Hi Al,
Thank you for the feedback and fair criticism :-)
Though, not so good launches and restricted bandwidth of the GMS-parameters
are not relevant to the subject of anisotropy. The layered structure of
dielectrics cause difference in even and odd mode phase delays. The result
can be observed as the FEXT. FEXT is compared for four different materials
up to 50 GHz on slide 19 at
https://www.simberian.com/Presentations/14-WE2Presentation_LessonsLearnedHow
ToMake_v3.pdf That is the way to identify inhomogeneity of dielectric that
can be modeled as anisotropy. Traces in homogeneous isotropic dielectric
will have very small FEXT.
Considering the launches, they are not masking anything. For the material
characterization with GMS-parameters, the launches have to be substantially
identical and not very sensitive to the manufacturing variations. It means
no small gaps between pads and antipads and no resonances (stubs).
Non-reflectiveness is not a necessary condition. What reduced the bandwidth
in the paper was inconsistent backdrilling and variations of geometry of
tiny traces - the bandwidth of GMS-parameters was reduced to 15-20 GHz, but
the extracted model is usable up to 50 GHz - that was essentially
demonstrated in the correlation section. In fact, looking back, I think that
was the main result of that paper - models extracted with the reduced
bandwidth measurement can be usable to predict interconnect behavior up to
much higher frequency!
It is not point by point extraction as suggested in some standards.
Best regards,
Yuriy
Yuriy Shlepnev, Ph.D.
President, Simberian Inc.
www.simberian.com
Simbeor - Accurate, Productive and Cost-Effective Electromagnetic Signal
Integrity Software to Design Predictable Interconnects!
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On
Behalf Of al@xxxxxxxxxxxxxxxxx
Sent: Friday, December 13, 2019 11:42 AM
To: shlepnev@xxxxxxxxxxxxx; jason.j.ellison@xxxxxxxx; 'si-list'
Subject: [SI-LIST] Re: PCB Isotropy
Yuriy,
The Rambus paper you co-authored 2014-01 demonstrates a platform with at
most 15GHz of useful bandwidth, and without being harsh or trying to be
denigrating is very poor overall signal integrity based on IEEE P370 TG1.
Your figure 8 and 10 in the paper clearly outlines that along with having
10ohm impedance bump launches (again actually very poor SI). As you
suggested in your post, the signal integrity issues of poor launch (excess
capacitance launch pad for example) can mask the anisotropic effect. How
exactly did you achieve high confidence 50GHz unimpeachable material
identification as summarized in your conclusion?
My thinking is 50GHz material identification should have significant SI
margin as reflected in 11psec TDR and return loss at Fmax, with very low
delay noise. Thanks in advance for helping me to understand.
- Al
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx <si-list-bounce@xxxxxxxxxxxxx> On Behalf
Of Yuriy Shlepnev
Sent: Friday, December 13, 2019 10:58 AM
To: jason.j.ellison@xxxxxxxx; 'si-list' <si-list@xxxxxxxxxxxxx>
Subject: [SI-LIST] Re: PCB Isotropy
Hi Jason,
The anisotropy of layered PCB dielectrics shows up on the measurements for
differential strip lines as difference in phase or group delay of the odd
and even modes. So, you can use any data or publications that allow or
compare odd and even modes to see the effect of the layering. The outcome
will depend on a particular material - for instance in #2014_01 at
https://www.simberian.com/AppNotes.php some materials showed anisotropy and
some (Nelco) did not. If difference between odd and even modes is
negligible, no reason to simulate it. Otherwise, either layered dielectric
model (physical, preferred) or averaged anisotropic dielectric model
(permittivity tensor) can be used, to take it into account. See more in
2016_01 (Material world tutorial) at
https://www.simberian.com/TechnicalPresentations.php
Considering the vias, periodic capacitive loading of the via barrels can
slow down the wave quite a bit (barrel to planes). If you measure delay
through it and compute "effective" dielectric constant, it may come out
higher than the Dk of the dielectric it is filled with. Sharp
discontinuities on the surface of vias and on the edges of the planes also
increase the capacitance and the effective dielectric constant (slows down
wave). Some of those effects can be simulated, but some may require an
anisotropic dielectric models.
Best regards,
Yuriy
Yuriy Shlepnev, Ph.D.
President, Simberian Inc.
www.simberian.com
Simbeor - Accurate, Productive and Cost-Effective Electromagnetic Signal
Integrity Software to Design Predictable Interconnects!
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On
Behalf Of Jason Ellison
Sent: Friday, December 13, 2019 10:19 AM
To: si-list
Subject: [SI-LIST] PCB Isotropy
Hello SI Community!
Does anyone know of a PCB laminate permittivity isotropy study?
Specifically, I'd like to know if the permittivity is different when the
signal is propagating through the board vias (Z-direction) vs along traces
within the board (X-Y direction).
Cheers,
Jason
Employed by and affiliated with Amphenol AICC (https://www.amphenol-icc.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 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
------------------------------------------------------------------
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
------------------------------------------------------------------
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
