I simulated a bunch of QHAs and dual QHA combinations using the NEC2 simulation software. To aid in generating the NEC2 models, I wrote a simple and stupid (and incredibly ugly) program in C to automatically generate NEC2 source code according to given QHA dimensions (height, diameter, number of twists and so on). It can generate either a single QHA or multiple QHAs stacked according to your preference. It does not generate QHA dimensions for your design frequency—for that you'll want to use John Coppens's Quadrifilar online calculator. Take the dimensions it gives, and plug them into helix2nec.
If you like, you can download helix2nec.c. It's not pretty. You can also download an example input file, dual_435.helix, which creates a NEC2 model of the 2 m and 70 cm combination, and tells NEC2 to simulate it in the UHF range.
The input file's structure is as follows:
Compiling the program:
gcc -lm -ohelix2nec helix2nec.cUsing the program:
helix2nec blah.helix blah.nec xnecview blah.nec nec2c -iblah.nec -oblah.out xnecview blah.outDo not complain about what the program does, or doesn't do. I do not support it in any way.
|Direction of the helices: This is rather simple. The picture on the left shows one QHA with left-hand twists. They are in the opposite direction as the threads on a screw. A QHA with twists going in the same direction as a screw, as shown in the images on the right, has right-hand twists.||
Connection of the feedpoint: This is a bit more subtle. When the
antenna feedpoint is at the top of the antenna (which for some
reason seems to be the commonly used arrangement, perhaps because it
enables making sleeve baluns for the feed; something I've never done
myself) the two loops can be connected according to the diagram on the
left. Note that the diagram is looking down on the antenna feedpoint from
above; as labeled on the diagram, it is
a top view. The coaxial feedline is connected to the two thick black
areas (they represent conductors on circuit board).
It does not matter which one is the center conductor and which one
the coax shield. I shall hereby call this connection the "Standard
connection" (just because this is how I connected the QHAs at my
These feedpoint connections can also be mirrored. Be extra careful here, because the change is subtle but has major consequences. Still keeping the feedpoint at the top of the antenna, and still viewing the feedpoint from the top, if you connect the smaller loop where it says "Large loop" and the larger loop where it says "Small loop", you get what I shall call the "Anti-standard connection". (This is actually the mirror image of the standard connection.)
The first row represents my station's QHA antennas: skyward radiation with right-hand circular polarization. The second row, where both the twist direction and the feedpoint have been mirrored, is a complete mirror image of the entire antenna. It is therefore logical that it should otherwise work the same as the original, but have opposite (left-hand) polarization.
The third and fourth row describe antennas that differ from the first one in only one parameter (twist or feedpoint). They provide either RHCP or LHCP polarization, but they radiate downward, into the ground. That might seem useless—unless the antenna is hung from its feedpoint (feedline going up), and used to transmit telemetry from a high-altitude balloon! Alternatively, you can mount the entire antenna upside-down, i.e. with the feedpoint at the bottom, and get skyward radiation. Try this, if you think a bottom-fed antenna would be more convenient for your application.
If you build a QHA for satellite use and find that it does not work, please
check, double check and triple check the twist direction and
especially the feedpoint
connection, and their resulting radiation direction with respect to the
location of the feedpoint. It is all too easy to design a QHA for sky
coverage, but build one that radiates down into the ground.
Various double QHA designs
Here are some designs I simulated before building my pair of QHAs. Since the
QHA structure is "hollow", I was most interested in the possibility
of placing the small 70 cm QHA inside the larger 2 m QHA,
and wanted to
see if any unwelcome interaction would result between the two antennas.
Note that these models and simulation results are for LHCP antennas,
whereas I finally chose to build RHCP ones! The RHCP versions are
just mirror images of these antennas
how the helix direction and feedpoint connection affect the radiation
pattern and polarization), and the simulation results are equally valid
for either case.
(Click on the images to see bigger ones!)
|This is a NEC2 model of a single QHA. The feedpoint is at the top.|
|This picture shows what I wanted to make. The two QHAs are completely independent, but the 70 cm QHA is built inside the 2 m QHA. Since the 2 m antenna is resonant also on 70 cm, I was worried the outer antenna might shield the inner one, or at least change its impedance drastically. Therefore I created NEC2 models of a single 2 m QHA, a single 70 cm QHA, and various combinations of the two. Some main results are presented below.|
|This is what I eventually built. As it turns out, the two helices can be placed as close together as you like, without much interaction. With just 10 cm space between them, the radiation patterns of both are essentially unchanged. The final reason why I chose this design instead of the one above, was ease of construction, plain and simple. I still think the above design would have been more elegant, though.|
Some simulation results
Only one QHA will be fed at a time, so in the dual helix models, I had only a single feedpoint, feeding the antenna whose performance I was interested in. The other antenna was either terminated in 50 ohms, left open, or shorted, to mimic what would happen when it's connected to a diplexer or an antenna relay. Apparently this had very little effect on anything. I terminated the antenna in the following simulations. (I did not try what would happen if the two antennas were fed in parallel.)
I started by simulating a single QHA. Once I had all the bugs sorted out in my code, the simulation results were in good agreement with Coppens's design tool. The resonant frequency was just a tad off, and SWR was not exactly 1.0, but this could just as well be a non-ideality of the NEC2 simulation. So rather than focus on the SWR curves produced by any one simulation, I compared the results of single helices to those of dual helix arrangements.
|On 2 meters, the SWR curve of the dual helix arrangement was practically identical to a single 2 m QHA. This is to be expected, as the UHF QHA has no resonance on 2 m. However, the VHF QHA does do something funny on 70 cm, which is its third harmonic. It looks like it's not a good enough resonance to be useful (though when testing the completed antenna I noticed it performs just fine also on 70 cm), but I was afraid it would mess up the 70 cm antenna's performance. But as it turns out, the SWR curve of the dual QHA combo at 70 cm (right-hand image) remains nearly identical to a single 70 cm QHA (left-hand image). The minimum SWR frequency does move a few MHz, from 440 to 442 MHz.|
So far so good: the SWR of either antenna remains unaffected by the presence of the other, and even the resonance frequency is unchanged. Also the radiation pattern of the 2 m antenna stays clean, but what does change is the radiation pattern of the 70 cm antenna.
|Shown here are the 70 cm radiation patterns of a single helix (left), the dual helix combo at its resonance frequency (middle) and the dual helix at its design frequency (right). As can be seen, the radiation pattern seems to change quite dramatically, but the biggest effects are close to the horizon. Overall, the helix still radiates primarily into the upper hemisphere without any major lobes or nulls.|
So in conclusion, the changes in the radiation pattern aren't really that bad. The polarization does not change by much either. So for a satellite antenna this arrangement just might work! However, with my luck, the relatively deep nulls in the horizon plane would end up pointing straight at the terrestrial repeaters which I also wanted to access... One thing I thought might decrease coupling between the antennas is mirroring one of them, i.e. changing its polarization. This, however, made the radiation pattern on 70 cm even more screwy. Also its polarization became significantly less circular.
So I ended up building the two QHAs one above the other, rather than inside each other. With just 10 cm space between the two, both radiation patterns were practically perfect.
Simulation vs. reality
I don't know about the radiation patterns in real life, but Coppens's online calculator gives very accurate dimensions. Mostly the dimensions are just a tad on the large side—to get the resonance spot on, a centimeter or two may need to be trimmed off the loops of a 2 m QHA. I don't know whether this comes from feedline interactions or other non-ideal garbage, but that has been the trend in most QHAs I've built for any frequency range.