A random wire antenna’s feedpoint impedance is 450 Ohms: True or False?
It must be written on a stone tablet somewhere that an end-fed random wire antenna should be fed through a 9:1 impedance transformer. Common knowledge says a 9:1 transformer, in conjunction with a tuner will bring the impedance of a random wire antenna down from 450 ohms to a nice 50 ohms to keep your radio happy. And this shall be true for any band that the wire is long enough to tune.
Once something becomes “common knowledge” it is rarely questioned, but at Ham Radio Outside the Box nothing is accepted without first examining the facts, sifting the evidence and remaining skeptical until we can confidently say “quod erat demonstrandum” (QED: roughly translated as: there I told you so!)
It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.
Mark Twain
Is a 9:1 transformer (sometimes referred to as an unun) really necessary – or even desirable?
In an attempt to gather the data, I threw up an 84ft wire in the Outback (out in the backyard) and measured the impedance over several bands with a RigExpert AA-55 Zoom antenna analyzer.
A counterpoise was added and a very short (about 3ft) RG-58 coax was used to connect the wire to the RigExpert. The coax was used simply to stop the weight of the wire from pulling directly on the RigExpert.
Traditionally, a random wire antenna was connected directly to a radio without a 9:1 impedance transformer or coax feedline. An example of this can be found in the history books. Special Operations Executive (SOE) agents in WW2 used this method to send their reports from behind enemy lines back to command posts in the UK.
The direct connect method has one problem for which there is a simple solution. The problem is “RF in the shack”. The problem can be solved very simply by reducing power. At QRP levels the RF voltages fed back into the shack are significantly lower and less of a danger.
The benefit of the direct connect method is no feedline losses. This arrangement works well for a QRP field operation out in the Big Blue Sky Shack but may be less desirable for the home shack.
TABLE 1: Here is what Ham Radio Outside the Box’s rig expert had to say:
| Band | SWR (:1) | Z ohms | R ohms | X ohms |
|---|---|---|---|---|
| 80m | 7.6 | 307 | 256.9 | 168.6 |
| 40m | 34 | 123.5 | 11.9 | -122.9 |
| 30m | 18 | 150 | 27.1 | -147.5 |
| 20m | 5.6 | 87.8 | 35.3 | -80.4 |
| 15m | 6.2 | 52.7 | 17.6 | -49.7 |
| 10m | 16 | 21.2 | 3.58 | -20.9 |
These numbers are for a specific installation and may well be completely different for other wire lengths etc.
A 9:1 impedance transformer (unun) will transform the 50 ohm impedance of the radio to 450 ohms to match the antenna. But will the antenna actually have a 450 ohm impedance? The measured data for this 84ft wire shows that the impedance is less than 450 ohms on each band.
A good antenna tuner has a wide impedance matching range; for example, the LDG Z-11ProII can match impedances up to 1000 ohms, so can this particular antenna be used without a 9:1 impedance transformer?
Table 2: What happens if we do add a 9:1 transformer?
| Band | Z original | SWR original | Z with 9:1 | SWR with 9:1 |
|---|---|---|---|---|
| 80 | 307 | 7.6 | 34 | 2.09 |
| 40 | 123.5 | 34 | 13.7 | 41.35 |
| 30 | 150 | 18 | 16.7 | 18.46 |
| 20 | 87.8 | 5.6 | 9.8 | 13.23 |
| 15 | 52.7 | 6.2 | 5.85 | 25.30 |
| 10 | 21.2 | 16 | 2.4 | 125.27 |
NB: The 9:1 transformer divides both the R and X components of the impedance by 9 which can be shown as: Z = SQRT(R^2+X^2) so Z/9 = SQRT((R^2)/9 +( X^2)/9). This can be verified by plugging in the data from the first table into these formulas.
LDG advises that, although their tuners can provide a match for almost any antenna impedance within their range, an antenna should be as close to resonance as possible to avoid losses in the tuner.
