It’s fun to take a minimum of equipment out to a high place, lob a half-wave long wire antenna up into a tree, or support it with a fibreglass pole, and off you go – but such a long wire typically has a high (2.5kΩ to 3kΩ) impedance. To make it work, this needs transforming down to match the 50Ω output of a typical transceiver. This is done using a 49:1 autotransformer, built on a small toroid.

This post details my construction of the end-fed half-wave autotransformer and antenna for the 10m, 20m and 40m bands.

The antenna is very effective, can be erected in a variety of ways, exhibits low noise, doesn’t cause any RFI in the shack (use a choke balun in addition), places the feedpoint very near to your transceiver and it’s pretty easy to build.


There are many other articles out there on this subject:

There are also several other pictures of builds I’ve used to assist in this project, particularly from Ian MW0IAN and Michael G0POT. Thanks!


I hope to add some detail and pictures for the terrified constructor in this article, but the above are my source materials. This antenna and its autotransformer are a great project for those new to construction – the toroid winding isn’t hard.

One problem I find in many projects is that they go from diagram to finished construction very quickly with insufficient detail. It’s a bit like “How to Draw an Owl”….

How To Draw An Owl

So I wanted to document what I’d done, in the hope it might be useful to others, especially if you’ve not wound toroids before. The key point is that the count of ‘turns’ relates to the number of times the wire goes through the inside of the toroid.

Parts and materials

I’m building this autotransformer on a FT-140-43 toroid, purchased as a pack of two on eBay from Spectrum Communications, for £9.50 incl P&P.

I first wrapped this in plumber’s PTFE joint sealing tape.

I’m using Maplin 0.9mm 20SWG enamelled copper wire, code YN82D. £8.49 for a small reel (the picture on their page is misleading; it’s not as small as indicated). The source articles used 1.0mm wire; this will allow higher power to be transmitted – I’m mostly QRP, and without calculating it, 0.9mm will be fine for this.

For connectors, I’m using a BNC socket, £2.49 for Maplin’s code HH18U, and a small terminal post, again £2.49 for Maplin’s code FD69A. A small black plastic box, for £3.39 code KC91Y houses the completed autotransformer. This box is a bit ‘tight’, but it does all fit. You’ll also need a 100pF 3kV ceramic capacitor, available for £0.131 each (min order, pack of 10) from RS Components.

Building the Autotransformer

Wire lengths for the toroid winding given here come from reading other articles which describe builds on the larger FT-240-43 toroid. After I’d wound on the smaller toroid as shown here, I had some waste wire that required trimming – you could start with shorter pieces of wire, and I’ve estimated possible lengths for this [in square brackets]. However please note I haven’t tested those lengths.

Measure two lengths of wire, 1.0m, and 22cm. [For the FT-140-43 toroid, and a small box as given above, you could probably use lengths of 80cm and 18cm.] Strip the enamel from the ends for about 2-3mm, clean with sandpaper and tin with solder. Solder the two right-hand ends together.

Solder the right-hand ends together

Holding the soldered end in pliers, twist the two lengths tightly together for 13cm [for the FT-140-43, 9cm may suffice], to form the bifilar winding. ‘Bifilar’ meaning ‘two filaments’, I presume. The remaining length of the shorter piece of wire (at the end of the bifilar winding) being about 6-7cm long. Let’s call this length ‘the tail’.

Now to wind the bifilar part onto the toroid.

Pass the bifilar part through the toroid from the back, so that the ‘tail’ passes under the bottom edge, and the start of the bifilar part is entirely running through the inner of the toroid.

Starting the Windings

Wind the bifilar part back round the top and outer edge of the toroid, and back through the hole. Bend it again over the top. You should have two bifilar passes through the toroid. The end of the bifilar winding (that’s the soldered-together end) will be connected to the outer/shield/chassis connector of the BNC socket. The end of the tail will be connected to the inner/centre connector of the BNC. I’ll get back to the connections later.

The Bifilar Winding

Now, there’s the rest of the toroid to wind.

Wind the long single strand of wire through the toroid so that you have six passes on the inside.

The first part of the rest of the winding

Then pass the wire over the top, and over to the other side, going under the far edge. Loop over the top, and go back through the toroid so that on the far side, you have seven passes, with the remaining end (‘the antenna end’) going under and out.

Ensure that the windings are spaced as evenly as you can.

The final winding

Now tin the connections of the BNC socket and terminal post, and fit them in the box.

Carefully trim the three wires coming from the toroid to fit; scrape off enamel (might need to unwind the end of the bifilar winding a couple of mm to do this properly), clean and tin the ends.

