A (Relatively) Simple H.F. Polarimeter

By

Eric P. Nichols, KL7AJ

Here's a more advanced project than some of the others I have written about, but still within the capability of any competent home constructor. It's certainly different from anything you've ever built, but not any harder.

With this instrument, you will be able to prove to yourself something that's been known by ionospheric scientists for many decades, but still denied (or ignored) by most of the amateur radio community: that ALL ionospheric H.F. signals are circularly polarized.

About a decade ago, I wrote an article in QEX called “How the Ionosphere REALLY Works” and had a clever subtitle “Gimme an X, Gimme an O, What's that spell? RADIO!”

Well, it's just as true now as when I wrote it, but you will be hard pressed to find ANY amateur literature that mentions X and O mode propagation. You will be able to settle the question once and for all in your very own home, with my simple polarimeter, which will easily show you whether an incoming signal is clockwise circular (O-mode in the northern hemisphere) or counterclockwise circular (X-mode in the northern hemisphere). To keep things simple, this device will be fixed tuned to WWV, either, 10 or 15 MHZ. (If you like, you can modify it to be tunable, as long as you aware of the calibration requirements in doing so).

The point here is not to go into great detail on the actual plasma physics involved in the polarization conversion of the ionosphere. It's fairly complicated and not necessary to understanding and taking advantage of the end results here on the ground. We simply want to show that it not only exists, but that it always exists. You will stun and amaze yourself! You will also see the fallacy of some amateur rhetoric that the polarization of H.F. is somehow scrambled along the path. No, it is not scrambled, it is cleanly divided into two different propagation paths, one being X-mode and one being O-mode. What IS being scrambled is your knowledge of which of the two signals you're actually receiving! This device will tell you!

Because it's a little awkward to send attachments on this particular forum, I will instead direct you to another site where you can download schematics and block diagrams. However, using my finely-honed command of the English language, you should be fairly able to figure out the basics by description alone. (Ahem).

The most expensive part of this project is a good 20-30 MHz oscilloscope, with a well-matched X-Y plot capability. You may have to beg borrow or steal a scope to do this, but you can also make this a club project, and split the cost, if you don't have a good scope on hand. The bandwidth of the scope isn't too demanding, but you NEED to be able to get a nice straight diagonal line with a 15MHZ signal simultaneously applied to the X and Y channels. In fact, this test is the first step of the overall calibration process.

The Antenna

The “ears” of this project is a simple circular polarized (CPOL) crossed dipole. If you have the room to put up a full-sized 15 (or 10) MHZ dipole, more power to you, but it's certainly not necessary. My own version uses four 4-foot Fiberglas CB whips). In fact, it's MUCH easier to get the phasing just right if the antennas are NOT anywhere near resonant, as long as both of them are identical. Even more wonderful, we don't have to do anything tricky to get the 90 degree phase shift normally necessary for a CPOL antenna...the X-Y oscilloscope does that all for you!

As we work our way through the project, keep in mind that nothing is critical except the MATCHING between the two channels. I want to avoid calling them the “vertical” and “horizontal” channels, because we're actually going to erect our antenna with both dipoles at 45 degrees to the horizon (but 90 degrees to each other). This is so that neither dipole favors vertical polarization; it will allow us to “calibrate out” any error. So...physically speaking, your antenna will be an “X” instead of a cross. Another reason we do this is that we will bring our two coaxial cables directly down from the feedpoint, any unintentional coupling to the antennas will be equal. We do want to use two matched high-isolation choke baluns for this, however. Since there is minuscule power, you can use some really high permeability beads here, achieving lots of decoupling without having to worry about power loss.

The Receivers

This is such a clever design; I wish I'd thought of it myself. However, someone beat me to it by a few decades, for use in remote frequency monitors for broadcast stations. What we need is a receiver that contributes no frequency change during the process. Actually we need TWO of them. YIKES! How can one pull THAT off?

Real simple. We build a superhet receiver, using a crystal oscillator to convert the signal to a low, highly selective I.F., and then using the VERY SAME local oscillator to convert that I.F. back up to where we were! VOILA! Any error in the oscillator is cancelled out. In our particular application, we will use the same local oscillator for BOTH receivers.

If we're super careful about matching our receivers, the phase difference going through the I.F. amplifiers should be minimal, but we have made provisions for final tweaking of phase during the calibration process.