Note: a resonant antenna is considered as having an R component of 50 ohms and an X component of zero.
Table 3: So how do the two scenarios measure up on that basis?
| Band | R ohms | X ohms | R/9 ohms | X/9 ohms |
|---|---|---|---|---|
| 80 | 256.9 | 168.6 | 28.5 | 18.7 |
| 40 | 11.9 | -122.9 | 1.3 | -13.7 |
| 30 | 27.1 | -147.5 | 3 | -16.4 |
| 20 | 35.3 | -80.4 | 3.9 | -8.9 |
| 15 | 17.6 | -49.7 | 2 | -5.5 |
| 10 | 3.58 | -20.9 | 0.4 | -2.3 |
Now if we compare the data obtained without a 9:1 transformer against the data with a transformer, we have to ask which is easier to match?
Referring to Table 2, the 9:1 transformer makes the SWR worse on all but 80m and 30m. Table 3 shows the 9:1 transformer has not really got us any nearer to resonance so we may still experience losses in the tuner.
A commenter on one of the online antenna forums offered the opinion that when using a multiband wire antenna, the results will be all over the Smith Chart, so why not forget the unun and just let the tuner do its job?
Random wire antennas have been used with great success for a very long time, but it is well known that some combinations of wire length, band and tuner get “ornery”. Perhaps this analysis explains why. You can probably still continue to enjoy using a random wire antenna by changing the wire length/band/tuner. Or alternatively, use an antenna that is close to resonance and use a tuner only if it necessary to “touch up” the SWR to keep your radio happy – for example a linked End-Fed Half-Wave.
Disclaimer: this analysis may be completely wrong. I thought I was wrong once before but I was mistaken 😉 Seriously, if you disagree with any of the points in this post please offer corrections in the comments. I have thick skin.
SWR calculations relied on an online calculator at: https://chemandy.com/calculators/return-loss-and-mismatch-calculator.htm to whom I am very grateful for helping keep the math in this post down to a minimum.
Help support HamRadioOutsidetheBox
No “tip-jar”, “buy me a coffee”, Patreon, or Amazon links here. I enjoy my hobby and I enjoy writing about it. If you would like to support this blog please follow/subscribe using the link at the bottom of my home page, or like, comment (links at the bottom of each post), repost or share links to my posts on social media. If you would like to email me directly you will find my email address on my QRZ.com page. Thank you!
The following copyright notice applies to all content on this blog.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Discover more from Ham Radio Outside the Box
Subscribe to get the latest posts sent to your email.
Ham Radio Outside the Box 
I’m not saying you’re wrong or even mistaken; I’m just saying you may not have been using a truly random length wire. At least consider your results in light of the recommendations on page 3-4 of this document on the Balun Designs site.
-73 de K3FZT / Steve
LikeLike
Thanks for the feedback Steve. I read the reference you provided. If I interpreted that information correctly, they are using the coax shield as a counterpoise. My measurements were made at the end of the wire (except for 3ft of coax which would make little difference) and I used two 18ft wires on the ground as a counterpoise. They also recommend a wire length of 88.5ft but at https://www.hamuniverse.com/randomwireantennalengths.html 88ft is to be avoided and 84ft is recommended. 88ft and 88.5ft are too close for comfort. So I guess opinions are divided as is often the case. Perhaps next time I deploy my 84ft EFRW I will try measuring the SWR, Z, R and X at the end of a 50ft coax and see if I get better results. Anyway, I appreciate you taking the time to comment.
LikeLike
I recognized the difference in configuration and should have acknowledged that. I think as you appropriately note there’s plenty of “common knowledge” not supported by data. Your test setup was implememted to evaluate the random wire with and without the 9:1 Unun. It was not a test of a usable antenna for ham radio operation. Thanks for sharing your reference. 73 de K3FZT / Steve
LikeLike
On a recent QRP operation in the Big Blue Sky shack I was using the wire element from a 20m EFHW on 20 meters, BUT I didn’t have the 64:1 un-un that is used to bring the high impedance of the wire end down to 50-ish ohms.
