A crossed dipole or Turnstile antenna is said to not only be omnidirectional in the plane of the two dipoles, but its pattern is roughly spherical, there's no null along the axial (boresight) direction.

There are no monopole waves; in order to couple to plane waves in the far field an antenna must have a dipole moment, but a dipole field will always have a null somewhere.

Question: How does this actually work? What's the "secret sauce" for getting a spherical radiation pattern from a couple of dipoles?

I'm wondering if there is magic involved somehow. The author of the following does mention having a tool that can make wires longer after cutting them too short though at the time of writing it had been "misplaced" ;-)

With the initial dimensions, best return loss was at 1.14 GHz, so I pruned the dipoles proportionally to move it up to 1296 MHz. As can be seen in Figure 3, I trimmed a bit too much, so it ended up tuned to 1340, but the Return Loss is still a very good 22 dB at 1296 MHz. I couldn’t find my tool for putting stuff back on, so I figured this is good enough.

cf. Quick and Cheap Omni Antenna for 1296 MHz from W1GHZ.org found at http://www.w1ghz.org

Simple Omnidirectional Antenna for 1296 MHz

  • 1
    \$\begingroup\$ (background: there's more involved designs with multiple dipoles that only achieve an approximation, and there's amateur articles out there that claim they build fantastic antennas, but no simulation can reproduce their results. I've grown extremely wary of anything that comes out of the amateur radio community, and is an antenna design without any simulation or trustworthy testing.) \$\endgroup\$ Commented Sep 13, 2020 at 10:37
  • 1
    \$\begingroup\$ @MarcusMüller despite your qualified trust of amateur radio operators, we'll likely see such simulations or measurements in the answer(s). This is a known thing and answers will explain why it works the way it does. \$\endgroup\$
    – uhoh
    Commented Sep 13, 2020 at 10:39
  • 1
    \$\begingroup\$ I don't mistrust amateur radio operators, at all! It's just that these "I've designed a great antenna articles" are essentially rumors: easier to spread than to disprove. If you come with a new design, it's your "duty" to prove it works when publishing it. And unrepeatable stuff has been written and published before, in magazines, not the unedited private web pages of a single person. \$\endgroup\$ Commented Sep 13, 2020 at 10:42
  • 1
    \$\begingroup\$ @MarcusMüller I can gather from your comments that you haven't actually read it yet. \$\endgroup\$
    – uhoh
    Commented Sep 13, 2020 at 10:48
  • 2
    \$\begingroup\$ @MarcusMüller and others will vote on that answer, and other answers will be posted, and in a day or so we'll understand why this works. \$\endgroup\$
    – uhoh
    Commented Sep 13, 2020 at 10:58

3 Answers 3


The author of that paper themselves didn't even fully read the survey paper they're citing: The antenna they replicated was a circularly polarized one, not an omnidirectional one.

When you think about symmetry and what circular polarization means, you'll quickly realize you can't have omnidirectional emission and consistent circular polarization on a sphere; you need the polarization to change, which means that the field can't have the same magnitude in all polarization planes all over the sphere.

In fact, that's a corollary to the pretty intuitive "Hairy Ball Theorem" by Poincaré (the same person of Poincaré Sphere fame, the thing you've learned when studying mixed polarizations).

So, this literally can't be an omnidirectional antenna.

Hope this is authorative enough. For general commentary on how to approach papers that claim great things about antennas, but don't offer proof, see my other answer.

By the way, here's the figure from the paper 53 that they say they can't access:

Don't fuck with science

More than 30 dB attenuation at 135° doesn't look like an omni antenna, does it?

The authors themselves don't claim this antenna to be three dimensional omni. Quite the contrary

A three-dimensional (3-D) plot of the far-field resembles a doughnut with an omnidirectional distribution in the plane perpendicular to the dipole axis and a figure eight pattern in the orthogonal cut.

TL;DR: doesn't work.

  • 1
    \$\begingroup\$ Hairy ball theorem will show this works if you'd just reconsider your assumptions You shouldn't post two answers in rapid succession, this isn't appropriate SE behavior and I've flagged this post accordingly. \$\endgroup\$
    – uhoh
    Commented Sep 13, 2020 at 11:11
  • 1
    \$\begingroup\$ You're hiding the LP power, only showing the CP power. Why not show both, or simply total power? These days receiver front-ends can handle full quadrature. \$\endgroup\$
    – uhoh
    Commented Sep 13, 2020 at 11:14
  • 1
    \$\begingroup\$ @uhoh I don't really know what your deal is. The people that invented the antenna that you're referring to literally themselves say "this is not 3D omnidirectional", see what I quoted. You yourself notice how any dipole has a null. Every single resource but your paper without measurements nor simulations says this doesn't work. Can we put it to rest already? \$\endgroup\$ Commented Sep 13, 2020 at 11:31
  • 2
    \$\begingroup\$ There's nothing to wait for, in all honesty. I wouldn't have posted this answer if it weren't right, to the best of my knowledge, the survey paper that your paper cites, and the original paper. I get the feeling you'd only accept an answer that confirms your bias. \$\endgroup\$ Commented Sep 13, 2020 at 11:34
  • 1
    \$\begingroup\$ You might want to read that correctly: Not every single one of them produces has all of these properties... Some do one or the other. The one we talk about is not among those that have the anisotropic property. \$\endgroup\$ Commented Sep 13, 2020 at 11:54

I don't see a reason why this would yield an approximately spheric wave front.

Unless you have a simulation or measurement that proves working of this, you'll have to assume it doesn't work. It's the scientific way – assume the null hypothesis, unless evidence or theory prove the opposite.

