Cascaded RL Calculations? - RF Cafe Forums
Post subject: Cascaded RL Calculations?
Wed Aug 11, 2004 5:13 pm
Wed Aug 11, 2004 5:11 pm
I’m trying to characterize
the return loss of a cascaded path. The path elements are very simple…
cables, attenuators and switches. I have SWR and insertion loss
data for each element, but am unsure how to calculate the return
loss or SWR of the entire path. Do you know of a tool or method
that may be useful?
Thank-you very much, any help would be
Unread postPosted: Thu Aug 12, 2004 12:49 pm
Do you have (complex) s-parameter data for the components?
If so, the s-parameters of the cascade can be computed by converting
the s-parameters of each of the elements to their ABCD matrices,
then computing the ABCD matrix of the system by multiplying the
individual ABCD matrices, then converting back to s-parameters.
A math package can be very helpful for this type of calculation.
There is also free MW software that will compute this easily, though
I've never used it.
Unread postPosted: Thu Aug 12, 2004 2:06 pm
If your elements are cascaded then you should multiply the SWR
of each element and you will get the cascaded SWR. From this it
is easy to obtain the relflection coefficient and RL, as follows:
SWR=1+rho/1-rho (rho- Reflection Coefficient)
Hope this helps,
Unread postPosted: Thu Aug 12, 2004
Thanks for your replies!
I have already calculated
the cascaded VSWR and RL using the RSS method (very similar result
to simply multiplying each components VSWR as Itay mentioned), but
the problem is that the result is not even close to the measured
value of return loss. The calculated value is MUCH worse. I am using
the specified max VSWR for each device which would of course cause
the resulting calculated VSWR to be higher than would typically
be measured. My paths are also up to 12 devices long which is compounding
the errors due to the assumptions made in the RSS method.
I think part of the problem, if I understand this method correctly,
is that it assumes no device losses and only looks at the 1st order
reflections. In this model the reflected waves do not experience
any attenuation because device losses have not been considered.
This results in much larger SWRs and much smaller return loss values
than would actually occur.
Of course the RSS method calculates
the absolute max possible VSWR only. I will use the information
on http://www.rfcafe.com/references/electr ... smatch.htm to calculate
the minimum...hopefully my measurements are somewhere between the
two ;-) Otherwise I suppose I’ll need to get a hold of the complex
s-parameters for each device from their manufacturers and try Vuk
Borich's suggested method.
Unread postPosted: Thu Aug 12, 2004 4:09 pm
Why make your life complicated? Use a VNA to measure exactly
the Return Loss/VSWR at the wanted frequency range for each element
in your cascade (don't meaure it outside the frequency band you
are using). Then apply the method you named, which is similar to
the one I noted before. If you start to convert ABCD matrices to
S-parameters you will end the calculation next year ... long and
winded calculations are prone to errors. :!:
Also check that
all your elements are in good shape, a cable that is faulty might
give you a short circuit or an open circuit and then of course you
will end with a wrong measurements. Check that the insertion losses
of the cables are in the range and that the cables are matched to
50-ohm (Use the Smith Chart for that).
Thu Aug 12, 2004 6:54 pm
Wed Aug 11, 2004 5:11 pm
I'm actually trying to
determine the limits that a measured result should be between. I'm
more or less satified with my current results. The RSS method is
providing a maximum SWR (minimum RL) and the method of deviding
the larger VSWR by the smaller VSWR thought the cascade (used here
http://www.rfcafe.com/references/electr ... smatch.htm) is providing
the min SWR (max RL).
Anyone have any details of the theory
behind this min VSWR calculation?