New??? Antenna Design, 4NEC2 Modeling

[jonwienke]

I had an idea for a wideband antenna design that I think is new, or at least I haven't encountered it in my travels on Google and other sources researching antenna theory.

The basic idea is to take a folded 1/4-wave ground plane antenna whose overall length is set to be resonant at the lowest frequency you're trying to receive, and starting a short distance above the feedpoint, periodically shunt the parallel vertical conductors with capacitors, starting with small value capacitors and small intervals, and gradually increasing the capacitor values and the intervals at which they are spaced.

So if you wanted to operate 100-800MHz, the overall length of the verticals would be selected to resonate at 100MHz. But to allow the antenna to perform at 800MHz, the first shunt capacitor would be placed approximately 1/8 of the way up the verticals, and its value chosen so that a 800MHz signal goes primarily up one vertical from the feedpoint, through the capacitor, and down the other vertical to ground. Other frequencies between 100 and 800MHz would be tuned by proportionately larger shunt capacitors placed proportionately farther up the antenna verticals. The idea is to make a more elegant fan dipole (or rather the top half of one) that can be designed to be resonant at more frequencies without having a Christmas tree of vertical elements interacting with each other.

I'm trying to model the concept, but I don't see in 4NEC how to model the shunt capacitors as part of the active antenna structure.

 

[prcguy]

A folded grounded ground plane is a parallel transmission line that is shorted at one end. It just happens at 1/4 wavelength long the folded section that is grounded is an inverse of the section that goes from the feedpoint to the far end and the folded grounded section is transparent to the radio and it simply operates as a 1/4 wave monopole. Shunting across the folded element would introduce a bunch of problems similar to shunting a parallel transmission line with capacitors and would probably not produce any useful results for a multiband antenna.

There are other ways to cut a vertical monopole and insert parallel resonant LC circuits (traps) to make a multiband resonant antenna. Some companies like Shakespeare have used parallel inductive/restive networks placed strategically along the vertical element to make it resonate at many different frequencies and this technique is used on some of their broadband military whips.

One of the best, if not the best scanner antenna I have encountered was a guy that made a very fat dipole from 4" wide circuit board with 4" dia capacity hats on each end and over the length of this circuit board dipole he made strategic cuts and bridged that with LC networks similar to traps.

This antenna had a fundimental resonate frequency around 50Mhz then the various cut sections resonated in the popular amateur, aircraft and public service bands up through 800MHz. He then had a nice preamp at the base and the entire antenna was housed inside a PVC radome several feet long. This antenna was designed in the 1980s and they designer mentioned his antenna computer program at the time had to run about 6 hours to crank out the dimensions and LC values for this design..

Maybe a few dozen of these antennas were produced and they work better than any other scanner type antenna I've ever used. The output level was a little hot but when run through a 4-way splitter it was just fine for any scanner. The preamp chips used did not seem to have any IMD problems when used in my RF hell location near Los Angeles.

I've since broken the example I have and I've lost contact with the designer but someday I hope to contact him again to see if more of these can be made.

 

[jonwienke]

I understand your point, and realize that antenna designs not in common use may well be not in common use for a reason. But for educational purposes, I'd still like to model the design and run performance simulations, if for no reason than to better understand why it won't work. How do you model an antenna with L/C components integrated into the structure of the antenna?

 

[Lauri-CoyoteFrostbite]

Hi Jon....
.
I am going to throw out another idea you might consider in your wide band antenna designing quest - the Fractal.
.
My designs were used in quite a number of wide bandwidth, high UHF/microwave systems -- and they are quite remarkable antennas. I'll spare you the operational parameters but if you take a look, I think you'll be impressed. Fractal antennas are infinitely easier to engineer than series lumped verticals, etc.
.

.
.
Lauri
.
.

 

[prcguy]

I've not analyzed a trap antenna with EZNEC yet but there is a LOAD button in the program that brings up another menu and you can select RLC as the load. You will need to know some stuff about your trap to enter in the program.
Unfortunately one of my antenna mentors and EZNEC experts is dead, so I'm a little slow at analyzing antennas these days.

