Richard,
In this case the generic answer "it depends" is more true then ever. I
assume your
question is based on the need to somehow capture these variations in
simulations.
As Yuriy pointed out, there are reasonably accurate ways to get decent
agreement
between simulated and measured data, but even in those cases, which can
be done
in an R&D phase but not necessarily in the fast-paced product
development, the
limitation is that we get the average electrical properties over the
length of the
test traces.
In principle, once we know the electrical properties of the glass and
resin, the average
behavior can be approximated based on the glass-resin ratio (neglecting
anisotropy).
Here is where the absolute size of the problem starts to matter: If we
need the
dielectric properties for the end-to-end delay and attenuation, simple
averages will give
good starting point. However, if the scale is smaller, namely if we
need the variation
over smaller distances, the numbers may vary substantially with
location, and the
best we can do is to get an estimate of the statistical distribution.
The complex routing
patterns of real high-density boards can further increase the
uncertainty and we can
conclude that without obtaining detailed and accurate geometry data of
the traces
at sufficiently close intervals either by cross sectioning (which is
destructive) or by
Xray imaging (not practical on large production volumes), all what we
can predict in
worst case is a fairly wide range of performance. Among others this was
illustrated
with the estimated TDR profiles of lossy traces following different
patterns of parameter
variations:
http://www.electrical-integrity.com/Paper_download_files/DC13_Determining_PCB_Trace_Impedance.pdf
Figure 26 illustrates the huge range of possible TDR impedance profiles
for the same
10% impedance variations, just rearranging the sequence of the segment with
different impedances.
Regards,
Istvan Novak
Oracle
On 5/29/2016 9:03 AM, Richard Allred wrote:
Greetings,
I am aware that typical PCB manufacturers usually guarantee some impedance
target (+/- 10 or 15%) for high speed traces and that they may use any
number of manufacturing controls to achieve this. The result is that the
only way to know the geometric dimensions of a given sample is to
cross-section it.
What I am interested in is, what kind of variation can be expected in the
effective dielectric properties due to the PCB manufacturers tweaking of
the glass/resin ratio and the geometrical variation? Is anyone aware of a
published study that reports this?
Kind regards,
Richard Allred
www.SiSoft.com