RF signal averaging

[frogg]

Assuming N phase-locked receivers, each with their own antennas (assuming uncorrelated noise), does signal averaging increase gain for the sampled signals?

I use this technique for optical microscopy. Does it apply to RF, and if not, then why not?

[ejeffrey]

Yes.

[FenTiger]

How valid is the "uncorrelated noise" assumption, in practice?

[Bud]

It will increase signal to noise ratio, not gain.

[TheUnnamedNewbie]

Is this not what a phased array/beam steering is? Use multiple RX or TX with certain phase shifts, locked to one LO, to create a directive response. Only in certain directions will signals interfere constructively (IE, gain), in all others it will cause destructive interference (directivity). Even correlated noise (ie, external blockers/interferers) will be lowerd unless they are in the direction of the beam.

Your received noise floor will not be better, but your signal will appear stronger, hence better SNR.

You do need to take this into account when designing the antennas, esp. if you are using very low-gain, omnidirectional antennas to form the array. They will influence each other, and cause a change in their beam pattern, phase response and impedance.

[radiolistener]

It will increase signal to noise ratio, not gain.

and it leads to gain (relative to noise level)

[TheUnnamedNewbie]

It will increase signal to noise ratio, not gain.

and it leads to gain (relative to noise level)

But it does lead to actual gain. The received power will be larger, as multiple antennas are the same as getting a larger aperture area, which will get more of the energy distributed over space.

[RoGeorge]

Yes.

[radiolistener]

But it does lead to actual gain. The received power will be larger, as multiple antennas are the same as getting a larger aperture area, which will get more of the energy distributed over space.

yeah, in any case, this is a gain. When you use single antenna and average for N times longer reception time, or using N antennas for 1x time, the result will be the same - gain. The signal amplitude will be increased N times, while noise floor will remains the same (due to random nature of the noise).

The same gain happens in high Q antennas (shortened resonant antennas, such as magnetic loop). The output amplitude of high Q antenna will be Q times higher in cost of more narrow bandwidth BW = f / Q, because high Q antenna needs Q oscillating cycles in order to accumulate RF energy. The same averaging which is performed by nature... 

[RadioNerd]

It will actually have the same effect as combining 4 antennas in phase with a hardware 4:1 power combiner to a fixed antenna array.
This will influence the radiation pattern of your receiver. The gain in "forward" direction (perpendicular to the array) will increase while reducing the sensitivity to signals incoming from other directions. Depending on the physical distance of the four antennas to to each other you can have various side lobes and nulls

[frogg]

It will increase signal to noise ratio, not gain.

Right...I understand Bud's point. To be very strict about my terminology, and to be in harmony with others posts about gain being valid, maybe I should say that the Signal to Noise Ratio Gain is increased.

I believe the term "Signal to Noise Ratio Gain" is probably more strictly accurate and academically valid, at least in this context?

[ogden]

I believe the term "Signal to Noise Ratio Gain" is probably more strictly accurate and academically valid, at least in this context?

I am not sure that "Signal to Noise Ratio Gain" is academically valid because can't recall/find any use of it. Obviously I may be wrong & corrected. Better just say that SNR increases (becomes better).

[radiolistener]

in signal processing such feature is known as "process gain" - the gain caused by signal processing