clanon

parallel stripes (transmission line ) width...separation ? distance to reflector...?

the green PCB seems DOUBLE SIDED with lots of THROUGH HOLES...

https://www.mdpi.com/2079-9292/9/10/1574/htm

looks a lot like EIA dual dipole feed ...but wideband

Yeah , you're RIGHT... maybe ANGLED slots in a little box (cube or sphere)

Petra BB the best (i'll try two of those at 90 and see)

MagnetoElectric Dipole (MED) more bandwidth , less gain and low efficiency on Dielectric ( ALL metal should be better) PS : and the TUBE DIAMETER doesn't help...either

12 to 15 dbi and a nice 50 ohms curve in the whole bandwidth... show more details on the INFINITE like shape... so it behaves like 3 half waves SLOTS...? (wideband) What happens if that shape is multiplied into a flat BOX...? (4 , 6 , 8 of those)

Backfiring circular patch (CHINISHIT) NOBODY ever got more than 27 -28 dbi at those frequencies (800 mhz to 2700 mhz ) you'll need more than 2 meters diameter to get over 30 dbi ... ALL lies and good looks

1mm aluminum size should be increased X,Y by 10% (1.1) Dielectric used must be known (Boom and spacers) impedance close to 50 from 2 to 2.5 ghz Vswr should be adjusted (exact materials needed and Bandwidth) "More efficient (NO DIELECTRIC) , a little more gain"

More efficient (NO DIELECTRIC) , a little more gain , imho

it should work a BALUN 4:1 is NECESSARY

Pre made , wonderful...and all dimensions should be available before buying it , for trial and error Jomy the supports seems to be PTFE (teflon)

Similar

the Blue and RED coax should be NEW CABLES straight to the card...(if possible to solder in there) High losses

BANDWIDTH at 0.4mm diameter would be THIN (some Mhz) thin wire High Q Factor , small bandwidth

bigger diameter most of the times means wider bandwidth (wich is GOOD) some changes in impedance may occur (folded dipole won't change much) it's always 4 times that of a single dipole in free space 288ohm...(that's the reason why you need a BALUN 4:1 most of the times) "The impedance of this antenna is around 300 Ω, so you may want to use a transformer or a λ/4 long piece of 120 Ω coaxial cable to match this impedance to 50 Ω. " Use styrofoam to hold it (polystyrene FOAM , the lighter the better)

GOOD INFO...i read somewhere that elliptical was very good for tree penetration...(most of the time two dipoles angled and 90 degrees delay)

NEMO i mean

What's the difference between directivity (any polarization) and circular...? MEMO are you still trying with the 3 layer PCB...?

i would put the CENTER line of the oval (folded dipole) aligned with the directors...if possible https://www.wifi-antennas.com/uploads/monthly_2021_01/Captura.JPG.30e5234d97b2075b2469fad7ad1a6c18.JPG

well any knock off with an AR9271 should work...but firmware makes a big difference I have a 722 and a 721 (both ar9271 , same hardware different firmware) HUGE difference then the rtl8187B was digging up all the APs the 722 left unseen and hacking them on the fly finnally the wn7200 (r3070) with some issues with Linux scripts...

wikidevi having 0+1+2+aux antennas doesn't seem like the BEST candidate for a single antenna...firmware hardware conflicts...and all for hacking you should try a in 2,4ghz the best are (in that order) TP WN722 (version 1.0 1.1 ONLY) RTL8187 (any flavor but imho 8187 B ,BvE ,SE, L) and any r3070 card