Announcement

Collapse
No announcement yet.

Building an CANTENNA

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Building an CANTENNA

    Hi. I just build a cantenna today. But the signal is actually worst /similar result than the orgianl one come with the dvb-t dongle. Any suggestion?
    I have tested the orginal one and the cantenna outside and inside. They both are the same result.
    The photo are inside because it is more easier to take.
    Here are some photo.
    WhatsApp Image 2017-05-05 at 14.39.33.jpg
    WhatsApp Image 2017-05-05 at 14.39.33 (1).jpg

  • #2
    Originally posted by edmundhei View Post
    Hi. I just build a cantenna today. But the signal is actually worst /similar result than the orgianl one come with the dvb-t dongle. Any suggestion?
    Remember that the theoretical gains for a 1/2w dipole and a 1/4w whip (on a decent ground plane) are about 2 and 5 dBi respectively.

    A cantenna is effectively a 1/2w dipole, so the difference in performance might be explained by that alone: your cantenna has less gain and the loss due to lack of resonance in the 1/4w whip is less than the extra gain it gets from being a 1/4w whip.

    This is complex stuff as you'll see if you read the hundreds of pages long threads about antenna design. A detuned antenna doesn't necessarily perform poorly, but the length of the feeder and impedance mismatches come into play in unpredictable way.

    You can try building a co-linear (aka "co-co"; again, look at the threads). They're tricky to get right, but they're inexpensive. Just be sure to use decent quality RG6 coax with a known velocity factor.

    Bottom line: try it and use what works best.

    Comment


    • #3
      Originally posted by Strix technica View Post
      Remember that the theoretical gains for a 1/2w dipole and a 1/4w whip (on a decent ground plane) are about 2 and 5 dBi respectively.

      A cantenna is effectively a 1/2w dipole, so the difference in performance might be explained by that alone: your cantenna has less gain and the loss due to lack of resonance in the 1/4w whip is less than the extra gain it gets from being a 1/4w whip.

      This is complex stuff as you'll see if you read the hundreds of pages long threads about antenna design. A detuned antenna doesn't necessarily perform poorly, but the length of the feeder and impedance mismatches come into play in unpredictable way.

      You can try building a co-linear (aka "co-co"; again, look at the threads). They're tricky to get right, but they're inexpensive. Just be sure to use decent quality RG6 coax with a known velocity factor.

      Bottom line: try it and use what works best.
      I am also using an mcx to coax coax to f connector. Will this affect the coverage?

      DVB dongle - mcx to coax - coax to f

      Comment


      • #4
        Originally posted by edmundhei View Post
        I am also using an mcx to coax coax to f connector. Will this affect the coverage?

        DVB dongle - mcx to coax - coax to f
        Minimally, less than the difference in gain between the whip and dipole cantenna.

        Comment


        • #5
          Originally posted by edmundhei View Post
          Hi. I just build a cantenna today. But the signal is actually worst /similar result than the orgianl one come with the dvb-t dongle. Any suggestion?
          I have tested the orginal one and the cantenna outside and inside. They both are the same result.
          The photo are inside because it is more easier to take.
          Here are some photo.
          [ATTACH=CONFIG]8638[/ATTACH]
          [ATTACH=CONFIG]8639[/ATTACH]
          The ADS-B transmission is vertically polarized.
          Keep your Cantenna vertical. If it is tilted, you will get reduced reception. If it is horizontal, you will get almost zero reception.




          Originally posted by edmundhei View Post
          I am also using an mcx to coax coax to f connector. Will this affect the coverage?

          DVB dongle - mcx to coax - coax to f
          The MCX to F Pigtail has negligible effect due to its short length

          Comment


          • #6
            Originally posted by Strix technica View Post
            Remember that the theoretical gains for a 1/2w dipole and a 1/4w whip (on a decent ground plane) are about 2 and 5 dBi respectively.

