Thanks for your interest in my HWEF article from 2012. Even though it’s from october 2012 and it’s outdated because I’ve come to many new insights (mostly thanks to the knowledgable Mr. Owen Duffy), I think it’s still a good starting point to get some understanding about this antenna. If you’re interested in HWEF antennas, I recommend you read the more recent articles I’ve written about it:
Articles from Owen Duffy that are worth reading:
RF transformer design with ferrite cores – initial steps (Owen Duffy 2019)
WW1WW’s matching transformer for an EFHW (Owen Duffy 2018)
End fed matching – VK3IL design on LO1238 (Owen Duffy 2017)
End fed matching – PA3HHO design review #2 (Owen Duffy 2016)
End fed matching – PA3HHO design review (Owen Duffy 2015)
End fed wires – new hams love ’em (2014)
Following the original article I wrote back in 2012. Please be aware that…
- … the described antenna in this article is NOT “my design”!
- … this is just a compilation of information from Dutch forum http://www.zendamateur.com
- … the antenna design is based on a PAR EndFedz that was available back in 2012
Multiband half wave end-fed HWEF (2012)
Yes it’s cheap, easy to build, doesn’t need radials and it works GREAT! It sounds to good to be true. Below the result of a 30 minute daytime session on 40m in WSPR mode with 0.6 Watt ERP using my trapped 80/40/20 HWEF. Even though the WSPR mode is very efficient and claims a gain of +30dB compared to a “normal” signal, I think the results are impressive.
In the Netherlands, the HyEndFed antenna by PA3EKE, in principle an improved copy of the PAR end-fedz is a very popular antenna. The antenna seems effective / efficient, can be put up as a vertical, horizontally or as a sloper, is low on QRM/QRN, does not cause any RFI in the shack, is very practical to put up (feedpoint very near to your transceiver) and it’s pretty easy to build. On the (Dutch) forum www.zendamateur.com there’s a lengthy thread on how to build this antenna and many Dutch hams have succeeded in building one themselves. I use one for my holidays and from my home QTH I have put up a multiband end-fed aerial for 80/40 (which is doing OK on 20m as well ). When I’m working international stations on 80/40, many people are interested in my 80/40 antenna, therefore I made this artible with some explanation and a summary of the info I got from the forum.
How it works…
The heart of the multiband (you could as well call it “anyband”) end-fed is a 1:50 (to 1:60) impedance transformer that feeds half a wave of wire fed at the end (or beginning) where impedance is (very) high (aroun 3k Ohms). You could consider the end-fed half wave an “off center fed” dipole that’s fed VERY off center, like a Windom taken to the extreme: total length half a wave divided over one very long leg and one very short leg. The short leg is so short, that we actually leave it out. I have seen some end-feds that have about 1m of wire hanging from the transformer. Just like a windom, it doesn’t seem to require any radials.
Many people are claiming it can’t work without, but people are using this antenne everyday are kinda proving it does. In my perception, VERY well actually. I reckon it needs some ground, but not much because of the high impedance that is “Voltage fed antenna”. Many people claim the outside of the coax will function as a radial and the coax will be radiating causing RFI probems. I’ve used my 80/40 end-fed regularly with 400W with the feedpoint (litterally) 3 meters above my head and I never experienced any RFI problems. Yes it will use a (very short?) length of the coax braid as a counterpoise… so what? I’m using 10m of military grade RG58 that passes an FT-240-43 core eight times before it enters my shack. No problems with RFI whatsoever.
