Thursday, January 29, 2009

Seeing the Magic

If you're like me, antenna theory has always been firmly in the realm of the occult. Oh sure, I know the basics. 1/4 wavelengths, dipoles, ground planes, building antennas following instructions in an article, etc. But when you get beyond the most basic designs I'm always mystified about WHY it works (or more likely why it doesn't work). At long last, however, I have found the wisdom. Perhaps not so much understanding but at least now I can actually see what's going on. The two-part solution is fairly simple. The first requirement is a decent antenna analyzer. VE8DW and I went in together to purchase an MFJ-259B analyzer. Ever since we got it (well, actually, since we got it back as it had to be returned to MFJ to fix a defect) it has revolutionized antenna work around the shack. All the unseen things that a simple SWR meter won't tell you are right there in front of our eyes. Did putting out all those radials under the vertical work? Obviously they did because I can see that our resistive component on 80m is about 30 ohms. How did we make out with that trap? Yup, the meter says its an open circuit at the design frequency and low resistance everywhere else. Being able to see the R and X components makes a huge difference! I can't even imagine how much time and effort I had wasted in the past doing things by trial and error.

The other piece of the puzzle is antenna modelling software. For a long time now I've wanted to be able to design and analyze antennas on the computer. After trying the demo version of
EZNEC by W7EL I was instantly hooked. $89 and a quick web download later I had the full version. The learning curve wasn't too bad and after some heavy reading both of the online manual and the antenna modelling sections of the Antenna Book and Low-Band DXing I think I've got a pretty good idea of how to do things and, more importantly, what the limitations of the underlying NEC2 software are. With EZNEC you can actually SEE the magic. Antenna patterns are displayed in 3D. You can see the RF currents flowing in individual conductors. You can sweep the antenna with a virtual antenna analyzer to calculate feedpoint impedance and design matching networks. The best part is that you don't have to learn the theory to begin understanding how stuff actually works. For example, I was playing around with a multi-element wire antenna. I've seen antennas before that had multiple driven elements connected in series by a section of ladder line so I just tried it. By adjusting the length of the virtual ladder line I could instanly tell what sort of effect it had. I was quickly able to optimize the length to produce the highest gain with a reasonable antenna pattern. Which works best, a director or a reflector? It only takes a minute to evaluate both and decide which has the advantage. How will the antenna perform in the real world? I was looking for a DXpedition antenna that would be able to slope down from a single 30 foot high support. Several of the first designs I tried had good pattern and gain when they were straight and level but tanked when I brought one end close to the ground. An other design worked better but had an unrealistic feedpoint impedance. Every trick I tried to fix the feedpoint would distort the pattern. A guy could mess around FOREVER with real wires and never find the right combination that was actually useful as more than just a dummy load. I finally found a design that doesn't get all squirrely when one end gets close to the ground, has a nice 50 ohm match, a flat SWR across 20 meters and lots of gain. One thing that I did note in my research is that actual gain predictions tend to be somewhat inflated as the program doesn't take the far field ground into account. It is useful though for comparing one design to another. The next step will be to build some of these designs and see if they perform the same in the real world as they do in the virual one.

No comments: