, KS 66061
v: 913-393-1305
f: 913-393-1306
pager: 888-775-1138
e: [email protected]
web: <http://www.bogatinenterprises.com/>

> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]]On Behalf Of Teddy Chou
> Sent: Friday, January 14, 2000 3:48 AM
> To: '[email protected]'
> Subject: RE: [SI-LIST] : questions about Spicelink/Ansoft tools
>
>
> Dear Nirmal,
>
> Thanks for your response. I've got many suggestions from your
> company. Thanks for all.
> One of these suggestions is that, for the return path of microstrip
> lines, a 10% more in the width of the ground plane than the trace is
> enough. For different structures, we often need to select or build an
> adequate return path model by ourselves. Why can't this tool help us
> to do that according to its field distribution or something else?
>
>
> Best regards,
>
> Teddy Chou(Hou-Yuan Chou)
> Signal Integrity Engineer,
> VIA Technologies, Inc. Taipei, Taiwan, ROC
> TEL : 886-2-22185452 ext : 6046
> mailto:[email protected]
> Very Innovative Architecture
>
>
>
>
> -----Original Message-----
> From: Nirmal Jain [mailto:[email protected]]
> Sent: Tuesday, January 11, 2000 3:33 AM
> To: [email protected]
> Subject: Re:[SI-LIST] : questions about Spicelink/Ansoft tools
>
>
> Dear Teddy,
>
> When you use "Return Path" in Maxwell Quick 3D Extractor, it is actually
> computing loop inductance assuming the return current is flowing through
> the
> assigned return path conductor. So if I have a signal conductor over a
> gnd plane, the original matrix would look like
>
> L11 L12
> L21 L22
>
> where L11 represents the partial inductance of the signal conductor, L22
> is the partial inductance of the gnd plane and L12 (or L21) represents
> the mutual
> inductance between the signal conductor and the gnd plane.
>
> When you use "Return Path" operation, the software is computing the loop
> inductance using the expression
>
> Lloop = L11 + L22 - 2 * L12
>
> The Loop inductance would be higher than the partial inductance of the
> signal conductor if
>
> 1) Ground plane is far off, hence my L12 is very small
> 2) or if my mutual inductance number is negative in the original matrix,
> which means when I do my return path operation, my current in the gnd
> plane is not
> flowing in the opposite direction to the current in the signal
> conductor. In this case the L12 term would added instead of being
> subtracted ( negative of
> negative is +ve). So if your mutual number is -ve, you could either
> re-solve the model by swapping sink/source terminal or swapping the
> sink/source
> terminals in the reduce matrix operation. I prefer the reduce matrix
> step as it is much faster. After this operation, the mutual between the
> gnd plane and
> the signal should be +ve, and it would actually get subtracted in the
> return path operation instead of being added and you would see a much
> lower number!
>
> 3) the solution may not be converged
>
>
> Hope it helps. Let me know if you have any further questions. If you
> still have problems feel free to send me the model.
>
> Best regards
>
> Nirmal
>
>
> > > -----Original Message-----
> > > From: Teddy Chou
> > > Sent: Friday, January 07, 2000 11:00 AM
> > > To: '[email protected]'
> > > Subject: HFSS: questions about Spicelink/Ansoft tools
> > >
> > > Hi,
> > > Recently, I have started to use the Spicelink / Ansoft
> > > simulation tools( it resembles the HFSS/Ansoft, but only the
> > > quasi-TEM waves guided by transmission lines are
> > > considered ). I found some problems when using the
> > > "REDUCE MATRIX" item in Maxwell 3D Extractor. I needed to
> > > get some characteristics of microstrip lines. However, when I
> > > assigned the ground plane to the "return path" or "ground"
> > > ( part of the return path ), the inductance would increase as
> > > the size of ground plane increases. Even when the ratio of
> > > the width of ground plane and the width of a trace is over 100,
> > > it never got a stable value. For avoiding this phenomenon, I
> > > tried to decrease the area of the I_sink and I_source surface
> > > in the ground plane, it still got the resembling results. Who can
> > > tell me how to assign the reasonable size of the ground plane
> > > for microstrip lines.
> > >
> > >
> > > Best Wishes,
> > >
> > > Teddy Chou
> > > Signal Integrity Engineer,
> > > VIA Technologies, Inc. Taipei, Taiwan, ROC
> > > TEL : 886-2-22185452 ext : 6046
> > > mailto:[email protected]
> > > Very Innovative Architecture
> > >
> > >
>
> Nirmal Jain
> Senior Application Engineer
> Signal Integrity Products
>
>
> Ansoft Corp.
> 2560 Oak Rd. #322
> Walnut Creek, CA 94596
> (925) 938-6490 (Tel)
> (925) 215-2477 (Fax)
>
>
> Pittsburgh Office:
>
> Four Station Square
> Pittsburgh, PA 15219
> (412) 261-3200 (Tel)
> (412) 471-9427 (Fax)
>
> **** To unsubscribe from si-list: send e-mail to
> [email protected]. In the BODY of message put: UNSUBSCRIBE
> si-list, for more help, put HELP.
> si-list archives are accessible at http://www.qsl.net/wb6tpu/si-list
> ****
>
> **** To unsubscribe from si-list: send e-mail to
> [email protected]. In the BODY of message put:
> UNSUBSCRIBE si-list, for more help, put HELP.
> si-list archives are accessible at http://www.qsl.net/wb6tpu
> ****
>



**** To unsubscribe from si-list: send e-mail to [email protected]. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
****


About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Thu Apr 20 2000 - 11:34:41 PDT