AM Dial Question for R-D's Engineering Specialists

[Steve Green NEPA]

With some recent -- chronic -- talk elsewhere about the vagabond fate of 1560 WFME in NYC, I ask here about the capabilities of multi-tower arrays. See, back in Queens near JFK Airport, their towers were the closest ones to our DX dens, nostalgia now affording it a form of pity. Despite their proximity and its occasional band-wide sear of countless symphony violins I logged a combined 54 stations on 1550, 1560 and 1570 -- and I was the worst of the four DXers in the clique.
Anyway: it appears that this STAtion Without a Home has been looking for a xmtr site since well before The Rapture. Considering the numerous multi-tower arrays in North Jersey -- 570, 620, 710, 930 et al ad nauseum -- my question goes 'Is there a minimum number of towers that can be used, diplexed or otherwise, to send a non-Euclidean signal just about anywhere within reason?'
I've seen 2-towers limited to a schmoo-shaped pattern -- equal power sent out on both sides of its axis.
3 towers in a line can get nulls, a main lobe. plus a back lobe and a few minor juts. But they also seem to appear most often in that Rorschach inkblot way, with whatever 'information' going out of one side of the symmetry line being identical to the stuff on the other side of the axis.
4 towers in a line seem to get the same balances 3 do, only stricter and tighter.
Obviously (to me, anyway) 3 towers NOT in a line and four towers in a box or parallelogram have greater capabilities for mixing, phasing, reinforcing augmentation, destructive rigging and so forth.
Now, I really don't give a flying one where of if WFME's signal goes anywhere. It seems to me that, with the towers of WOR (3 in a triangle), WWRL (4 in a trapezoid), WLIB (4 irregular), WEPN (3 crooked), WSNR (God knows how many) already up, WFME will work out some arrangement.
If they actually want an arrangement.
So, as far as DXing is concerned; you folks with the holstered soldering guns : are cockeyed 3-tower arrays and orderly 4-tower arrays capable of sending out bizarre lobes and nulls just about anywhere you want them to go? I'd love to see the oddest, most unlikely patterns you can find, from the 3- and 4-tower sites. Those 8-and 12-tower ones have begun to outlive their uselessness.

 

[PTBoardOp93]

There are limits on what you can do given a specific amount of towers and land.
The most general statement I can make is that more towers allows a less symmetrical pattern and/or a sharper null.

The two towers of WFME's day pattern produced a carotid pattern that was symmetrical and was mostly used to send less signal toward the Bahamas. The three tower night array had a much sharper null toward the west, with a small back side lobe to cover Staten Issland and New Jersey.

 

[Steve Green NEPA]

>>'There are limits on what you can do given a specific amount of towers and land.'<<
Thx, PTB!
Fwiw, this non-engineer, on an associated muse today (tower configuration) looked up a few sources that gave some credence to the limitations of in-line-towers.
Our local WPPA 1360 Pottsville owned a 5-tower array, like the 5 dots on a die. I'm surmising, as they are omni now 24/7, that with their old day pattern of 5,000 watts having gone E-W, most of it West, they used the three sticks lined up along that path. At night they'd drop to 500 watts and send the bulk South, with a wee nub North, the whole figure-8 pulled in away from both Cincy and Hartford. I'm figuring that they used that middle tower for each pattern.
WGNY 1220 Newburgh NY has very much the same daytime signal as WPPA had .... the same lobes and the same loose nulls.
WGNY-AM Radio Station Coverage Map
At night, WGNY drops power and sends the Oldies East, with a small bubble West.

Radio-Locator.com

Radio Station Search Engine

radio-locator.com

WGNY has three towers, lined up in both those directions. But with 'only' three sticks operating I'm thumbnail-guessing that they are as permanently soldered to go East-West, as their towers point. They can only aim matters 270 degrees or 90.
WPPA's five towers gave them the capability to beam 270-90 or 0-180.
Just one more question/question of ye, and then I'll both cut power and sign off.
'It seems obvious to me that three towers NOT aligned ..... instead, like in a triangle, or a dog-leg Big Dipper handle ..... can create far more possible patterns and directions than 3-in-a-row can. Not as many as four towers, but pushing real close to those possibilities."
As many have said here, available land is a big issue. Not all stations have Ponderosas available like CHML had, or WNLC New London CT, The Big Ape Jacksonville or KLIF Dallas.
In any case, the learning experience has been food for thought. 73!