[SI-LIST] w element table model - methodology
- From: "John Lipsius" <jlipsius@xxxxxxxxxxx>
- To: "SI-LIST" <si-list@xxxxxxxxxxxxx>
- Date: Fri, 3 Jun 2005 19:56:07 -0400
Hello,
I'm looking into the table model for W-element. I'm requesting experiences
from those who've performed some or all of this. Thanks in advance.
The details are: diffpair, 2.5Gbps using SiGe CML and about 8 inches FR4
plus Teradyne VHDM-HDS connector. Obvious xmt and rcv pkg, and PDS, issues
aside...
1. The need for f-dependent model when targeting 2.5Gbps.
Perhaps an obvious "yes"... Assume edge-coupled stripline. At 2.5G
signaling, there is enough loss and phase changes in 8in. so that w/o
pre-emph. there's significant eye collapse, from trusted measurements I've
seen. Therefore, I assume a freq.-dependent model is advisable. Yes?
2. f-dependent Gd.
This is maybe an obvious "yes", too, but since Gd=wC(tand), where C=
cap/unit_length, if I have an assumed accurate C and tand (which is fairly
const. up to 8 or 10GHz) then I can simply calculate Gd for use in a
w-element table model, using a single tand value.
3. Passivity & causality when using w-elem table.
Are there problems encountered with those two conditions due to slightly
inaccurate Df and Dk over frequency? Ie: if I casually construct a w-elem
table w/o enough regard to accuracy of Dk and Df over freq.? What
precautions are advised in using the W table model, both from HSpice and
from physics viewpoints?
4. Model BW.
Wouldn't I then have to filter out any stimulus harmonics above the model BW
since actually the Df saturates and then decreases above some frequency
(like around 10G maybe)? This may be related to #3.
5. Any issues still outstanding with the W elem. field solver?
Regards,
John
------------------- <<< attachment >>> ----------------------------------
[SI-LIST] FW: HSPICE W element field solver model
From: "Eric Bogatin" <eric@xxxxxxxxxxxx>
To: <si-list@xxxxxxxxxxxxx>
Date: Fri, 8 Jun 2001 15:58:50 -0500
Hi folks-
In my limited experience looking at measurements of a few selected FR4
samples, I conclude that the dielectric constant and dissipation
factor are pretty constant from DC up to 10 GHz. For the few samples
we've looked at, the values are 4.0 for dielectric constant and
0.014 - 0.02 (depending on sample) for dissipation factor, constant up
to 10 GHz.
In my presentation to the Rocky Mountain EMC Society, in late May, I
showed that if you do not account for the vias in the transmission
line sample under test, you might interpret the response as a
frequency dependent dielectric constant and dissipation factor.
A number of papers and copies of slide presentations we've given on
this topic are posted in the application notes section of our web site
(www.gigatest.com). We show the test structures, the measured data and
the comparison to simulated response assuming a constant dielectric
constant and dissipation factor. We also show what would have happened
if you did not include the vias. If you want more info on this topic,
we are doing a class on related techniques, in late July.
If anyone has measured results from FR4 that shows frequency dependent
effects from 100 MHz to 10 GHz, I'd be interested in seeing the data.
Of course, always keep in mind that not all FR4's are the same. This
is why every end user of an FR4 laminate for high speed applications
where the dissipation factor is important, should make sure they have
measured their laminate at least once to verify it will meet their
performance spec.
Even though the dissipation factor is constant with frequency, the
attenuation due to dielectric loss is frequency dependent and
increases linearly with frequency. Higher frequency components in the
signal will be attenuated more than low frequency components. This
will give rise to rise time degradation and collapse of the eye
diagram.
comments are always welcome.
--eric
**************************************
Eric Bogatin
CTO, Giga Test Labs
v: 913-393-1305
f: 913-393-1306
e: eric@xxxxxxxxxxxx
corporate office:
408-524-2700
134 S. Wolfe Rd Sunnyvale, CA 94086
web: www.gigatest.com <http://www.gigatest.com>
**************************************
--
------------------------------------------------------------------
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 FAQ wiki page is located at:
http://si-list.org/wiki/wiki.pl?Si-List_FAQ
List technical documents are available at:
http://www.si-list.org
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:
- » [SI-LIST] w element table model - methodology
