magnetic loop antenna
NOTA BENE: It is September 2011. After years of blindly trusting my antenna design software (while always suspecting that the aluminium foil I was using was probably too thin), I have finally done some investigation. Ordinary foil might be about 0.016mm thick. At frequencies where this would be only a small fraction of the skin depth of a non-hollow conductor, its higher resistance would increase the loop’s bandwidth and make it less efficient. Therefore, before using foil as a conductor again, I would like to be sure of adequate thickness (or layers numerous enough to overcome the deficiency). And I would like to use a magnetic loop again. Its narrow bandwidth is wonderful.
About a week ago [end of 2007?], I abandoned my magnetic loop antenna for 10 to 30 megahertz, because I was unhappy with its high SWR near 30 megahertz. I had achieved low SWR with a previous incarnation of the magloop, but that might be because I tested it with the full operational length of coaxial feeder (about 30 metres) instead of the 5 metres I used for testing this one.
The advice I would offer anyone wishing to construct an antenna like this is that it should be tested (and eventually installed) away from metallic objects; any such object within a loop diameter can give misleading results. Also, I think the 30 megahertz SWR was lower when I temporarily substituted a smaller capacitor for my home-made unit, so perhaps that is an important factor. And, if you find yourself making the coupling loop very much bigger than 1/5 of the main loop diameter, stop and find another solution. I tried to fix the SWR by enlarging the coupling loop; I came close, but by that time the coupling loop was as large as the main loop and the bandwidth of the antenna had ballooned.
One of the first magloop pages I saw was that of ON4CEQ. At about the same time, I found the antenna modelling software of G4FGQ. I have used that software (magloop4 and others) ever since, though a simpler, os-independent alternative for magloops is available from AA5TB. Information about motor-driven capacitor designs is available from LA6NCA. SM0VPO designed an alternative servo-based circuit.
Edit 2010-06-13: I recently constructed an experimental magloop from aluminium foil, along the lines of previous magloops on this page. As before, though the tuning was characteristically sharp, I noticed that the SWR 2:1 bandwidth was much wider than that suggested by ‘magloop4’. I don’t know why. Perhaps common kitchen foil is not a good material. A side-by-side comparison with a copper pipe magloop might be a good idea.
Edit 2010-09-12: (I thought I might build some aluminium foil / copper pipe magloops and compare their outputs against each other and against a simple wire antenna, but I haven’t got around to it; anyway...) I don’t remember whether I mentioned this anywhere else on this page, so I thought I’d add it here. It’s about capacitor construction. It was fun to build all the capacitors I’ve attempted. The aluminium plate capacitor worked, but my lack of precision in constructing it meant that I had to make more space between the plates in order to prevent their touching. That resulted in considerable size/weight gain. Similar gains were noticable in the capacitors I made from copper-clad epoxy board. Maintaining the spacing between the copper plates was not an issue (the epoxy board was the spacing), but the warping of the epoxy board increased the friction between the pieces, making them more difficult to turn. The linear capacitor avoided these friction issues. I never enjoyed running the noisy motor/gearbox here in suburbia, especially at night; the servo was much quieter.
(entries from blog follow)
The antenna is a magnetic loop, constructed from 4.5 metres of 500mm aluminium foil, folded once. It’s hanging from the curtain rail in the next room, the rather heavy, high-value capacitor perched precariously in the middle. A 32cm diameter coupling loop sits in the cradle of the foil loop, forcing its larger sibling into a rather uncomfortable triangle. Being unlicensed, I tuned the unit by turning up the computer speakers and adjusting the capacitor plates by hand, moving them and taking my hand away many times to achieve a reasonable signal of background noise. But still the noise level is very low. If only some stations were around.
I heard a Scottish amateur station on the 40m band earlier, where an ordinary wire gave me mostly noise.
According to ‘magloop4’, I ought to be able to tune down to 2.690 MHz, with this 2000pF capacitor. However, I could swear that I managed something lower than that earlier: 2.3?? MHz, with less than full capacitance.
Anyway, it should prove an interesting toy, as long as I don’t receive too many more shocks from the PL-259, when I try to unscrew it to compare signals. :)
Oops - I forgot to mention that the pre-amp is inline.
