Simulation Fun: Rotary 5-Band Antennas - Let's Compare Them

In this thread, I would like to discuss pros and cons of various multiband rotary antennas for the 10 m through 20 m bands. But I would like to discuss facts and figures rather than opinions. So, I will try to model the antennas that you think should be included.
One reason of starting this topic is my own need for such an antenna. I am not in a hurry but I would like to build such an antenna in the near future.
Moreover, I am doing a small favor to W7CJD who expressed interest in this sort of discussion in the other thread in this forum. So, one reader and active debater is practically guaranteed.
Let's start.

Fan Dipole

This is the simplest 5-band antenna that comes to my mind. A picture below shows how you can construct a rotary version of a classical fan dipole. You just need two fishing poles and some spreaders.



I copied the photo from http://www.g8njj.org.uk/index.php/antennas/96-the-batwing-fan-dipole

The first question I had to answer when modeling was: how close to one another the dipoles should be. The closest separation I could found on the Internet was 4 centimeters see the picture below. I have found this picture onhttp://www.angelfire.com/mb/amandx/dipole.html.


My first model looked like that:



Such antenna with radiator lengths optimized for the centers of bands had very high SWR at 21.0, 28.0 and 29.0 MHz (7, 497 and 6 respectively)

Please note that during all simulations, each time I was optimizing length of the radiators for the best combination of gain and SWR at the following frequencies: 14.175, 18.118, 21.225, 24.94 and 28.5 MHz. I wanted to cover whole bands except for the 10 m band where I regarded 28.0-29.0 MHz coverage as sufficient.

Back to the antenna. I increased the separation 2.5 times to 10 cm and the 15 m band was now fully covered but SWR at 28.0 MHz, though somewhat reduced, was still much too high (>240). When I analyzed radiator currents at this frequency, it became obvious that the proximity of the 10 m and 12 m band radiators created the problem. See the picture below.



Perhaps that’s why the fan dipole presented in the picture above lacks the12 m band.

Increasing vertical separation between those two radiators to 20 cm helped only marginally. So, I decided to shift the 10 m band radiator from the bottom to the top - above the 20 m band radiator. Improvement was significant but still the SWR at 28 MHz was too high. I re-arranged the order of all radiators to reduce interactions. The new order was: 10m, 15m, 20m, 17m, 12m.
I placed the feed point in the central radiator (20 m band). Thanks to that change, bending of the radiators near the feed point was as low as possible.
I gradually increased the separations between the radiators to: 25 cm, 30 cm and 35 cm. 30 cm (or 1 foot) proved to be enough separation to ensure the bandwidth I wanted to achieve. The final model has 30 cm vertical separations between the radiators and looks like that:


The dimensions are:
d=0.3, a14=5.121, a18=4.028, a21=3.446, a25=2.802, a28=2.451meters

Overall length: 11.242 m, height: 1.2 m, turn radius: 5.621 m.
Required balun: 1:1.

Find below the simulation results.
SWR - free space


SWR - real average ground, h=8 m


Note that the minimum of SWR does not agree with the optimization frequency because my settings of the 4nec2 optimizer were: 50% for SWR minmalization and 50% for gain maximization. Evidently, SWR minimum does not agree with antenna gain maximum.

So, for example, when trimming 10 m band you should aim at getting SWR minimum not at 28.5 MHz but at 28.15 MHz as you can clearly see in the graphs above.

Radiation patterns of the fan dipole are practically the same as those of half wave dipoles and I will not present them.

In the table below, I gathered calculations showing how the fan dipole compares to a half wave dipole in free space. I compared not only the antenna itself but also the system: an antenna with a transmission line (A+TL). That's because the antennas under comparison differ in SWR values so the losses in a real transmission line will be different. I assumed 30 meters (ca. 100 ft.) of RG58 cable and used AC6LA's TLDetails program to calculate the SWR as seen by a transmitter “SWR (A+TL)”, and the total loss in the RG58 cable “TL Loss”.

The final result is in the rightmost column. The fan dipole with 30 meters of RG58 when compared to a HW dipole with the same transmission line is worse by 0.04-0.74 dB. I think we can totally neglect such small difference and consider the fan dipole as equal to the HW dipole.



To summarize:
- ensure at least 30 cm (1 foot) separation between radiators
- ensure maximum physical separation between 10 m and 12 radiators
- to win extra split decibel, rearrange the order of radiators to: 10m, 15m, 20m, 17m, 12m and place the feed point in the 20 m band radiator center.

That's it.

The NEC2 models are attached of the dipoles and the fan dipole in two versions: free space and with real average ground.

The next antenna I want to analyze is a broadband fan dipole for the 14-30 MHz range. I designed it some time ago. I think you will be interested in comparing it with a classical fan dipole. But I need a few days to prepare the materials.

What should be the third rotary multiband antenna to be compared?

 

And this is my answer to Ai6KX/JS6TMW and W7CJD expressed surprise about so big separation needed.

When I reduced separation to 3 cm in my final design, the 4nec2 optimizer could not find an optimum for 12 m band. I ran SWR plot 14-30 MHz and there was not even a slightest dip near 25 MHz:

So no wonder the optimizer failed. And take a look again at the picture of the 40 mm separation dipole in my previous posting. This band is missing in the design. Evidently, the constructor was unable to tune his antenna in this band.

