EFHW matching unit

My first couple of SOTA activations used a SOTABeams Band Hopper III linked dipole. This worked well, but with two wires, a 10m coax and an extra guy, it took a bit of time to set up and it was easy to get all the wires tangled if the wind got up.  I started reading about alternatives and decided an End-Fed Half Wave (EFHW) looked like a good option.20131221_204618The benefits are:

  • One matching unit could feed a range of elements that were just simple wires (or could be multi-band with links or traps)
  • The matching unit was low to the ground in an inverted V configuration and hence could be fed with a short coax
  • The antenna could be used as either an inverted V or a half wave vertical on the higher bands

The challenge with the EFHW is feeding it as the end is a high impedance point (apparently between 2000 – 6000 ohms depending on the surrounding environment). I explored a range of matching unit designs. They seem to fall into two main categories:

  • A transformer with a tuned circuit for the secondary
  • A simple impedance transformer with a fixed capacitor across the primary

The first model is well described by AA5TB (Steve Yates) along with several example designs. The main limitation of this design is the need to tune the tank circuit for the operating frequency in use. Designing a single matching unit that will cover a broad frequency range is challenging. It also requires a tuning capacitor that is capable of quite high voltages that appear at the end of a half wave antenna. The typical construction of such a matcher uses a Polyvaricon capacitor which limits the power to QRP levels only (noting the high voltages present at the end of the antenna).

The second design is that used by PAR EndFed antennas. It is a broadband match design with no tuning required and this appealed to me. The trade off is that you can’t get the SWR down to 1:1 except by luck, but it is possible to get a SWR below 2:1 for most frequencies of interest with a typical half wave wire. I found a couple of good references to such designs by Dutch hams here and here. They include details of creating multi-band EFHW antennas using this matching unit.

I followed the Dutch designs quite closely, but experimented a bit to see if the 150pF capacitor is the right size. In my case, I used a FT140-43 toroid (wound with a layer of Teflon tape for protection of the toroid and enamel coating) with a 3:24 turns ratio wound using 1 mm enameled copper wire. Here’s a photo of the competed unit:

My EFHW match box. 3:24 turns ration on a FT140-43 toroid with a 150pF capacitor across the input.

My EFHW match box. 3:24 turns ratio on a FT140-43 toroid with a 150pF capacitor across the input.

I experimented by attaching a load resistor between the output and ground for a range of possible antenna impedances from 2200 to 6800 ohms. Here’s a series of SWR plots across the HF bands from 80m to 10m with first no capacitor and then a range of capacitor values as detailed on the images.

Image Image Image Image

As you can see, the capacitor is clearly necessary in this design and 100 pF allows a better than 2:1 match across most HF bands and impedance values. Whilst 150 pF provides a better match at some frequencies, the 100 pF seems to provide a more even performance over a wider range of loads and frequencies. Values above 150 pF resulted in much worse plots and weren’t recorded.

It would be possible to replace the fixed capacitor with a variable capacitor (such as a Polyvaricon). This may allow the SWR to be reduced through tuning for the specific frequency and antenna impedance at the time. I haven’t tried this as the goal was to get adequate performance with no tuning needed. The advantage of this design with a variable capacitor is that the voltage across the capacitor is much lower than if the capacitor is on the high impedance side of the transformer. Hence a Polyvaricon in this position would likely be fine with 50W or possibly even 100W (with the sections wired in series).

You’ll see that there is no counterpoise with this matcher and the coax to the rig (and the rig and operator) performs that function. If this were a permanent high power set up, it may be necessary to add a counterpoise of some kind, but the Dutch amateurs seem to indicate that this does not appear to be necessary in practice even at high power.

The 3:24 turns ratio is better for the lower frequency bands and if I was building for the upper bands, I would probably have tried a 2:16 ratio instead. As you can see this design should be OK for 80m to 12m bands (which is what I wanted for SOTA activations).

In terms of power handling, I expect that the FT140-43 should be good to at least 50W. Above that, the core will likely saturate and losses will grow rapidly. I haven’t tested the loss of the unit, but hope to get round to that at some point. In practice it seems to work well with no problems making contacts with my 20W SSB transceiver. I’ve also driven it with my IC7000 in a field day contest (JMMFD) with up to about 70W with no obvious problems.