000042e0 | 2005-05-19T16:20:46Z | view in blog
having established a capacitance range of 11 to 190 pf and made the motion reasonably smooth, i took the skeleton of the old loop to make a new one; there is more foil in this main loop, using 300 mm kitchen foil, folded twice, for a 75mm effective (?) width; the previous loop used the same foil cut in half, before folding just once, for the same final width; the motion is generally good enough to make tuning simple; however, tuning on some bands is a little fiddly, due to the capacitor ’sticking’; (it’s really friction brought about by warped epoxy board, rather than jam or something); at 10.1 mhz, where the capacitance is highest, the sticking is quite pronounced, but tuning can be achieved in time; tuning, yes, but not a low swr; not with significant power used; not on any frequency; the old loop required a much larger coupling loop than was forecast; 90 cm perimeter, rather than the 60 cm predicted; however, having adjusted the coupling loop of this ’thicker’ loop four times, i perceive that the coupling loop would have to be larger still; perhaps quite a lot larger; a motor instead of the servo would prevent the sticking; the only other matter i can think of now is the foil tags which connect the static plates to each other; they are not very big and are currently only taped to the plate surfaces; it wouldn’t take much (hopefully) to dismantle the capacitor a bit to attach the tags to the threaded rod for a better connection; but the tags would still be small; but that should be alright; and there is another thing; i was able to compare some 14 mhz signals on the magloop and the skywire; the magloop (for whatever reasons) removed the strange noises, which seemed to be local computery/electronic stuff; but the signals were way down; 9 and below compared to 9+20 (all with pre-amp); perhaps the magloop was suffering from proximity to other conductors
00004d42 | 2007-05-24T22:12:00Z | view in blog
after disassembling, connecting, reassembling and rebalancing the capacitor plates, i remeasured the capacitance range; it is now 7 - 200 pf; which is nice; 150mm folded once is now the main loop; the coupling loop perimeter is now about half that of the main loop; 1.4 metres; i don’t know where the difference has come from; or why the original coupling loop was so far in excess of the recommended size, being 0.9m as opposed to 0.6m, which recommendation would have been by magloop4 for 10 mhz, i suppose; the only difference should be the capacitor
00004d4a | 2007-05-25T14:03:00Z | view in blog
when g4crt called out earlier, i was testing the router; i used the skywire loop; ma had returned from town with a couple of ferrite rings from maplin; i tried the rings on both patch cables and the phone cable; at one point, noticing no difference between the noise level with the cables and that without them, i concluded that the noise pertained to the dsl connection, rather than the intranet; later, using the largest capacitor available, i tuned the 10-30 mhz magnetic loop to the same frequency; it was hung on my shower room door; turning the loop, so that the null was in the direction of the router (door open 90 degrees), the speakers up and opening (from the other side) the other door, i returned to the router to test it again; this time, i noticed that emissions were significantly lower with the dsl connection running and no intranet cables attached; the addition of the cable to my room returned the noise to a high level; and then the removal of the phone cable resulted in nearly no reduction in noise, if any; i therefore conclude that the intranet is manageable, provided that the antenna is situated at a distance from it; that’s not a good conclusion, because the intranet cable ’antenna’ might be more potent than the dsl connection; one more thing; i did not think to try the ’dsl connection only’ configuration with the antenna turned through 90 degrees; that might have affected the noise level
00004de4 | 2007-05-29T19:23:00Z | view in blog
ant magloop, perimeter 10m, in bay tree, coupling loop perimeter 5m; some signs of reasonable swr (1.5:1), and the range appears to be as predicted (down to 3.5 mhz or nearly so) (not sure about the higher limit, though), but the loop picks up mostly servo circuit noise on the 5 mhz frequencies and below, which precludes tuning by ear (as was the case with the motor-driven circuit); also, the application of power appears to render the servo circuit impotent (not a problem with the motor)
00004e64 | 2007-06-23T10:43:04Z | view in blog
the noise on this channel seems too much; perhaps the larger loop is more susceptible to electrical signals; perhaps that’s because it’s slightly closer to the router circuit; perhaps because there are metal objects nearby; in any case, this does not feel like a good project; (also, that loose servo control box connection is playing up again, making this experiment more of a pain;) granted, the noise of the router on the skywire is greater; oh, and one of the top bamboos broke; probably because i gambled that the string connecting top and bottom would not be required
00004e67 | 2007-06-23T10:43:10Z | view in blog
Home-made motorised capacitor
Rotation of a capacitor built from copper-clad epoxy board and powered by a MFA Como Drills motor. Wired up to rig power supply and multimeter for capacitance measurement. Looks like minimum ~10 picofarads and maximum ~200 picofarads. This design worked alright, but the plastic gears seemed flimsy when coupled and the motor was too noisy for comfort.