            A cantenna is effectively a 1/2w dipole, so the difference in performance might be explained by that alone: your cantenna has less gain and the loss due to lack of resonance in the 1/4w whip is less than the extra gain it gets from being a 1/4w whip.
            Cantenna is NOT a 1/2w dipole.
            It is a 1/4w monopole. The 66mm dia bottom of soda can provided a ground plane disc, and soda can's cylindrical wall provides a 1/4w long decoupling sleeve (sleeve balun), and helps in suppressing common-mode currents on coax shield.

            Please see diagram below.

            Last edited by abcd567; 2017-05-05, 21:24.

            Comment


            • #7
              Originally posted by abcd567 View Post
              The ADS-B transmission is vertically polarized.
              Keep your Cantenna vertical. If it is tilted, you will get reduced reception. If it is horizontal, you will get almost zero reception.






              The MCX to F Pigtail has negligible effect due to its short length

              Hi. I have repostion the Cantenna, my max distance is 109 nm. When I drill the hole for the connector, i used the wrong drill. Therefore the can is in an angle. As show in the picture. Shall I make a new one?
              WhatsApp Image 2017-05-06 at 11.44.38.jpeg

              Comment


              • #8
                Originally posted by edmundhei View Post
                Hi. I have repostion the Cantenna, my max distance is 109 nm. When I drill the hole for the connector, i used the wrong drill. Therefore the can is in an angle. As show in the picture. Shall I make a new one?
                [ATTACH=CONFIG]8645[/ATTACH]
                Since angle of tilt is not big in this one, the improvement by making another better cantenna is not likely to be big.

                It is easy to make a Cantenna. No harm making and trying another more accurate one.

                What is the length of coax between Cantenna and DVB-T? A long coax cable will result in reduction of reception.

                Comment


                • #9
                  Originally posted by abcd567 View Post
                  Cantenna is NOT a 1/2w dipole.
                  It is a 1/4w monopole, with a ground plane disc (bottom of the soda can), and a 1/4w long decoupling sleeve (cylindrical wall of the soda can).
                  I'm not formally trained in antenna design so I claim no particular expertise with antenna physics however, from what I understand of the theory, the cantenna probably lies somewhere between, depending on the angle of incident radiation.

                  Let's see whether we agree on what 1/4w monopoles and 1/2w dipoles actually are:
                  • A 1/2w dipole is a balanced antenna (both 1/4w poles actively collect and contribute to signal), requires no ground reference and has a gain of ~2 dBi with a pattern perpendicular to the axis of the dipole. The natural impedance of a dipole is 72Ω but, being balanced, they require a 1:1 balun (into a 75Ω feeder, as is ideal with RTL SDRs) to work efficiently.
                  • A 1/4w monopole is an unbalanced antenna that requires an RF ground plane that is at least 1/4w in diameter for efficient operation. The ground plane needn't be directly coupled (and often isn't) to the feeder shield, and it reflects incident radiation to the monopole for which reason its gain is ~5 dBi with a pattern that is slightly elevated relative to the perpendicular ground plane. Because there is only one driven element (the feeder shield goes nowhere), no balun is required and its natural impedance is 36Ω, half that of a dipole. Pole-mounted 1/4w monopoles include an RF reference, usually in the form of 4 or more straight rods which are often somewhat declined relative to the horizon to improve impedance matching into 50Ω coax.


                  The cantenna's ground disc is not 1/4w wide, in fact it's probably < 1/8w for soda can of average diameter (~60mm), which means it isn't going to be a very efficient ground plane to the extent that it presents as a ground plane at all. Many soda cans have concave bottoms, though I have no idea what effect that might have.

                  The 1/4w decoupling sleeve is DC coupled to the feeder shield at the bottom of the 1/4w core. It extends 1/4w back over the feeder and ends in an open impedance with respect to the core. Because of this, radiation (of any frequency) incident on the feeder will have less of an influence than might otherwise be the case.

