Antenna match woes

[mayor]

Hello! I've matched a couple antennas before using the same methodology. In this particular instance, I am not able to get where I need to and am turning to people with more experience (you!) for some insight. Trying to tune a PCB antenna to 915 MHz. Pi filter footprints are in place.

A semi-rigid cable is attached at a 50 ohm point (just after a balun), just before the pi filter to the antenna, so:

IC -> balun -> 50ohm point (sma cable here) -> pi filter -> antenna

This image shows the calibrated VNA with a 0 ohm resistor in the series position of the pi filter:



OK, so from that, replacing the 0 ohm resistor for a ~5pF capacitor should bring the impedance down along the constant resistance circle to here, or just about, right?




Here's what I get instead:




As stated, in the past, I got results (mostly with inductors) that matched theory more closely. In this case, not too sure why I'm observing this, and what to do next.

[tszaboo]

This might be obvious, but did you remove the balun?

[mayor]

Hi, yeah, anything before the measurement point is out of circuit.

[tszaboo]

Hi, yeah, anything before the measurement point is out of circuit.

OK, I have very limited experience in antenna tuning, that is mostly for 2.4GHz PCB antennas. I'm using 0402 RF rated capacitors and inductors from WE and Murata. I've found that regular 0 Ohm resistors might be quite terrible, so I usually ended up shorting the filter tabs with a small piece of copper, or start from the smallest inductor instead of the 0R resistor. Also found that my tuning software (Megiq VNA) is unreliable with the results, so -50%+100% easily can be on the component calculates. So if it tells me that I should place a 3pF capacitor somewhere, i do that in small steps, otherwise the lines seemingly jump around, and I dont know what happens. So maybe take measurement with 1, 2.2 and 3.6pF first, and see that if it is moving the right direction. But then again, for lower frequency, these might be too small steps.

[mayor]

Alright, will try something other than the 0 ohm resistor to begin.
I am using Kemet HF caps and Johanson inductors suited to the purpose. I think there's always a lot of play regardless of the parts, in terms of how repeatable measurements are, but I am "used" to that.

[thinkfat]

Are you trying to match an antenna that is not resonant at your target frequency? I'd have expected a dip in the S11 logmag plot somewhere.

[Gribo]

What is the footprint of your components? If it is a monopole, do you have a sufficient ground plane? Can you post your layout?

[mayor]

Hi guys,

first, thank you for taking time to help. This is an updated layout for a board which is known to be working well (though ground plane is small!). Components are 0402, antenna is resonant around 915MHz. Here's the "updated" layout:



Now. It turns out I had messed up my port extension, which explains the odd behaviour I was describing above. With that sorted out, matching components behaved as expected:



Once again, thanks, and apologies for wasted bandwidth!