VNA measurements not matching smith chart matching network design

[grg183]

I have a PCB trace antenna (915MHz monopole meander) and trying to tune the matching circuit for it. I have a FieldFox N9932A VNA (and also a miniVNA Tiny). Now I have a good understanding ( or at least I thought so ) of the matching procedure, smith charts, etc... but I'm quite baffled by what I am seeing on this one. Basically what I am getting is that any component that I try to place in the circuit moves the impedance to a totally different position on the smith chart (as measured on the FieldFox) compared to what I should get based on the smith chart calculation (using a software tool - and verified to be correct). Clearly I would expect some discrepancy due to tolerances, parasitics, etc.. but it seems that the impedance positions I am getting should not even be possible for the added component. So for example, at one point I had the impedance in the top right section of the smith chart (almost on the 50ohm circle) so I added a series capacitor to go down in an anticlockwise arc that should have brought me very close to the 50ohm target, but instead, after adding the capacitor I ended up in the top left section of the chart - which shouldn't be possible. Parallel components seem to behave better than the series ones but still some results don't seem to make sense.

FYI: I calibrated the FieldFox using mechanical calibration at the end of an SMA cable attached to port 1 and then added port extensions to compensate for an SMA pigtail with its cut end soldered directly to the board and generally seems to be measuring fine. I tried so many things that I decided to compare the results also using the miniVNA and the behavior I observed was the same.

I know this is rather vague but any idea why I might be getting these weird results?

Thanks

[D3f1ant]

What does your antenna look like? I've had almost exactly the same problem with a meandering 300mhz antenna. It simulated perfectly, but in the real world I couldn't get it to work and matching made no sense at all.

[grg183]

It's based on this: http://www.ti.com/lit/an/swra227e/swra227e.pdf
As such the antenna works and got a good range on it (have not measured range properly yet).

Also I was able to cut it down to get it to tune to the correct frequency (ignoring the matching network - just the VNA connected straight to the antenna feed point) and this length matches perfectly with the simulation.

It is just the matching network that is not making any sense.

[yl3akb]

Measure just the parallel capacitor with all the setup the same (calibration, pigtails, port extensions). Do You get what You expect?

[ConKbot]

Does the port extension take care of the group delay? Iirc one of the analyzers I've used could do port extensions (or adapter removal? Can't remember which) as amplitude only, amplitude and group delay, or amplitude, group delay and mismatch correction.   Smith charts are indeed hard to work with the phase off.

[tautech]

Just been through an exercise with a homemade J-Pole antenna for 315 MHz and the one thing I learnt was to get the emissitivity Log Mag values right on frequency before anything else. Then adjust feed point for good SWR and the rest mostly falls into place only needing small tweaks to optimize it. Smith chart gives confirmation that it's all tickedy bo.

[Wolfgang]

Hi,

could it be that the cable from your VNA to the antenna radiates ? What you could do is to add ferrite sleeves to kill RF meandering along the cable shield.

[Bud]

Did you clean the board area from flux after soldering matching components, try that.

[Bud]

Measure just the parallel capacitor with all the setup the same (calibration, pigtails, port extensions). Do You get what You expect?

That needs to be done, too.

[grg183]

Thanks for the replies and sorry for not getting back, crazy busy week.
I'll try out these ideas and will report back soon

[mik4el]

Any progress for you? I have very similar issues so really appreciate to learn from your experience.

[xzswq21]

miniVNA Tiny is reliable?
Actually I have measured a response of a filter with Keysight Field Fox VNA from 1MHz to 2GHz.
but I don't have any VNA at home. I should check the response of some filters.
do you recommend me to buy a MINIVIA Tiny?

Thanks

[clockspectrum]

Would be helpful if you could show us your design at least from the transmission line out to the antenna w matching network.

matt

[xzswq21]

Would be helpful if you could show us your design at least from the transmission line out to the antenna w matching network.

matt

I have been working on instrumentation and measurement fields. usually I design and assemble lumped filters but I should verify the response of the filters. I would check S11 and S21 up-to from 1MHz to 1GHz. now I'm checking DG8SAQ VNWA3. wish I could find a second hand one. but the near year is coming soon. is it possible to declare lower price by manufacturers?
https://www.sdr-kits.net/index.php?route=web/pages&page_id=26

[grg183]

Any progress for you? I have very similar issues so really appreciate to learn from your experience.

@mik4el actually I stopped trying With the matching circuit that I had I was getting more range than needed for the application and sadly didn't have more time to dedicate to this so I just went ahead with what I had. When I have some free time I want to get back to it and figure it out. Did you figure out the issue in your case?