The toroid, trimmed, with box

Solder the bifilar winding to the outer of the BNC; the ‘tail’ at the end of the bifilar winding to the centre pin of the BNC; the ‘antenna end’ to the terminal post.

Solder a 100pF 3kV ceramic capacitor between outer and centre connections of the BNC. Use a hot glue gun to fix the toroid to the inside of the box (not shown).

The assembled autotransformer

Put the top on the box, and there you have it.

The completed autotransformer

Building the antenna

There’s a long antenna wire attached to the autotransformer terminal post, followed by a coil and a further bit of wire. I used fairly rugged plastic-insulated wire; the outer diameter is 2mm, bought years ago on a 100m reel, probably from Maplin.

auto |         Length L1           Coil in uH    Length L2  (fibre-
xfor-|=----------------------------|\\\\\\\\\|------------|  glass
mer  |                                                    |   pole)
-----+                                                    |

For 80m+40m+20m+15m+10m, L1 is 20.35m, the coil impedance is 110μH, and L2 is 2.39m.

For 40m+20m+10m, L1 is 10.1m, the coil is 34μH, and L2 is 1.85m. I built this variant.

These lengths are taken from the diagrams in the PD7MAA article. Also see PA3FRP’s notes.

I had a length of plastic tube with an outer diameter of 15mm; I wound 141 turns of the Maplin 0.9mm 20SWG copper wire on this, and measured 34μH. This was very much a trial and error exercise; I had no idea how many turns it would take, so the initial length was extended by soldering more on, several times.


I attached the far end of the wire to the top of my 7m SOTABeams fibreglass pole, and fed the wire out so that the autotransformer was on the ground – imagine a right-angled triangle with a bowed hypotenuse; I also lifted the autotransformer up onto a small table.

Measurements with an antenna analyser show very favourable SWR readings for the three bands I’m interested in:

  • At 7.305MHz, SWR is 1.21:1
  • At 14.318MHz, SWR is 1.14:1
  • At 27.780MHz, SWR is 1.15:1

With a bit of adjustment, and appropriate positioning – it does seem sensitive to its location above ground (this is explained in the test article by John Huggins) – I think it’d be fine to transmit into this. I’d ideally like to tune it a little more, so I can use it on the exact frequencies I frequent without an ATU.


This is part 2 of my series on building a cheap loop antenna. If you haven’t read part one, it’s here. In this second part, I’ll cover the building of the tuning capacitor.

Obligatory Safety Notice

This article describes machinery and tools, workshop practice and construction that if not performed carefully and with the appropriate protective equipment (goggles, etc.) – could lead to serious injury. Please take care. I won’t be held responsible for your accidents!

Tools Required

In addition to keeping the material costs of the antenna as low as possible, it should also be buildable with a minimum of specialist tools. The following were used in constructing the capacitor:

  • Superfine permanent marker (Staedtler Lumocolor S)
  • Compass with universal adapter to take these pens (Staedtler Noris Club 550 01)
  • Metal / plastics drill (Dremel 3000)
  • Drill bit 3.2mm (Dremel)
  • Drill press (Dremel 220 workstation)
  • Thin circular metal file, flat metal file (Stanley warding file set)
  • 8mm spanners for M5 nuts
  • Junior hacksaw
  • Fret/coping saw (Stanley fatmax)
  • Centre punch
  • Two C-clamps for holding things whilst cutting (a vice would be far better)
  • Straight aviation tin snips


Item Cost each Total cost
9mm thick IKEA Legitim Chopping board £1.00 £1.00
Stainless steel washer M5 100 pack (Toolstation 79167), x2 £0.96 £1.92
Stainless steel Threaded Bar M5 (Toolstation 61139) £1.78 £1.78
Stainless steel Lock Nut M5 100 pack (Toolstation 64158) £1.58 £1.58
Stainless steel Spring Washer M5 100 pack (Toolstation 92679) £0.97 £0.97
0.9mm thick Aluminium (grade 1050) plate sheet, 300mm x 200mm, protective coating on one side £3.80 £3.80

Well, there you go, I wanted CHEAP, and you can’t get much better than £11.05!

I chose M5 threaded rod, nuts and washers as it would be mechanically substantial, and not present too much resistance to RF. I considered M3, but it would introduce greater loss, despite being easier to drill for. The Dremel drill I use is excellent, but does not drill holes with a diameter greater than 3.2mm. I thought of drilling this size, then increasing the hole size with a burr, but this did not work out well, so I reverted to filing out to 5mm. Another reason for going with M5 hardware is that Toolstation in Tunbridge Wells stock M5, but not M3, and they’re really quite reasonable as the table above shows.