Signal reports I get on 40m are *very* good, comparably with and mostly better than stations that have a full size Windom or dipole for 40. I don’t want to brag, but in many local rag chewing QSO’s I often was one of the louder stations. On 80m my signal reports compare to a friend who used a 2 x 14m open line fed aerial. The end-fed for me however is way more practial because you can span it from an (attic) window and let it slope down to a tree, small mast or garden shed in your back garden. Anyway, the typical impedance at the end of half a wave is around 2500 – 3500 Ohms. The 1:50 auto transformer transforms this high impedance to much lower 50 – 70 Ohms. Once built, you can connect half a wave length for *any* band as the auto transformer just transforms 2k5 impedance to 50 Ohm impedance. Hook up 10m of wire for 20m band, 15m of wire for 30m band, 20m of wire for 40m band; what you like. The “multiband” feature of the antenna is that e.g. half a wave for 20m happens to be a full wave for 10m so the 20m half wave will be resonant as a full wave on 10m. Another trick to add a third band is to add a trap or high impedance coil at the end that “disconnects” the wire that’s after the trap/coil on high bands and acts as a loading coil on the lowest band. Here’s a drawing of the principle for the only 12m long 40/20/10 version:
My 80/40 version (2012)
My 80/40 version that I use on a daily basis, regulary with 400W PEP has a 1:60 autotransformer using two stack FT-240-43 cores. The core has three bifilair windings (= 3 primary + 3 secundary) Than 8 more secundary windings, than it crosses the stacked cores and than there’s 12 more secundary windings. The winding ratio is 3 : 23 or 1:7.7. The impedance transformation will be the squared ratio so a little less than 1:60. My first version had 21m of wire, than a 115 uH coil (1mm enamelled wire on a 50mm PVC pipe) and then only 1.5m of wire. This worked great on 40m and “acceptable” (far from great) on 80m. The bandwidth on 80m however was *very* limited: only 50 – 60 kHz. Because I could put up a little more wire, I peeled down the coil to 80 uH coil. To my surprise, only 3m of wire were sufficient to get resonance on 80m. Again, great on 40m, “reasonably well” on 80m but still limited bandwidth (only 50 – 60 kHz). From the tree that held the end, I could go down another 4m so I replaced the 80 uH core with a 40m coax trap (11 wdg of RG58 coax on a 40mm PVC pipe). After the coax trap around 12m of wire were required for resonance at 3.7 MHz. Just a little more than I could span away. With some coiling around the tree, I managed to get rid of all the wire, but not so neat solution. The bandwidth was now 200 kHz so from 3.6 – 3.8 I was able to operate it without tuner (SWR better than 1:1.5). Total length around 33. Because the resonance dip on 40m was a bit high at 7.15 MHz, I added a length of 30 cm of wire before the coax trap just hanging off the trap. This is a great trick to fine tune resonance if you’re 40m half wave wire is a little too short. Latest version of this antenne has a home brew L/C trap made of a 7 uH coil (on 50mm PVC pipe) combined with a 15 kV 68 pF door knob capacitor. After the trap, there’s now around 11m of wire. I used 8m and than some end loading with 2 wires in a V-shape. Ends of the V-shape are at 2m AGL (safe height for my kids). Bandwidth is now around 150 kHz on 80m.
The “original” (PAR / HyEndFed) 40/20/10 version
This antenna is a 12m long multiband antenna for 40m, 20m and 10m. It has a 1:60 autotransformer with a single FT-140-43 core for 100W PEP or a bigger FT-240 for 400W PEP or 100W CW/RTTY. The transformer has two bifilar windings (= 2 primary + 2 secundary), 5 or 6 more secundary windings, than wire crosses the core and than 7 or 8 more secundary windings. The winding ratio should be between 2:14 (1:7) and 2:16 (1:8) for a 1:50 – 1:64 impedance transformation. My experience is that you need 1:60 to get the impedance of the highest band around 60 Ohms. After the transformer there’s around 10.2m of wire, then there’s a 34 uH coil and abt 2 more meters of wire. This antenne will work as a full wave on 10m. You will need to solder a 150 pF capacitor on the INPUT of the transformer to get the SWR down on 10m. On 20m the antenna will be a full half wave. On 40m the antenna electrically acts as half a wave but physically is dramatically shortened. Bandwidth on 40m is limited to 50 / 70 kHz. If you need more bandwidth and can put out a little more wire, replace the 30 – 40 uH coil with a 20m coax trap or normal L/C trap. Depending on the inductance of the trap, you will need 5 – 6 meters of wire after the trap, total legth 16m. With this version, bandwidth on 40m will then be around 150 – 200 kHz. If you want to go the full monty, just hook up 20.4 meters of wire and you will have excellent effiency and full bandwidth on 40m and some gain on 20 and 10. For a 30/15 meter version, just hook up 15+ meter of wire. For a 80 – 10 version, make an autotransformer like my 80/40 version (with 3 bifilar turns on an FT-240-43) and hook up 20+ meters of wire, a 70 uH coil and 2.to 3 meters of wire (or 40+ meters if you have enough space).
