power divider tactics for phased array

[yefj]

Hello i have the following table shown bellow.
i want to design a power divider network for the given nomalised amplitudes.
amplitude is in unit voltage so the unnormalised form is to multiply each voltage by sum X value
so assuming input power is 1[Watt]
X^2(a1^2+a2^2..+a8^2)=1

1.first step for me is to imagine the nomalised weights in Watt{power form) where the input is 1 Watt.
so i take the sum of the whole amplitude and divide each member by the sum of the whole column as shown bellow in the matlab code.

in the end my power input in each port is as shown bellow:


0.0371 0.0676 0.2043 0.3595 0.2197 0.0368 0.0368 0.0382

how do you reccomend to do a power dividing network to get these power values in the end?
Thanks.

amplitude_norm_V=[0.3213,0.4336,0.7539,1,0.7817,0.3201,0.32,0.3261];
x=sqrt(1/(sum(amplitude_norm_V.^2)));
real_voltage=amplitude_norm_V.*x;
real_power=(real_voltage.^2);
sum(real_power)

 

[FvM]

I don't see a sense in scaling power to 1W total. First step is to decide about a power divider topology, e.g. binary tree, second step to calculate the divider ratios for each divider.

--- Updated Feb 20, 2022 ---

For design of unequal power dividers, see e.g. https://www.microwaves101.com/calculators/871-unequal-split-power-divider-calculator

--- Updated Feb 20, 2022 ---

If you don't need the crosstalk suppression of Wilkinson divider, you can use a simplified version with two instead of four TL segments per divider and no resistors.

 

[yefj]

Hello FVM,is there some tactics regarding distributing the phase on the network so i wont need to make 170 degree phase shifter befire the out put.
i prefer to make the big phase shifts in the root of the tree ,s there some tactics you could reccomend?
thanks