Topband: Eznec/ 4sq

[Tom W8JI]

I then bring in the other three element and enter sources of 0 degrees 2 
elements at -90 and the front element at -180 when I check the resonance of 
the array the frequency drops far below 3.795.

- at this point I would add that I was modelling a reduces height element of 
41.5ft with 4 11ft 10inch top loading, could this be caused by the top 
loading increasing the mutual coupling and lowering the array 
frequency?>>>>>>

When you couple multiple elements in a unidirectional array, the elements 
will considerably shift reactance and impedance. What you see when you test 
one element is nothing like what you have when they are driven. There is no 
way to stop this, other than swamping the elements with a very high loss 
resistance. (The passive receiving arrays I developed work this way. They 
have so much element loss the effects of mutual coupling are diluted to the 
point of being meaningless.)

I'm assuming you built an uncompensated textbook hybrid. Since you have no 
compensation for the effect of mutuals, you should expect a very noticeable 
frequency shift.

Since the impedance changes and the hybrid is not terminated in the design 
impedance, phase shift is not actually -90 and 180. It is  something other 
than that, because elements are reactive with very different impedances. The 
only two elements with the same impedance are the center two elements. That 
pair is different than the front element (that shifts up in impedance) and 
the rear (that shifts down in impedance).

With a forced 90 shift, which the 4 square you built really does not have, 
the impedances are roughly:

Rear element Impedance = -1.696 - J 15.96 ohms

Middle 1 Impedance = 39.35 - J 19.69 ohms

Middle 2  Impedance = 39.35 - J 19.7 ohms

Front  Impedance = 59.58 + J 54.55 ohms

Notice the rear element is a negative impedance, which means it puts more 
power back into the feedline than it takes from the feedline.

Since the impedances are not what you planned on having with a hybrid, you 
do not have the current distribution you think. The array also does not have 
the phase shift you think, because the impedances are not even close to what 
the hybrid calls for.

<<<Second question - is how do I phase two four squares? do they have to be 
in line for my preferred direct NW from the UK which gives me in line SE for 
LP.>>>

The highest gain is with two arrays broadside, ideally at least 5/8th wave 
apart broadside.  In that case you can get roughly around 3 dB gain, 
depending on the base patterns of each cell and exact spacing.

You can get end fire gain with 0 shift between cells, and that would take 
about 1/2 wave center-to-center cell spacing in line with the target. The 
array would have about 2 dB gain, but one of the elements would go highly 
negative in impedance. Since the impedance errors get even nastier than with 
a regular uncompensated 4-square, you are very unlikely to see any real gain 
in a typical array. You might even see loss.

If you compensated impedances, you could get the ~2 dB. But.... the array 
would waste elements and physical space. There are other phasing methods for 
in-line (end fire) but all of them complicate distribution because the array 
actually wants a binomial current distribution.

I am looking to achieve more forward gain than my 8 element array and two 
shortie type 4 sq's would be a good start.>>>

........or a bad start. I'm not sure how the eight element array was 
planned, but most Ham arrays work far less than optimally. :-)

If the eight element is poorly planned or implemented, then a pair of poorly 
planned 4 squares could be better. If you really want gain, then you might 
want to start with a better initial directive cell or just improve the eight 
element.

I hate to sound so negative about this, but this is just how stuff really 
is. Commercial 4-squares, at least those I have had contact with (like the 
current Comtek and the DXE), are not the same as slapping a 50-ohm 90 degree 
hybrid on four elements and calling it done. They have some planning to 
correct for mutual coupling shifting element impedances around. A 4-square 
built with a perfect textbook hybrid is a dismal performer.

73 Tom