Antenna Array Design separation of the elements in the array
Just was curious to know that λ/2 spacing should be pitch of the array or the kerf of the element spacing?
We know that:
pitch(centre to centre spacing)= 2*a(half of the width of each element) + kerf(edge to edge spacing)
If the spacing is less than λ/2 then there is increase of mutual coupling and interference and if the spacing is more than the main lobe is divided in more than one side lobes thereby formation of grating lobes.
The spacing should be from center to center of each element.
Ideally the λ/2 spacing should be adjusted (tuned) for best antenna performance using particular antenna elements.
Is it necessary to have same inter element spacing for both the x and y direction in a planar array?
No is not necessary to have the same element spacing on x and y directions, mainly because the mutual coupling between antenna elements varies for x and y.
Tuning both spacings, on x and y directions, for best antenna performance is the art of antenna array design.
Spacing also may be increased as the element count increases. It all depends on what you are trying to accomplish. If you are scanning the array then spacing must usually be decreased with increasing scan angles. Microwave Engineers Handbook (The two black books that have been around for years) have plots of gain, spacing and number of elements that you might find helpful. There is also a plot of spacing, scan angle, and the appearance of grating lobes. I often find that a spacing of around .7 wavelengths as convenient. One other thought. Use a binary number of elements. That keeps the feed network from getting too crazy.
There is a pretty good online calculator that will let you play with different spacings, phasings and feed coefficients. This will give you a opportunity to play with various settings to see the effect on the main lobe and side lobe levels.
http://antennaarraycalculator.blogsp...alculator.html
There is also an Excel array calculator on microwave101's encycopedia site.
https://www.microwaves101.com/download-area
Neither of these take into account the pattern of the individual elements. It is not to hard to combine this with the gain pattern of the individual elements(assuming they are all identical). Also these simple tools don't have any way of accounting to mutual coupling or other imperfections. That said I still find them useful and vastly quicker than a full EM simulation.