In practice, I’ve used this matcher mainly for 80m-20m to date and it has worked very well and makes antenna setup a breeze with no tuning necessary. I’ve also now tried it with my 40m-15m trapped EFHW and it seems to provide an acceptable SWR for all 5 bands.

Update June 2018:

I’m still using this matching unit regularly and most recently have started using it with another multi-band end-fed design which is even simpler than the trapped EFHW. It does however rely on having an ATU in the rig (e.g KX3/KX2). Details are here.

15 thoughts on “EFHW matching unit

  1. Great article. I enjoyed it. I also found that 3 turns on the primary and 24 on the secondary worked better than the 2 turn primary 16 secondary. The loss figures
    almost are cut in half by using the 3 – 24 winding.
    I put 100 watts through my UnUn (FT-140-43) and did not see any heating issues.

  2. I built an unun based upon your unit, but due to my interest being in the 20M
    down to 160M range, I decided to add additional turns. I retained your ratio
    of 1:8 (3:24) by using a 5 turn radio winding and 40 turns for the antenna.
    Early tests indicate that it is working well.

    73, Bruce, VE3EAR

    • Hi Bruce,

      Yes, that should work fine. I too built a variation for 160m-40m which used 4:32 ratio and this seems to work fine too. I found that for 160m and 40m, the capacitor isn’t needed, but it is required if you want to use it on higher bands.



  3. A transformer wound like that will have a fair bit of leakage reactance, maybe that is what the capacitor is compensating. It’s a fine solution . Be sure to remember bigger caps will have more current capacity, also

    • Yes, I agree. Owen Duffy has just published a design variation that should have lower leakage. I plan to try that too. Yes, the capacitor is only suitable for QRP in reality and I think I will replace it with a 500V silver mica which should be OK to around 100W.

    • Daniel,

      If you are only building for 40m or below, then you don’t need the capacitor. However for 30m and up, the capacitor will lower the SWR enough to be worthwhile I think.

      David VK3IL

  4. You never state in the article what length of wire you generally use with this matching unit. What length is required for 160-10m matching?

    • This matcher is designed for feeding the high impedance of end-fed half wave wires, so the length of wire is determined by the operating band and needs to be cut for a half wavelength on that band. As noted, this design is optimised for the 80-15m bands. For 12/10m you should use a 2:16 turns ratio and for 160m, a 4:32 ratio will work better. If you want a multi-band solution, you either need to use traps (see http://vk3il.net/projects/trapped-five-band-efhw-sota-antenna/ ) or one of the variations of loaded EFHW (e.g http://pa-11019.blogspot.com.au/2012/04/149-transformer-for-endfed-antennas-35.html ). Note also that you can use any multiple of half wave-length too, so cutting a wire for a half wavelength at 80m will also work on 40/20/15/10m (and may work with much higher SWR on the WARC bands), but the pattern will become multi-lobed on the higher bands.

  5. OK1FTA
    Funguje to super, použil jsem pevnou 150pF kapacitu, toroidní kroužek z PC zdroje(žlutý), na malé výkony do 20 Wattů to nemá chybu. S lepším toroidem o průměru 45 mm a s ladícím kondenzátorem jsou parametry přeladění pásem
    ještě lepší.
    Díky za info a 73!

  6. Hi David
    Can you tell answer the following questions?
    1. If I use an 80M half-wave I assume it has a figure 8 radiation pattern
    and on 40M it might be a clover leaf pattern?
    2. Have you tried to use two radiators, e.g. 80M and 40M, to try to keep
    the figure 8 pattern on both bands?
    Tom Dixon

    • Hi Tom,
      Yes, nominally the pattern for 80m and 40m would be as you describe. However, that’s only the case if the antenna is a half wavelength or more above the ground. In practice for portable use, they will usually be much closer to the ground and this tends to result in a pattern which is largely radiating straight up (i.e an NVIS antenna). While it’s a different antenna, if you look at the 3D modelling I did for my home loop antenna here: you will get the general idea of what happens as the frequency increases.

      To answer your second question, I haven’t bothered with separate readiators for each band as it’s not very practical during a portable operation to keep changing antennas each time you want to change bands. However, if you were doing a more permanent installation and could get the radiators higher off the ground, it would be worth considering.

      David VK3IL

Leave a Reply

Your email address will not be published. Required fields are marked *