                  The core extends 1/4w in the other direction, also open impedance at its end. Radiation incident perpendicular to the cantenna will see two 1/4w conductors — which looks very much like a 1/2w dipole, albeit one where the feeder shield is somewhat decoupled from the active antenna. But the decoupling sleeve is not a balun because it's still connected to RF ground so, in that sense, it's more like a 1/2w dipole without a balun.

                  You're right that if the sleeve Ø ≈ coax Ø, there would be some capacitive coupling between the sleeve and the feeder, but all that does is increase the influence of the feeder shield on the active antenna. Regardless of the diameter of the sleeve, the sleeve is still at RF ground, just as a 1/4w monopole's feeder shield and RF ground plane are, so only the core is active but without the benefit of a 1/4w ground plane that a monopole would normally have (I'm still talking perpendicular radiation for now, which doesn't see the disc at all).

                  The above is also somewhat true for radiation incident at 45° to plane and core, but some of the incident radiation will also be reflected back onto the core as in the case of a monopole. The extent to which the cantenna has the characteristics of a 1/2w dipole or 1/4w monopole will vary according to angle but, as the angle to the core decreases, gain will drop because neither possess any gain on axis.

                  The remotest of sources requiring the greatest gain will be on the horizon, so the cantenna probably looks like a 1/2w dipole without balun where it matters most.

                  Leastways, that's my inexpert analysis. I could well be wrong but if I am, I'd like to know how I've misapplied theory.

                  Don't get me wrong, I'm not saying your cantenna design is a bad one, on the contrary, you proved that it outperformed your dipole, but at least part of that is probably because your dipole construction didn't appear to have a balun in it in order to make the arm coupled with the shield an active element.
                  Last edited by Strix technica; 2017-05-06, 14:51.

                  Comment


                  • #10
                    Originally posted by abcd567 View Post
                    What is the length of coax between Cantenna and DVB-T? A long coax cable will result in reduction of reception.
                    So will the type of coax. From memory, good RG6 is about -2 dB/10 metres where RG59 is about -7–-9 dB/10 metres. Even RG8, if you can get it, is something like -4 dB/10 metres.

                    Comment


                    • #11
                      Originally posted by Strix technica View Post
                      The core extends 1/4w in the other direction, also open impedance at its end. Radiation incident perpendicular to the cantenna will see two 1/4w conductors — which looks very much like a 1/2w dipole, albeit one where the feeder shield is somewhat decoupled from the active antenna. But the decoupling sleeve is not a balun because it's still connected to RF ground so, in that sense, it's more like a 1/2w dipole without a balun.
                      Correction: apparently, a sleeve can act as a wavelength-specific symmetrising balun but can also act as a radiating (receiving) element, depending on which way it's oriented as described here.

                      This explains, sort of, how sleeve baluns "symetrise" (for a specific wavelength), but it's unclear whether the shield side of the dipole is actually active or whether it merely increases output on the driven element, and therefore how it compares with a ferrite core balun in terms of radiating/receiving efficiency. NB: ferrite baluns have some insertion loss that sleeve baluns won't because of losses in the ferrite core, so it might might, er, balance out in the end.

                      As applied to the cantenna, the decoupling sleeve is oriented to be a radiator/receiver, so my original point stands: the cantenna is a 1/2w dipole (largely balanced, apparently) for perpendicular incident radiation, ie for the most distant of signals.
                      Last edited by Strix technica; 2017-05-09, 12:43.

                      Comment


                      • #12
                        A 1/2w dipole has 2.15 dBi (0dBd) gain and a 1/4w GP has 2 dBi. Not sure how you got to 5dBi ?

