Antenna safe Distance

[Dummyeng]

Hi
i designed 5 antenna in different frequency 0.4G 0.9G up to 6G and they have different BW 7M to 100M
the power of each one is almost 20w

i bought different monopole antenna for each band

and i like to put all of in and stand

Coupling Matter

i dont like couple my antenna to each other so

first should i put freq with for ex 0.4 and 0.9 near each other or 6 and 0.4?

second how can find out the most near and safe zone that i can put all of them together?

 

[vfone]

In dB, what is the maximum acceptable coupling between antennas, which you are looking for?
You can do whatever innovation to minimize the coupling between antennas, but some coupling will still exist (more or less).

 

[Dummyeng]

I look for kind of sorting that make the smallest coupling between antennas for ex circular or rectangular or ... and i think simulation in hfss isnt match in reality in this kind of stuff
My antennas dont have ground plane

 

[biff44]

well it CAN be an issue.
An antenna that is transmitting 20 W could couple into a nearby antenna trying to receive a faint signal, and that faint signal would be jammed.
In a worse case, the receiver could be damaged by the high power if there was strong coupling.

i guess i would set up the antennas, hook up a signal source one at a time, and use a spectrum analyzer to see the leakage back out of the adjacent antennas. If it is too high a leakage signal, if the frequency bands are different a good bandpass filter might fix things.

One thing to realize, signals DO decay very rapidly with distance. Even just a few feet away will see many tens of dB of isolation.

 

[Azulykit]

I see coupling discussions relative to the safe distance above. Is that what you are after or is this a question about radiation hazard to people in the vicinity of the transmitting antennas? Radiation safety evaluation usually boils down to establishing a safe (surface) power density at a distance from the antennas. Near/far field modeling comes into the discussion. Unfortunately 1/ r^2 modeling is usually not sufficient.