You got it NB its voltage fed like a 1/2wave endfed.
The article i posted says infinite impedance at the feedpoint but that's only true for an infinitely thin wire.
Real world conductors have a lower end impedance. Size matters.
How can the two halfs cancel when all currents are in phase ?
I don't understand your logic. They would only cancel if they were out of phase.
You don't get the theoretical 3dbd because the two 1/2waves don't have enough physical distance between them.
That's where a 2 x .64wave comes in.
It spaces the in phase 1/2waves to increase gain even though it has 1/8wave of your "deconstructive" currents in each leg.
Imagine its like a 1/2wave vs .64wave vertical but it has a real other half rather than an imaginary other half mirrored in the ground.
The advantage of the .64wave is it raises the usefull upper 1/2wave further above its imaginary other half.
Even though the lower 1/8wave is deconstructive the extra height of current maxima more than compensates up to .64wave where your deconstructive currents start to dominate and high angle lobes take over.
OR your cloverleaf takes over if its a dipole.
Of course that also means that just raising a 1/2wave to the same tip height or spacing the two 1/2wave dipole legs wider using a harnes will produce a bit more low angle gain because they have the extra height or spacing without the deconstructive currents.
Are we on the same page NB.
if not what part do you think i have all wrong ?