I have been trying to design an equalizer (2GHz-4GHz) using stripline
structures. For this particular application, only the phase needed
adjusting and so I designed a phase shifter using Schiffman phase
shifters. The final equalizer was to be manufactured using
stripline. We use arlon 2.33 (+/- 0.02) dielectric boards....low
loss (loss tangent...0.002)
A Schiffmann phase shifter is simply a coupled line with one
end shorted together. An estimate of the design was performed
using Hewlett Pacard's touchstone program using ideal coupled line
structures. I then proceeded to make a layout of the structure
using the circuit models that touchstone has.
The problem I have is this: I used Ansoft's ensemble 6.0 moment
method program. The program that we bought was the 2.5D emulator
for infinitely thin circuit lines. I then refined my touchstone
results to get to what I thought would be a more accurate model.
I played with the mesh sizes so I found a stable regime in my
solutions.
When I had my boards manufactured, the results of Ensemble were
horrid--the de-embedded phase was off by circa 40%, while the
touchstone results were accurate to 2-3%.
Does anyone know why?? I can imagine that line thickness might play a
role in the solution, but has anyone else encountered such a result?
I am trying to equalize around 40 degrees of phase error in this band
and am using 8 Schiffman's...
Kindest of regards,
Craig
**** 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 ****