20m band seems to be OK but 15 m band has significantly greater SWR than in the 30 cm version:


And the 10 m band by no means cover 28-29 MHz:


So, you can build a 4 band fan dipole with 3 cm separation but without 12 m band, with increased SWR at the ends of 15 m band and with only ~500 kHz bandwidth in the 10 m band.

My goal was to make a fan dipole work on all 5 bands end to end with reasonably low SWR and good gain so that it can match a monoband dipole. If you also want to achieve it, you have to introduce large separation.

That's what the simulator predicts.

 

If you did 17-15-12 in a three band fan dipole, is 12-meter still the problem?

 

I'm not seeing any attachments. Perhaps they got lost when the qrz administrators switched forums for this thread. Or maybe I just need new glasses!

Dan, AC6LA

 

Hi Dan,

oh, no. This is not the administrator. I had to repost the thread myself and forgot to attach the models. I am doing it now.
I have tweaked the model to get the AGT as low as possible. And it is =<0.1 dB. So, practically the model is perfect.
As you will see, I connected the centers of dipoles with small segments of transmission lines to avoid direct connection of wires at sharp angles what fools the NEC2 engine. I also carefully adjusted segmentation. So, if you do not change it, it should also work fine in EZNEC.

 

And also Steve, JS6TMW had a following comment:

All I can say is that I believe my own experience. I only tried to get low SWRs on the CW portions of the bands, so I also cannot cover the entire 10-meter band with my close-spaced parallel vertical. But no way does the SWR climb as high as on the plots you made, and also for 15 meters where the SWR is very acceptable on the phone section. Granted it is only a 4-band antenna and perhaps a vertical groundplane is a different thing altogether than a parallel dipole.

I wonder how a simulation of a close-spaced 3-band dipole would look compared to my two versions (20/17/15 and 40/20/17).

Steve JS6TMW


So I would like to post a SELF CORRECTION w.r.t. a closely-spaced fan dipole (d=3 cm).

I could not explain the total lack of 12 m band in the results I posted before. It disturbed me that much that I analyzed it again. I discovered that I did not do the radiator length optimization in the right way. I wanted to reply fast to W7CJD and made an error. How do you say it in English? More haste, less speed?


So again, the SWR plot of a fan dipole with a 3 cm spacing after a correct optimization proces.
The whole frequency range:



15 m band:





10 m band results:




The SWR and Gain in the graphs above are the SWR(A) and Gain(A) values in the table posted before.

You can see now that wide spacing = wide bandwidth.

I put in the table the best values one can achieve with a classical resonant fan dipole. With 30 cm you get almost exactly the same performance as with monoband dipoles. If the spacing is small, the usable bandwitch is also small what is especially painful in 10 m band.



17-15-12: 3-band version with 3 cm spacing

Of course there is no problem with narrow WARC bands. Fortunately, also the 15 m band looks OK to me:





The dimensions of half radiators:
a18=4.001, a21=3.500, a25=2.976
The 15 m radiator is a central one in my model.

I am attaching the models for 5-band 3 cm version and 3-band 3 cm version.

And one more thing. Steve JS6TMW noticed that in practical measurements SWR does not go as high as in computer simulations. I remember reading in a manual of an antenna anlyzer that it was accurate only in some range of SWR (I do not remember exactly but probably it was from 1 through 10?). The manufacturer admitted their analyzer does not measure SWR accurately when it is very high. And i think there is no need to make an analyzer accurate at high SWR values.

 

I'v got a few multiband beam models, all of my own designe. Two models are 20-17-15-12-10m beams. Also there are 40-20-15-10 and 40-30-20-17-15-12-10m.
Since my models are MMANA-GAL ones, I'm not sure if there is any sence to attach them here

 

Why not?

Fan dipoles is the discussion.


SP3L,

I won't be able to look at nec files because it runs on Excel.

I do not have Excel, and I remove Office from my new tablet computers w/1GB RAM so I can run Visual Basic software.

3:1 and 2:1 SWR isn't so bad, I suppose.

I would like the WARC bands suspended as the photo you supplied as one rotating wire fan dipole.

However, I like better SWR to be built into the antenna.

Is the SWR improved, on 17-15-12 with wider than 3 cm spacing?

 

Thanks for all your hard work and self-correction, (insert name). The simulation and the realization seem to be converging.

Yes, commercial SWR meters are not accurate over high values, but the simulated impedance/SWR curves look like very high-Q resonant circuits. In real life wouldn't there be some R's lurking around that would soften those peaks and make the SWR plot for 10 meters look a bit friendlier?

Steve JS6TMW

PS - by coincidence I today noticed high SWR on 17M on the parallel 3-band dipole. It had had been blown around during our recent weeklong winter, and one side was hanging like a shallow "U". Part of the 17M wire was almost touching the 20M wire (they are both insulated though). A quick fix lowered the SW but it's still above what I trimmed it to, so tomorrow I have to add back some of the length I trimmed off.

 

Yes, my heart skips a beat when I see parallel elements used in a multiband beam antenna. It makes good use of a strong boom and saves on feedlines. Let's see your ideas (I can't open .nec files either)