00003ab3.ogv | 2007-06-27T14:04:20Z | video | alt | view in blog
anyway, about my magloop; i have photographed it in the condition in which an swr of 1:1 was achieved on 30m and 10m, using 30w (got fried again on 10m); 30m would object if the top of the coupling loop were too high; 10m would object if it were too low; perhaps the length of the capacitor was a factor in this discrepancy, a longer one being used for the 20 and 30m bands; since i did not want to do too much taping/untaping, i did not attempt any other adjustments of the coupling loop; however, i did end up remembering the twisting of the coupling loop ends, from online descriptions of abxkoppel; this had the effect of the swr 1:1 condition being maintained when i removed my hand from the capacitor knob; when i had first tried the twisting, 30m had objected, because the top of the coupling loop was too high or the coupling loop was too small or both; anyway, it’s nice not to have sacrificed the thickness of the main loop; i still wonder whether the aluminium material is also a factor in 3 metres of coupling loop wire being required; if i had some copper foil, i might be able to prove this
000054f4 | 2007-09-24T12:18:19Z | view in blog
i’ve been playing with the magloop again today; the most recent development is that i have cut the aluminium foil prototype from 3 metres to 2.8 metres, which is the perimeter of the magloop deployed previously; at nearly 30 mhz, the capacitance required is 13pf; magloop4 suggests that the maximum capacitance (for 10 mhz) would be ~ 190pf, with an efficiency at that lowest frequency of just over 50%, at 3m high over average ground; the minimum capacitance is the worry at the moment, since my preliminary fiddlings with the remains of my old linear capacitor have yielded minimum values of about 12pf; i would like more breathing space than 1 pf; smaller plates would cause less stray capacitance, and my current idea is to extend my linear capacitor design to multiple mobile plates; also, i wonder whether the so-called stators should be the mobile ones; anyway, it’s been a long day; goodnight
0000552c | 2007-09-25T20:16:19Z | view in blog
one more thing about yesterday’s endeavours with the magloop; i spent a lot of time unfurling the main loop and trying a smaller coupling loop and such things, in an attempt to explain the discrepancy between magloop4’s capacitance predictions and the values i was getting from my multimeter; as with the ~60cm perimeter loop i experimented with (before constructing my original magloop), i had forgotten about the long test leads connecting the main loop to the capacitor; with little or no test leads, results became normal; and one more ‘one more thing’; using magloop4 and my multimeter, i came to the conclusion that the equivalent diameter of my aluminium foil ‘flat tube’ is about 70mm
0000552d | 2007-09-26T06:39:47Z | view in blog
another option would be to add capacitance to the capacitor, probably by adding a small, fixed unit to the variable; anyway, i reached my limit yesterday when the coupling loop size/position issue reappeared; despite some fiddling, retaping and cutting, i was unable to achieve good swr at the top and the bottom of the range with 30 watts; (the retaping is surely unnecessary, by the way, so i’ll dig out some clothes pegs, next time;) that will be in the next few days (not something i thought i’d be saying yesterday, though i was conscious at the end of the experiment that i should not burn any bridges); i suppose i’ll resize the former and make the loop 3 metres in perimeter; i hope that will not be too much; i think there are a couple of picofarads to play with at the low end (high end of frequency range); before i dismantle the loop, however, i must check that the loop is properly connected to the capacitor and that there is continuity between capacitor plates (to validate yesterday’s testing); and then, since the lower mobile plate is now screwed to the upper, i’ll see whether the screws involved in the safety circuits are strong enough to halt the motion of the capacitor; i suspect this has already occurred, when the upper circuit came loose, but it needs to be checked; it would be good to obviate the safety circuit; it makes this unit rather complicated
000055e9 | 2007-10-04T07:16:46Z | view in blog