For this receiver design (which is also negotiable) I've chosen a 100KHz I.F. just because I just happened to have a couple of 100 KC calibration crystals lying around from the Dark Ages. These things make wonderful, sharp, single pole I.F. filters. The only downside of this is that I had to build a 15.1 MHz local oscillator from scratch. If you pick an L.O. frequency that happens to be in a ham band, you can use your H.F. transmitter (with appropriate power reduction and/or padding) as your local oscillator. For the up and downconverters, I use very cheap MiniCircuits labs DBMs (doubly balanced mixers). These are passive devices that can handle quite a bit of drive. If you run your transmitter through 4:1 splitter (another MiniCircuits device, in my particular case) you probably don't have to worry about torching the things if you inadvertently pump a watt or so out of your rig.

Actually, ANY I.F. that gives you adequate gain and selectivity for your particular location will work fine. You just need to be able to distinguish WWV from the cacophony, which isn't too hard, most of the time. Again, it's not the specs; it's the MATCH that matters. Whatever you do, do it EXACTLY twice. And of course that includes the lengths of your transmission lines from your antenna halves.

I.Q. Test

Now, with just what we have assembled up to this point, we can demonstrate the circularity of incoming signals. In other words, if we were to lash our antenna, our receiver and our oscilloscope together (one receiver into each of the scope's inputs) we would see an ellipse on the oscilloscope of the WWV signal, demonstrating non-linear polarization. However, we are still missing one ingredient, though. We have not addressed the phase ambiguity. In other words, we know it's circular, but we don't know if it's right-hand or left-hand. This requires one more 90 degree phase shift in the mix. After going through all that excruciating effort to make sure everything matches, phase wise, we're going to screw it all up with an I.Q. switch. (I.Q. in this case is not Intelligence Quotient, but In phase and Quadrature). We need to intentionally shift the phases of our signals by precisely 90 degrees. Over a wide range of frequencies, this is rather difficult; at ONE frequency, it's extraordinarily simple. Let's do the latter, shall we?

A simple R.C network does the trick for us. We also want to be able to flop this thing by 180 degrees so we can select X or O modes. (See schematic).

Now, at this time, I'd like to address your attention to the WWV website:

http://tf.nist.gov/timefreq/stations/wwv.html

There you'll find all kinds of great information on their antenna system(s). You'll see they use ideal vertical radiators, and very tightly controlled ERP. I recommend you peruse the entire web site.

Now, the following discussion assumes you're outside the WWV groundwave signal. If not, you will want to use WWVH. Or, if you're in Hawaii, you definitely want to use the “mainland” signal.

Tweaks and Twiddles

We want to perform the final calibration one of two ways. We either want to pick a time when propagation from WWV totally sucks. Or we can use a couple of matched high-loss attenuators. We recommend the latter.

We will also need a 15 MHZ signal generator of some sort, and a reference vertical antenna directly in front of our dipole some distance. This antenna can be any short whip, but it should be well isolated from its transmission line. Even BETTER though, is to build a little battery powered oscillator in a box with a whip on top, thus entirely eliminating the possibility of transmission line radiation.

Proceed as follows. Insert enough R.F. attenuation so that you can no longer see WWV on your scope. Turn on your test transmitter. Adjust the gain of your receivers for a match, as indicated by a 45 degree line across the scope screen. Flip the “X-O” switch. The 45 degree line should lean the other way (also by 45 degrees). In both cases, you should have a nice straight line. If you see an ellipse, you need to adjust the IQ phasing. You may have to alternate between the phasing and channel gains. (Also, be sure your scope itself is capable! When in doubt, plug your generator directly into your scope ports with a simple splitter.

If all is A-OK, remove your attenuators and your test generator. You should now see WWV in all its glory. It may be a circle, it may be an ellipse; it may be an ellipse that ROLLS slowly like a spun coin. The axial ratio of the ellipse may change.

Now, flip your X-O switch. You should still see an ellipse of some sort but the amplitude will likely be different. MOST of the time, the larger signal will be the O-mode. If you consistently see one mode stronger than the other, you can with reasonable confidence, label that the O-mode on your switch.

Here's one further embellishment you might try. It will require the addition of one more phase flipper switch, so you can look at both the X and O mode signals simultaneously with a dual-trace oscilloscope. (You'll lose the ellipticity function, but that's okay) If you observe WWV's time clicks (assuming you have enough I.F. bandwidth to SEE them!) you will see a noticeable delay of the X mode click after the 0-mode click. This is because the critical height of the X-mode signal is always higher than the O-mode signal. (This is also a good confirmation of your X-O switch label).

I really encourage anyone who can build this to do so. It will open up entire new realms of understanding for you, whether you're new to H.F. or an old timer.

73,

Eric P. Nichols, KL7AJ

(Feel free to email me at kl7aj@acsalaska.net)