                        /M
                        F-ESDF1, F-ESGG1, F-ESGP1, F-ESNK1, F-ESNV2, F-ESNV3 F-ESSL4, F-ESNZ7, F-LFMN3
                        T-ESNL1, T-ESNL2, T-ESGR15
                        P-ESIA, P-ESIB, P-ESGF, P-ESSN, P-EFMA
                        mrmac (a) fastest.cc

                        Comment


                        • #13
                          Originally posted by MrMac View Post
                          A 1/2w dipole has 2.15 dBi (0dBd) gain and a 1/4w GP has 2 dBi. Not sure how you got to 5dBi ?
                          I've read conflicting accounts about this but as far as I can make out, the principle goes as follows:

                          On the one hand, a 1/4w mono has half the aperture of a 1/2w dipole and also half the impedance of a 1/2w dipole (36Ω v. 72Ω):

                          P = V^2/Z. Substituting as factors of monopole/dipole, we get P = 0.5^2/0.5 = 0.5 = -3 dBd.

                          This much is intuitive: half the length, half the gain.

                          On the other hand, this considers only radiation incident on the whip and ignores the effect of the ground, the impedance of which can vary wildly from 0Ω to hundreds or thousands. On an essentially lossless metal ground plane that is at least 1/4w in radius, all of the power incident on the ground plane (along its azimuth) will be reflected onto the whip but still driven into half the impedance:

                          So that's P = (0.5+0.5)^2/0.5 = +3 dBd or +5.15 dBi.

                          But this only works for a whip directly on a perfect ground.

                          Just as the vertical dipole's lateral gain comes at the expense of gain on axis, this extra gain has to come from somewhere; in this case, it comes from what would have been the pattern below the midpoint of the dipole. This gives it a more tightly confined and somewhat elevated pattern relative to a dipole as illustrated here. Note the caveats made in that description concerning the size of the ground plane in wavelengths.

                          The extended pattern (gain) is welcome for ADS-B, but the fact that it is inclined is not because the weakest signals come from the horizon.

                          One might conclude that, for our purposes, co-linear > 1/4w monopole > 1/2w dipole — certainly, that's how it worked out in practice for me (though my dipole didn't have a balun) — because though the 1/2w has a perpendicular pattern, it is roughly spherical in cross-section (convolved about the axis of the antenna) which means a lot of the pattern isn't where we want it, including reaching into the ground rather than toward the horizon. Co-linears (with sufficient elements) have a likewise perpendicular pattern that is flatter and elongated toward the horizon, which makes them good for ADS-B.

                          How a 1/4w mono will behave depends entirely on its ground plane size and angle to horizon. For example, some pole-mounted monopoles have their RF ground radials declined to better match with 50Ω coax, however this induces some RF into the shield and so the signal will degrade by various means. That's not to say that an impedance matching unun would necessarily do any better instead because of the insertion loss they would introduce (if a 1:1.333 matching ratio were feasible anyway).
                          Last edited by Strix technica; 2017-05-10, 13:12.

                          Comment


                          • #14
                            Cantenna is neither a pure 1/2 WL Dipole, nor a pure 1/4 WL Monopole with ground-plane.
                            It is a hybrid.


                            IMAGE 1 of 2
                            On right side is the old-timer Koaxial Antennen (Coaxial Antenna), aka Coaxial Dipole. It is a 1/2 WL Dipole as it has almost zero ground-plane due to very narrow copper tube/pipe used. The copper tube/pipe has diameter only slightly bigger than the diameter of the coaxial cable.

                            On left is the wide-bodied variant of Coaxial Dipole. Due to very large diameter of the soda can (66mm, almost 1/4 WL), the bottom of soda can makes a substantial ground-plane. As a result, unlike Coaxial Dipole, the Cantenna does not behave as a pure 1/2 WL Dipole.

                            (Translation in English in red color is added by me)




                            IMAGE 2 of 2
                            These are 3 Cantennas with different diameter can/pipe. All other dimensions are identical.
                            The worst performer was the one with smallest dia (20mm dia copper water pipe with end cap), and best performer was the one with largest dia (66mm dia standard soda can).

                            Comment


                            • #15
                              Simulation of Cantenna
                              Gain = 1.71 dBi
                              SWR (50 ohms) = 1.35

                              Can dia: 68mm
                              Can length: 70mm
                              Whip length: 68mm
                              Environment: Free Space

                              Comment

                              Working...
                              X