Phasing closely spaced antennas 135 degrees – what went wrong?

[K1DNR]

Equal magnitude elemental source currents seems to be a popular assumption...

If you are feeding two verticals, end fire or broadside, with the common endeavor of achieving a null or gain in some predictable direction by using a delay line of x degrees - its not an assumption.

If you are working out some other kind of array, then its something else. I don't suppose the currents in each element of a log periodic array are equal. But its a different antenna...

[W5DXP]

- its not an assumption.

Sorry, I was just quoting The ARRL Antenna Book, 20th Edition, page 8-7: "... currents of equal magnitude are assumed to be flowing at the feed point of each element ..." That's the "popular assumption" that I was talking about.

[AA7EJ]

What is the impedance at the end of your phasing line? Does it terminate in the characteristic impedance of the transmission line?

Do you think that is important if you are calculating the phase delay/shift in the line in degrees????

What theory is this? Basic transmission lines and AC theory...

8 db gain in any direction from those two wire verticals seems a bit of a stretch to me. I'm not sure how you accurately model wire antennas in that environment either - did you build the steel ship into the model? Maybe its the salt water?

Here are some additional details.
I am taking “start with simple, known parameters.” approach.
Salt water, proximity of ship superstructure , smokestack at the end etc. are details which would be interesting in calculations, but cannot be changed so why bother for now
The array is modeled at ground level, over real ground.
The angle is not calculated, I have selected that for the purpose of directing the RF in one direction.
The 3D radiation pattern is OK, about 8db with -3db beamwidth around 90 degrees.
(I could post the picture here if interested)

Here is the actual impedance / SWR (@50 Ohms) data at each feed point from the MMANA :

PULSE U (V) I (mA) Z (Ohm) SWR
w1b 1.00+j0.00 0.38+j0.77 510.63-j1043.33 52.93
w3b -0.71+j0.71 -0.18+j0.03 4470.39-j3148.73 133.77

I am still learning how to interpret most of this data,
It looks as the “pulse” on wire 3 , the one lagging 135 degrees, indicates that the currents in the antennas are not equal. Contrary to the data presented here.

I'll change the feed, from 135 degree lag from one antenna , to point where the absolute phase difference between them is also 135 degrees.

Here is another though – if I model each antenna individually and than as an array, can I than calculate the mutual coupling impedance, complex , but still in parallel value?

But at this point I do not see the need to know that anyway, since the array impedance includes the mutual coupling and that is the value I need to work with.

[AA7EJ]

I did dusted off some antenna books and came to a startling discovery about the “equal current in phased antennas”. Both “antenna compendium 1” and an old (1970) ARRL antenna book propagates this “popular” opinion.
IMHO it applies only when the antennas are phased 0 or 180 degrees.
MMANA supports that.
Does not “compute” for any other phasing arrangement.

And BTW – the ARRL book also states that “as reflected impedance of both elements are in parallel,... the resistive impedance seen by the transmission line is 600/2 = 3000 ohms.” (sic)
And ARRL was never wrong, except in this case they just cannot divide by 2. Minor detail – typos happen.

Found nice Java applet for calculating stubs, another piece of the puzzle is in place.

[K4SAV]

When you apply a current source to a wire you force the current to be whatever is entered for the source. When you apply a current thru a length of coax, the current in the elements will become whatever the element will accept based on the impedances of the elements and the length of the coax. You can't assume any length of coax will give you 135 degrees or that any length will give you any specified amount of current in the elements. The model has to include all coax lines and all elements to yield anything useful.

Jerry, K4SAV

[AA7EJ]

When you apply a current source to a wire you force the current to be whatever is entered for the source. When you apply a current thru a length of coax, the current in the elements will become whatever the element will accept based on the impedances of the elements and the length of the coax. You can't assume any length of coax will give you 135 degrees or that any length will give you any specified amount of current in the elements. The model has to include all coax lines and all elements to yield anything useful.

Jerry, K4SAV

So far the model is just two sources - one with 0 degree phase and the other one with 135 degrees. I honestly do not know how to add real coax and what would change in the model, but I'll look into it.
I assume that to do it real I would have to include the matching stubs too.
I did calculate the coax length from the 135 degrees phase desired - about 28 feet * the velocity factor.

[K4SAV]

So far the model is just two sources - one with 0 degree phase and the other one with 135 degrees. I honestly do not know how to add real coax and what would change in the model, but I'll look into it.
I assume that to do it real I would have to include the matching stubs too.

Yes all the phasing lines, matching networks, and stubs (if you have them) need to be included into the model. When you do that you will discover nothing works like you think it does. Using two current sources doesn't simulate anything close to what you built.

I did calculate the coax length from the 135 degrees phase desired - about 28 feet * the velocity factor.

Which is the point several people have made several times and you seem to be ignoring, you only thought you calculated the delay.
If you don't consider the load impedance, the answer you get will be wrong.

If you include everything in a model you should get something approximately correct. The ship's unusual ground plane and other structures on the ship will make a difference which won't be accounted for in the model.

Jerry, K4SAV

[AA7EJ]

After consulting with folks who made suggestions instead of criticizing the idea and finding reasons why it is futile, I ended up with a simple array.
The north antenna is end fed, no problem matching it to the TX, and the south is free floating passive reflector.
The radiation pattern and desired direction is better that expected ( 70 degrees beamwidth), with max gain close to 9 db, and F/B ratio close to 12 db. Elevation / take off angle of 23 degrees.
It works from Galveston and does 75 sections in FD.
I am a happy camper.
Next project – electronically steerable 4 element sort-off vertical array.
Discussion closed and no longer monitored by me - no reply necesary.
Of course you can do whatever, it is a free country.

Happy B-day America.