So long story. I have had some trouble getting the Nano VNA to hold a calibration and display the correct information. I always had a -10db offset when using certain SMA leads i had there. It would display fine with semi rigid SMA leads, but, others would show -10db. Like the leads had loss in them.

Well anyway, I think i have resolved the issue and got the calibration right. 0.5db loss in my leads would be about right and that is what is showing now, they are after all cheapest crap leads from China, not high end leads you would use in a lab.

So anyway, popper calibration procedure.

Open: Open Load on S11 port.
Short: Short load on S11 port.
Load: 50 ohm load on S11 port.
Isolation: 50 ohm load on both S11 and S21 ports.
Through: Shortest high quality 50 ohm cable connecting S11 and S21 together.

Then save that. That is it, that should then give you fairly accurate, well as accurate as the NANO VNA is results. As you can see by the plot of a 40m bandpass filter above, it looks about what you would expect from a known design that has low insertion loss. A couple of dB, made up of the lead loss and the filter loss. This means my filter has about 1dB insertion loss. That is something I can live with.

Anyway, through no fault of my own, I broke the mini USB socket off the original NANO VNA i bought and tore the tracks off the board. Yeah not really happy with myself but it is what it is. So i bought another one, this time it seems to be one of the better clones, it even came with shielding and a battery. So anyway, i plan this time to ruggedise my NANO VNA and mount the whole think into an aluminium box, and have an panel mount USB port on the side that will take the abuse of me pulling and stretching and inserting a lead in and out on a regular basis. So i ordered one of these as well, yeah the new VNA came with USB type C not that stupid micro USB rubbish. Anyway, by the time i mount this in a box, i will never have to worry about breaking the damn socket off the board again.