DIY RF project, higher noise floor than anticipated

[mayor]

Hello,

I've created a board for 868MHz communication, and matched its pcb antenna with a VNA. Note that measurements were taken from a micro-connector connected to the transmission line, not at the transmission line itself.

According to my measurements, I have achieved a good match (appx 1.3:1) at the target frequency. When measuring RSSI, it _looks_ like I am receiving pretty decently. But looking at noise floor, I am roughly 15 dBm worse than another antenna (connected to the same connector as I used to measure my pcb antenna).

The question I have is: can a poorly matched antenna produce the observed high noise floor? Or could this be a factor of the pcb antenna itself?

[RadioNerd]

a good antenna is also more sensitive to rf interference in the same band...
It's possible that your antenna is just picking up other signals (eg. mobile phone base station?) that were too weak to detect before...

[mayor]

Ah so that's interesting. Is it theoretically possible that you'd get better SNR from a slightly detuned antenna?!

[MiDi]

If you want to measure SNR you will have to do it in an rf shielded test chamber.
As stated before the seen noise floor is not only noise 

[coppercone2]

Can you band limit both antenna with a filter to see if they act differently?

[mayor]

If you want to measure SNR you will have to do it in an rf shielded test chamber.
As stated before the seen noise floor is not only noise 

Well, that's kinda outta my reach :-) But I'm not looking to characterize the antenna, so measuring RSSI in the "real" environment vs. noise floor suits my purpose. The idea being, of course, to have the best possible link budget. The second antenna's noise floor is at around -110dBm (15 or so better than mine), and I think its RSSI under same transmission conditions is similar to mine (like -14dBm within a couple feet -- but I should double-check that).

[mayor]

Can you band limit both antenna with a filter to see if they act differently?

Not readily... I have a 2.4GHz AND the 868MHz antenna on the board. I could experiment with some diplexers in a future revision of the board.

But to the question at hand: would a bad match also manifest itself not only as bad RSSI when receiving, but as a higher noise floor? I have been told that, but I don't understand why/how that would be so.

[coppercone2]

well you do heat the amplifier with reflections ever so slightly, thermal might be one reason (probobly not dominant).

It might be getting heated in points that are odd/unnatural which causes a weird stress on the die compared to regular heating or something.