  • \$\begingroup\$ No, "assume it doesn't work" is not the scientific way. One should assume that one simply doesn't know, or that one has made some implicit assumption that can't be justified. \$\endgroup\$
    – uhoh
    Commented Sep 13, 2020 at 10:46
  • 1
    \$\begingroup\$ I contradict: When I have a new thing that I think could work, then I'll have to assume when considering all my observations with my thing that it doesn't work, unless evidence for the contrary is overwhelming. Assuming a "I don't know, so 50/50, maybe?" is not scientific; again, I could simply claim my dipole becomes an anisotropic radiator as soon as I paint flowers on it. Do I get to ask a question "why does flower painting lead to a 3D-omni antenna", too? \$\endgroup\$ Commented Sep 13, 2020 at 10:48
  • \$\begingroup\$ "I don't see a reason why" is because you haven't looked. How are the two dipoles driven? Is it how you've implicitly assumed when you wrote this answer, or did you miss something? Your "it can't work" reaction is similar to my first reaction which led me to post this. I think when you read others' answers you'll feel better. Right now you've just written an answer that says "I don't know" which isn't really a Stack Exchange answer. \$\endgroup\$
    – uhoh
    Commented Sep 13, 2020 at 10:49
  • 1
    \$\begingroup\$ I did not write that. I said "unless evidence to the contrary, and there is none, assume it doesn't work". That's the opposite of "I don't know", I said "I know this doesn't work, and I've found nothing that makes me assume that's not the case". \$\endgroup\$ Commented Sep 13, 2020 at 10:56
  • 1
    \$\begingroup\$ @VoltageSpike no, I intentionally kept the "general approach" from the "specific approach" answer, as they address different aspects of the question. \$\endgroup\$ Commented Sep 13, 2020 at 20:46

Front matter:

I'd thought this would have a quick answer I could accept, but instead two other answers (1, 2) by the same user are of the "it can't work" variety, but they are hard to parse due to a deep and rich array of caveats and distinctions.

So instead I decided to counter all this negativity with a simple explanation just to start things on the right track. I'd prefer to accept a more thorough, authoritative answer, so please feel free to post something more thorough and/or authoritative!

If nobody posts an answer with the math for the 3D far-field power of crossed dipoles driven at 90° in a day or so I'll add it here as well.

Short answer:

How can a crossed dipole (turnstile) antenna be omnidirectional in 3D?


Question: How does this actually work? What's the "secret sauce" for getting a spherical radiation pattern from a couple of dipoles?

The "secret sauce is that the two diples are orthogonal in both the time and space domains.

In plane, it looks like one dipole no matter which direction you look. At 0°, 90°... etc. you see only one dipole because the other is end-on; at 45°, 135°... you see two, but because they are 90 degrees out of phase there is no angle in-plane where there is a null in power.

Out of plane it produces circular polarization. At the poles (axial, boresight direction) it's a pure circular polarization emitter.

There's more at Ta et al. 2015 Crossed Dipole Antennas: A review (available here as well). It's first reference is G. Brown, “The turnstile,” Electron., vol. 9, pp. 14–17, Apr. 1936. which I can't find a link to.

It has been known for quite some time that these can provide power roughly isotropically. One challenge is of course the mast and/or feed line. That's solved if the transmitter and power source are both right at the feed point!

In 1958 a turnstile antenna was used on the Vanguard-1 satellite one of the first satellites orbited by the US in answer to Sputnik-1 because you'd like to receive its precious data no matter how it tumbled and turned in space! These days satellites can have attitude control systems, but back in the 1950's satellites could do precious little more than say beep!

From this answer to Why did Sputnik 1 have four antennas?

original artwork on the circa-1958 plastic model of a Vanguard series satellite

above: Image from here - original artwork on the circa-1958 plastic model of a Vanguard series satellite - antenna pattern is more easily visualize in this representation. More about Vanguard models can be found in Space Technology from Six Decades Ago: 1:5 Scale Hawk Project Vanguard Satellite.

From Vanguard 2; NSSDCA/COSPAR ID: 1959-001A:

Four 30-inch spring-loaded metal rods were folded along the equator of the sphere and would protrude radially outward when deployed, acting as a turnstile antenna. Radio communication was provided by a 1 W, 108.03 MHz telemetry transmitter triggered by the ground station and a 10 mW, 108 MHz Minitrack beacon transmitter that sent a continuous signal for tracking purposes.

  • 4
    \$\begingroup\$ that sadly doesn't answer your own question on how it works, but just, again, claims it works. \$\endgroup\$ Commented Sep 13, 2020 at 20:47
  • \$\begingroup\$ Also, again, the photo that you show in your question is not that of an isotropic radiator, which the survey paper doesn't claim, but the paper w1ghz paper simply assumes. \$\endgroup\$ Commented Sep 13, 2020 at 21:00
  • \$\begingroup\$ @MarcusMüller Cheer up, no need for sadness Like it says, "If nobody posts an answer with the math for the 3D far-field power of crossed dipoles driven at 90° in a day or so I'll add it here as well." I'll include the Hairy ball theorem as part of it as well. \$\endgroup\$
    – uhoh
    Commented Sep 13, 2020 at 22:03
  • \$\begingroup\$ cool. and how does that change the fact that the antenna your question specifically asks about isn't a turnstile designed for omnidirectionality at all? \$\endgroup\$ Commented Sep 13, 2020 at 22:06
  • 1
    \$\begingroup\$ Been nearly 4 months now and the math still hasn't been added to this answer....... \$\endgroup\$
    – MCG
    Commented Jan 20, 2021 at 15:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.