My 2012 holiday version 40/20/10 (only 7.2m long)
My holiday version is very similar to the normal 40/20/10 version but has two coils instead of one. First there’s around 4,8m of wire, then a 15 uH coil (to trap 10m frequencies), then around 1.2m of wire, then a 34 uH trap and another 120cm of wire. This antenna works great on 10m, very well on 20 and of course relatively poor on 40m. When conditions were up I worked whole of Europe from a camping site in France. When conditions are average, performance on 40m is defenitely a compromise obviously.
How to make the “magic box” (1:50 to 1:60 autotransformer)
The “magic box” holds an HF autotransformer (UNUN) that transforms 3000 Ohm to 50 Ohms. The primary : secundary winding ratio is 1:7 or 1:8. For 100W, an Amidon FT-140-43 core will do but for 100+ Watt SSB or long CW or digital mode QSO’s you’re better off with the larger FT-240-43 core. At www.kitsandparts.com (“the toroid king”) the FT-140-43 sells for $5 per 2 pieces and the FT-240-43 sells for $8 a piece. I’m using two stacked FT-240-43 cores for my 80/40 version which easily handles 400+ Watts, probably more. I reckon a single FT-240-43 can handle 250 Watts easily.
On the core, you wind two “bifilar” turns (wires need to be twisted), then 14 more single turns. The primary side of the transformer has 2 windings (one wire of the 2 bifilair windings) and the secundary side of the transformer has 2 (other bifilair wire) + 14 (normal) = 16 windings. The ratio of the windings is 1:8 so the secundary Voltage will 8x the primary AND the secundary current will be 1/8 of the primary. Therefore, the impedance will be transformed in a 1 : 8 x 8 = 1:64 ratio. That’s what we want to match 3kOhm to 50 Ohm right? The original auto transformer has a 1:7 winding ratio (2 bifilar windings + 12 extra secundary windings) which results in a 1:49 impedance transformance. My experience is dat 1:49 results in slightly too high impedance so that’s why I opted for 2 more secondary windings. In practice you will find that on 10m the impedance is too high, on 20m it’s spot on and on 40m it’s too low so you need to find a compromise to be able to work three bands with one transformer. Bear in mind that the bandwidth of the auto transformer is not unlimited, a transformer wound on an FT-140-43 will work well on 10 and 20 and reasonably on 40, a transformer wound on FT-240-43 will work reasonably on 10, very well on 20 and pretty good on 40. If you want to use it for lower bands (merely 40 + 80 or even 160) you need MORE inductance (bigger or multiple cores).
Note that this transformer is different to a 1:9 UNUN that’s popular for random length end-fed antennas. The 1:9 transformer transforms 450 Ohms to 50 Ohms so you could use it to feed a non-resonant (random) wire length (impedance typically between 200 and 600 Ohms). The 1:9 UNUN defenitely needs a counterpoise because it is not a resonant half wave and it needs a reference to ground to radiate properly. The performance of the end-fed is much better than an end fed wire with 1:9 UNUN. On Youtube there’s a nice comparison video (in Dutch, but the pictures paint a thousand words): http://youtu.be/Ybbv_h1HFw0.
