Original question:
We have recently received an extended range filter by Varian.
It aims to replace the familiar cylindrical bandpass filters that have a narower frequency range.
It is a box-shapped filter labelled WR19-216-BB-N.
Per rf filter terminology that would mean a wide range(?), box filter centered at 19 MHz and
extending up to 127 Mhz. Not sure what the final N means, perhaps notch (lock frequency?).
My question: how could I evaluate it, especially vs regular bandpass
filters? Probe tuning, sensitivity, even running inverse detection
experiments come to mind. I was wondering if any of you would share your
experience and thoughts. I will post a summary.
AMMRL suggestions:
Try attaching it to the tune port with a 50 ohm load. Then run qtune.
I have never tried this, but it should give a response curve.
A network analyzer can give quantitative evaluation, but you can also
use a sweeper or your spectrometer's "Q tune" or "wobble tune" if you
substitute the filter for the probe and put an authentic 50 ohm load on
the other side of the filter. If you have no sweep tune capability, but
you can measure forward and reflected voltage, you can map out the
"spectrum" of the filter point by point, which is tedious.
What you need to measure the performance of the two filters is a
spectrum analyzer with a tracking generator. This is what they use at
the factory to test these things. I would offer to do this for you, but
my spectrum analyzer / tracking generator combo doesn't have the plotter
/ printer interface option.
I do a crude evaluation...at least confirming that they cover range as
labeled by using our probe station
signal generator-rf circulator-scope). I either hook the filter up like
a probe and terminate it with a 50 ohm load
or put the filter between the output and input. This might be
quantitative enough to compare filters but I have never
done anything that exact.
If you have access to a sweep generator that will give a very clear
picture of the pass band. Alternatively you can step thru with a
synthesizer and a scope.
The obvious first choice would be to evaluate the filter with a spectrum
analyzer. If you had one available to you, you probably would already
have done that.
I have never done this, but you might try using qtune. Just connect the
filter to the tune port of the spectrometer and then to a 50 ohm load.
Set qtune to a wide sweep and see if anything can be learned.
Outcome:
According to the filter manufacturer (SFSY) the filter has one 50 dB
reject range (DC to 12 MHz), one bandpass 23-68 MHz range, one 80 dB
reject notch (76.7 MHz, lock), one bandpass 85.5-203 MHz range, and one
80 dB reject range (250-500 MHz. The filter labelling (WR19-216-BB-N) is
just a part number, unrelated to the filter's specs.
I tried to use the qtune of the spectrometer to look at the filter with
a 50 Ohm load attached to it.
I attached the filter to the tune port of the tune interface and ran qtune.
I also substituted the probe with the filter, rewired like you would
normally do for tuning the probe, and ran qtune.
In either case the response was uninformative and did not match the filter's rf response.
Many thanks to all who offered to help.
Best regards,
Lazaros Kakalis
Department of Chemistry
Olson Hall
73 Warren Street
Newark, NJ 07102-1811
Tel: (973) 353-5040
Fax: (973) 353-1264
kakalis_at_newark.rutgers.edu
Received on Tue Mar 15 2005 - 17:51:04 MST