On 2/6/23 9:26 AM, PDXer wrote:

That would likely be the S11 magnitude - in the passband, the filter
lets the RF through, so it sees the 50 ohms on the other side, so
reflected power is small (i.e. good match). Outside the passband, the
filter probably reflects the power back, so S11 is close to 1.

S21 should show the bandpass characteristic (i.e. small magnitude (large
negative dB) outside the passband and big magnitude (small negative dB))
inside the passband.

Do you set the frequency range say from 20 Mhz to 50 Mhz, and set the centre frequency to 30 Mhz, and then sweep the frequency to get the graph?
What about SPAN and RBW (band width) - what do you set it to?
And it is set to read LOGMAG of S21?
What is supposed to happen on S11?

I couldn't find a schematic for the test board, so I don't know what the BPF looks like.

Span is just the size of the sweep - either you set start and stop, or set center and span.

RBW leave at the default - it doesn't make any difference.

What I'd expect on logmag S21 is that you see the bandpass characteristic of the filter.
What I'd expect on logmag S11 is that you see the input return loss, typically sort of the inverse of the bandpass, but often, you can see the individual sections of the filter.

If I can find a schematic, I can simulate it.

If you're using Windows, you might try getting a copy of Elsie from Tonne Software. The student version is free, and it lets you build filters of various kinds, either from specs, or by knowing the component values, and it can give you plots of S21 and S11 (and lots of others)).

Elsie and a VNA is a great pair. You can design in Elsie, build the filter on a board, then sweep it with the VNA.

If you want real fun (?) you can build a 3 section LC filter with variable L or C, and use the VNA to tune it. (which is more art than science, btw) It's a LOT easier than, say, tuning the cans on a diplexer, because your run of the mill LC will have lower Q.