after a long evening with various lengths of abxkoppel yesterday and a similar amount of effort made today with faraday loops, i have found a degree of satisfaction with the following setup; main loop length 2.89m; coupling loop (abxkoppel) length 2.8m, starting nearly at the top of the main loop; this configuration will achieve swr 1:1 from bottom to top, using 30 watts; however, the length of coupling loop appears to have affected the tuning to such a degree that 10.1 mhz is now the lowest available frequency, whereas lower than ten was observed before; also, the 1:1 match at the top of the frequency range is only good as far as 30w; the swr begins to slip after that, albeit slightly; by the way, it had not occurred to me to try any of these coupling loops upside-down; the abxkoppel is fed from above; the faraday loop from above; one more experiment occurs to me; a reduction in the width of the main loop should lower the frequency range, and might also make the matching at the top end more comfortable
000055fa | 2007-10-05T21:43:02Z | view in blog
i’m about to take the ladder down; (again; i took it down before lashing the capacitor to the top of the bay tree;) before i do, let me share an important detail about the capacitor remote control; with the button on the left hand side, the capacitance goes up (frequency down) when the top of the rocker switch is pressed and down (frequency up) when the bottom of the rocker switch is pressed; i can imagine myself having forgotten that
00005694 | 2007-10-08T13:47:30Z | view in blog
i’m testing a 15 mm copper pipe magloop, length presumably 3 m; the coupling loop is a 60 cm perimeter faraday loop; required capacitance for tuning is 8 pf on the 12 metre band and 82 pf on the 30 metre band; like the ‘operational’ magloop, swr 1:1 was much easier to achieve at the lower frequencies; the bandwidth is not as wide on such frequencies as the operational magloop, but it is not as tight as the calculator (aa5tb, at the moment) predicts either (50 khz between -3 db points, using 30 watts)
000057f6 | 2007-10-11T10:33:00Z | view in blog
aa5tb says that the bandwidth should be that wide (about 60 khz) for a 1 mm diameter main conductor; i’ve just tested it and it’s true; as the diameter increases, so the bandwidth should decrease; so why does the copper pipe have the same bandwidth? perhaps the connections should be cleaner
000057f7 | 2007-10-11T11:01:29Z | view in blog
the ‘operational’ loop does not transmit very well; that could have something to do with its being hidden at the top of a bay tree, but it’s probably mostly due to the coupling loop being so large; when i was unable to get good swr at the top frequency during testing, it seems that i should have made the main loop smaller instead of the coupling loop larger; certainly, i have tried a 2.4 metre perimeter loop in the bedroom this evening and it has given good swr; perhaps i will be able to get it somewhat closer to the 3 metres i had originally intended, but that’s not too important; at least i have found a way to make progress; while the operational loop is down for adjustment, i might adjust the capacitor too; the mobile plates could do with a trim, i think, to increase the clearance between their sides and the rods that hold the static plates; i have observed that there might be flashover at maximum capacitance
00005817 | 2007-10-12T20:52:58Z | view in blog
i tried a few more tricks with the magloop today, before happening to swing the door on which it has been mounted for each testing session; the tuning changed dramatically, and i was left to surmise that the objects on the shelving behind the door could have been responsible for the poor results not just of this loop, but of the redeployment of the original loop and another loop since; i have recalled that, due to the magloop’s high sensitivity to metal objects, i carried out my initial testing on the first loop with it mounted on the plastic pole on which it was to be deployed; there is a plastic pole in the back yard; i’ll try using it tomorrow; at least my redesign/refurbishment of the capacitor was worthwhile today; i shaved the mobile plates as planned and moved the links between static plates from the sides to the top, reinforcing them with washers; the intention is that the loop ends will now connect by being squeezed beneath the bottom-most static plate, with the same pressure strengthening the inter-plate links
00005818 | 2007-10-13T21:10:00Z | view in blog
Sagging magnetic loop antenna dangling from the willow tree
As with previous magloops, this coupling loop was extended to achieve reasonable SWR, much larger in the end than the conventional 1/5 of main loop diameter. And as before, the bandwidth was not as narrow as it should be, and the results of transmitting less than spectacular. Sure, the thing was not very high off the ground, but I don’t suppose there would have been much success at any height.
0000a81e.jpg | 2014-11-01T13:53:16Z | photo | view in blog