On the Fractal antenna, is there a hobby type design program geared towards that?

I understand your point, and realize that antenna designs not in common use may well be not in common use for a reason. But for educational purposes, I'd still like to model the design and run performance simulations, if for no reason than to better understand why it won't work. How do you model an antenna with L/C components integrated into the structure of the antenna?

 

[jonwienke]

I've not analyzed a trap antenna with EZNEC yet but there is a LOAD button in the program that brings up another menu and you can select RLC as the load.

I've looked at that, but I don't see any way to connect it into the antenna at a specified location. It seems to assume that the RLC entity is connected at the feedpoint or something, which limits its usefulness.

 

[jonwienke]

OK the fractals are interesting. Most of the designs seem to be variants of flat panel directional antennas. The most interesting version to me seems to be the triangular Sierpinksi design, which appears to lend itself to being translated to a 3-D structure from a 2-D structure (pyramid vs a flat triangle).

Also, I wouldn't find the details boring, I kinda specialize in getting into the weeds on stuff I find interesting...

 

[prcguy]

You can assemble an antenna with traps, coils or just about anything in EZNEC exactly as it is in real life and get very accurate results. Maybe others can assist on how to make and connect a trap in the program.

I've looked at that, but I don't see any way to connect it into the antenna at a specified location. It seems to assume that the RLC entity is connected at the feedpoint or something, which limits its usefulness.

 

[jonwienke]

You can assemble an antenna with traps, coils or just about anything in EZNEC exactly as it is in real life and get very accurate results.

I've no doubt it's possible, I just am not seeing how to do it.

 

[N9PBD]

This might help a bit, Jon: http://www.w5ddl.org/files/ComputerModelingSimple6.pdf

 

[N9PBD]

You also might find a copy of the ARRL Antenna Handbook at your local library and take a look at some of the info. I've got the 23rd edition here, from 2015, and it has a chapter on computer modelling.

 

[jonwienke]

Unfortunately, no. It confirms that it's possible, but doesn't explain how to do it.

 

[Lauri-CoyoteFrostbite]

.
I didn't mean to change this subject--- the fractal systems designs I still have access to are not public domain... no cloak and dagger stuff- but that means nothing - for there is a plethora of other information available out there- the formulaes, programs-- etc. etc..... I will add in closing out my two-cents, however, that once you step away from the 2D designs, Fractal antennas get very (!) interesting....
.
.
Lauri

 

[jonwienke]

I figured out the inline LCR thing. It's not the most intuitive thing, but I got it working.

 

[k7ng]

I'd like to see your model.

 

[jonwienke]

Meh, it's not that great. With the right shunt capacitor values and placement, it does a little better than a folded dipole, but it throws RF lobes in all sorts of non-useful directions, so it's not a very good design.

BUT

I've been playing with a dual-biconical design that I'm still optimizing, but so far it keeps SWR under 4:1 from about 100MHz-1300MHz.

 

[jonwienke]

Here's the model:
https://1drv.ms/u/s!ApJIS-l4xqPtgvl1586lsiix8HJsCg

It has some ugliness in the gain pattern, but I was able to move it into the mandatory coverage gap between 512MHz and 758MHz in most scanners. So it has decent performance from about 125MHz to about 590MHz, and then has another band of good performance between 720MHz and 975MHz.

Constructive feedback welcome.

 

[jonwienke]

OK, 4NEC2 is driving me nuts. It's giving me results I know are bullcrap. It's giving me negative SWR numbers, reflective coefficients > 0, and negative R impedance values. I understand X impedance can be negative or positive, depending on which way the capacitance and inductance are imbalanced. But I don't see how R (the resistive component of the antenna's impedance) can be a negative number unless the antenna is violating the second law of thermodynamics.

Is this unfixed bugs in the software? Or do I have some geometry or segmentation error in my model that the error checks aren't seeing?