I initially considered Aluminium Warehouse for the plate, but their shipping cost was £15, so scratch that! I mentioned this to my wife, who suggested Amazon…. which turned out to be an excellent idea, as Hardware Outlet sell it for £3.80 with free postage. Nice.


The diagram below shows the design of the rotor/stator vane plate (six of which are cut from the aluminium sheet), and end plate (two of which are cut from the 9mm thick cutting board). This layout comes from the article Magnetic Loop Antennas, by Tony ON4CEQ. The rotor and stator are cut from a single piece of aluminium with a minimum of tricky cuts. Other articles have separated them with Dremel cutting disks, which I had bought, intending to try – but managed to cut out Tony’s design easily.


In hindsight, it would have been better to make the end plate a few mm taller, as the rotor is exactly the same diameter as the end plate’s height (8cm), and if not raised, touches the desk as I rotate the finished capacitor.

I cut the end plates out using the above fret/coping saw, along with some extra pieces used to clamp things in my drill press. The holes for the threaded rod, forming the stator arms and rotor were drilled out with the Dremel’s 3.2mm drill bit then filed out to 5mm diameter.

I marked out the aluminium on its protective layer, using the microfine permanent marker and compass as shown. I left a 5mm gap between each plate (see detail below), thinking that the extra metal would prevent the drill from tearing the thin aluminium when drilling or increasing the size of the holes from 3.2mm to 5mm. I need not have done this; increasing the size of the holes was best done with a file, which was tedious. The vane panels could have been laid out directly next to each other; cutting each panel out was easy with the junior hacksaw. Each hole was centre-punched, which made drilling far easier.

Drilling was straightforward, although the Dremel workstation tends to allow the drill to move slightly. Centre-punching is essential.

The stator and rotor are cut from the panel easily; the diagonal cuts toward the centre are easily done with a junior hacksaw; the short cuts near the rotor axle, joining the diagonal cuts are done with the fret saw shown above (which is probably only supposed to cut wood; it handles 0.9mm aluminium well too).

Tin snips are then used to cut the unwanted parts of the rotor blade away, to yield its circular edges.

Burrs are filed off, edges and corners smoothed to remove any sharp points that could a) cut you and b) offer a point of voltage breakdown and arcing.

I then cut four 85mm lengths of M5 threaded rod for the four corner stator attachments, and a single 105mm length for the rotor. I then assembled the frame with the plastic, rod, spring washers, washers and nuts.

I assembled the rotor on the longer length of threaded rod. The washers I’m using to separate the vanes are exactly 1mm thick. Initially I used five of them, to make absolutely sure I’d have the inter-vane gap I’d need, but found the capacitance to be too low. So I rebuilt with four washers between vanes, and achieved a capacitance much closer to that needed.

The stators were assembled onto the frame, and roughly aligned with each other.

The rotor then fits snugly inside the stators and its axle fits in the centre hole. It’s hard to make out the end plate due to the perspective here, but it’s there.

Flat washers are fixed on either side of the rotor/end-plate mountings, in lieu of bearings. The top plastic end plate is then added, and everything tightened up. There’s a certain amount of adjustment of the stators and rotor to ensure that the rotor rotates freely within the stator plates, and that there’s an equal gap between rotor and stators. This adjustment is mostly done via the nuts mounting the stators, but there might also need to be gentle bending of the various plates.


After construction, I measured the capacitance, and read 0.006nF (6pF) for completely unmeshed, increasing to 0.043nF (43pF) for completely meshed – close to the upper value I need – 46.311pF from the first article in this series.


There’s a bit more mechanical tweaking to do, but that’s the risky part of the magnetic loop project done. Next time, I’ll assemble the rest of the loop and think about manual/automatic tuning.

Here’s a short clip of the capacitor being tuned and measured…..

Stay tuned!

73 de Matt M0CUV

I’ve been wanting to try a different aerial for 20m for some time. I currently either use a dipole in the loft, or a temporary SOTAbeams linked dipole on a 7m fibreglass mast. The magnetic loop antenna appeals because it’s compact, and if designed/built for efficiency, gives excellent performance.

They’re also expensive. One well-known UK dealer is selling the Ciro Mazzoni baby loop for £1299; another dealer sells the MFJ-1782X for £439; the Chameleon CHA F-Loop is available from a third dealer for £399. All are far, far outside my budget.

It’s not called amateur radio for nothing, and what I describe in this series of articles is truly amateur. I’m going collate my research sources, and try building a magnetic loop antenna (properly called a small transmitting loop antenna) using bits from my junk box, making good use of the things that I find, things that the everyday folk leave behind. I don’t have all the parts, but whatever I need to buy must not break the bank. Similarly, tools: I’ll need some, but they must be cheap 🙂