Here’s the schematic of the 1:60 autotransformer:
Use 1mm enamelled wire. Twist a length of abt 6 – 8 inch of wire with a length of 40 inch of wire. It is good practice to tape the FT-240-43 core with (e.g. PVC or teflon) isolation (tape). Start winding (back) at point B. This is the point where the bifilair winding ends. Make two windings back. Solder the beginning of the two bifilair windings together. Later, this will be connected to the ground of the SO393 female jack (or BNC if you like). The end of the bililair winding will be the “hot” input; the center of the female jack chassis. Cut exces wire, leave about 1½ inch for soldering. Now from point B (where you started winding back) continue winding the single (long) wire forward and make 7 single windings. Then let the wire go THROUGH the core to the other side and make 7 more normal windings. Cut exces wire (leave 1½ extra for soldering). This end, will be the feedpoint of the wire.
On the INPUT, solder a 150 pF ceramic disc capacitor, preferably 500V or better. This cap “shortens” the antenna for 10m to compensate for the impedance of a full wave (slightly different for a full wave). I used a blue 3 kV cap, they’re available from many electronic shops.
Following some pictures taken from www.zendamateurforum.com (ham radio forum in Dutch) how the transformer should look like.
PA1RIK 40/20/10 version:
PG3N 40/20/10 version:
Close up of PA3GWO’s autotransformer:
Close up of PA3GWO’s loading coil (high impedance trap for 10/20m)
Here’s the transformer of my (PA3HHO) holiday version (only 7m long). I left out the 150 pF cap because my holiday version had TWO coils; one 15 uH coil at 4,8m and one 34 uH at 6m. If you understood the principle; you understand my holiday version is a half wave for 10m as well because the small coil also “disconnects” the wire + coil + wire that’s behind it on 10m.
This is the coil that I made for my first holiday version that was only meant for 20/10. It had the above transformer, then 4.8m of wire, then this coil (I think I recall removing some windings), then 1.2m of wire.
PA3GGI 4-band version (only 9,8m long) by DL4ABB:
How to tune?
First tune the band on highest “half wave band”. For the 40/20/10 version this is 20m (on 10m it is a full wave!!). You can tune by varying the length of the 10m wire between the transformer and the trap. You probably need less than 10m, this is normal, it’s a little less than a ½ wave because of the velocity factor of the wire you use. If SWR is best high in the band, you might need to add a little more length. If SWR is best low in the band, you might need to make it a little shorter. A good tool for tuning is an antenna analyser (I use a RigExpert AA-54; can’t live without it!), maybe you can borrow one from a friend. If 14.2 MHz is OK, the SWR on 28.5 should be OK as well. If not, you might want to “play” with different value capacitors, anywhere between 100 and 200 pF. My experience is that 150 pF seems just right. If the highest bands are OK, tune the shorter wire after the first trap or coil to tune the next half wave band (40m) and so on. Use an isolator at the end and fold some of the wire back (using tie wraps) this is an easy way to make it shorter or longer. The wire that’s fold back is more or less “almost invisible” electrically (not totally, but it works). If you have too much wire and SWR is optimal low in the band, just cut 3” and start tuning. The tuning might be different when erected to a fishing pole (use fiberglass poles only, NOT the ones with carbon!!) compared to put away horizontally or as a sloper, not much though.
Here’s a nice Youtube video of a Dutch novice (PD prefix) station working Canada with only 25 Watts using an end-fed. He almost burst into tears (He keeps repeating “I can’t believe it”)… What a nice hobby! http://youtu.be/F7kbBDftRRc
All pictures I used are “borrowed” from from the Dutch forum www.zendamateur.com, if you’re interested take a look at the (pictures in) the whole thread: http://www.zendamateur.com/viewtopic.php?f=5&t=6682. Some browsers can translate pages using Google Translate. There’s many pages in this thread with many more nice pictures. One guy even made a “mobile version” on his car, 12m long, can you believe it?!
If you have any questions, just let me know (leave a reply on this blog or send me an email, info is on QRZ.com).