measuring insertion loss of filter with network analyzer [help needed]
i am currently working on a harmonic localized motion focused ultrasound transducer experiment and i must design passive filters to remove the harmonics drawn in by the transducer.
one of the main components of my design is adhering to a low insertion loss (due to the low nature of the signal)
how would i accurately measure the insertion loss of my filter with a network analyzer? step by step instructions would be greatly appreciated.
if this helps, the filters operate in the 4~10Mhz region ( I have several notch filters, a low pass, and a high pass filter [the band pass had to high of an insertion loss - or so I was told])
I couldnt find any other thread that answered my question adequately...
Thanks in advance,
Joey
It's always a good idea to press reset first.
Set your start and stop frequency. Start as low as it goes, stop around 20 MHz.
Sweep type will be linear frequency. Measure will be S21, display log mag.
Connect coax cables to port 1 and port 2. Go to calibration menu. Since you're just measuring gain you can just do thru cal. Connect a barrel between your two coax. Run the thru cal. When you're done with the cal, the display line should be nice and flat and right at 0 all the way across the screen. Disconnect the barrel and connect your filter to the two coax. Adjust the vertical scale as necessary.
Thank you. I am familiar with the calibrations (although I would perform all of them, thru cal only would simplify and shorten my measurements) and the scatter setting (S21), it is only the measure for insertion loss.
What I am looking for is average insertion loss over my pass band.
Is it simply to measure different values across my band and take the average, or use the delta function?
Thanks again.
Yes, for insertion loss you only need to measure S21, and thru cal is sufficient (big time savings).
As far as the average in the passband, different analyzers may have different math functions built-in, but I would probably export the data and do the math in a separate program (Excel, Matlab, Mathcad, etc.). Although, if the passband is smooth, you might be able to get "close enough" by blowing up the vertical scale and eyeballing it. There's going to be some variability in filters anyway, so accuracy to within more than a couple tenths of dB is probably not necessary.