The radio has a built in tuner that was able to match the bare wire to about 1.8 : 1 swr seen at the radio. But then I thought, why not use the 9:1 un-un that I did have on hand to at least do some impedance transformation so the tuner doesn’t have to “work as hard”?
As far as I could tell anecdotally, just by operating, is that this had little if any positive effect. The swr “improved” to 1.7 – big deal!
Thinking about it, the whole interest in swr minimization is to reduce FEEDLINE LOSS. But this antenna is driven at the feedpoint, there is no feedline to have swr loss in! By putting an additional transformer directly between the antenna matching unit and the feedpoint, I was actually introducing a lossy element. The more ferrite cores between the output transistor and the radiator, the more loss of energy you’ll have, regardless of impedance match.
So if you’re stuck with a non-resonant network as a radiator and the “tuner” can make the radio happy, then it matters little what the swr is. The energy is going to either be radiated or consumed in loss. The less loss you bring to the party the better.
LikeLiked by 1 person
Well said Matthew!
LikeLike
I use a random wire with my KX2 and connect directly with a binding post to BNC adapter to the radio. The matching unit in the radio does the magic and I am making contacts. You are right, you do not need the 9:1 – as long as you have a good tuner. I do have a 9:1 and a 4:1 that I use with lesser tuners 😉
de Karl Heinz – K5KHK
LikeLike
Thanks for the feedback Karl. I believe the KX2 internal tuner has a wider range than most internal tuners.
LikeLike
I use a random wire with my KX2 and connect directly with a binding post to BNC adapter to the radio. The matching unit in the radio does the magic and I am making contacts. You are right, you do not need the 9:1 – as long as you have a good tuner. I do have a 9:1 and a 4:1 that I use with lesser tuners 😉
de Karl Heinz – K5KHK
LikeLike
Excellent topic! And my favorite Mark Twain quote, too!
I just gave a local SOTA operator the same advice, to connect her antenna wire directly to the tuner, rather than using an unun. She is going to try it out, as itvwould save some weight.
i do have some thoughts on your data, however. I am particularly surprised to see the very low resistance values. Generally, I’d assume that any wire longer than about 3/8 wavelength would have a resistive component above 50 ohms (unless it is very low to the ground perhaps). In running some other tests, I found that even a metre or less of coax between the antenna and the analyzer can make a significant difference in the measured R and X at high SWR. Your SWR readings, however, should still be good.
so I’d suggest finding a way to support the wire so you can connect it directly to the analyzer, rather than through a coax jumper, and repeat your measurements.
One issue with low resistive values is that antenna tuners are often more lossy matching lower impedances than higher ones, even with the same SWR. Low resistance requires higher current for the same power, so higher losses in the coil resistance. That’s another reason not to use the transformer if your tuner doesn’t need it.
LikeLike
Thanks for your authoritative input Dale. I will certainly try your suggestions. Actually, connecting a wire directly to my antenna tuner should be very easy. I use banana plugs on all my field antenna wires and those plugs fit nicely into a SO-239. I have big alligator clips that can be used to connect a counterpoise.
LikeLike
Hi John
Can I ask why you use a counterpoise on an end fed and then try measure the resistance as though it is now the end of the antenna? I use an EFHW with a 49:1 transformer as my main shack HF antenna, no counterpoise (although there is a short coax to a choke).
I know there is lots of debate about the need or otherwise for a counterpoise on an EF antenna. But is an EF antenna with a counterpoise not just an off centre fed dipole (OCFD)? In which case your resistance is going to reduce per band at your feedpoint because you are no longer feeding at the end but at a point along the antenna where impedance is going to vary with frequency.
I have not done the calculations and I don’t think you declare the lengh of your counterpoises, but your results align (except maybe 30m) with resistance lowering with band. This would align with the OCFD where your high impedance is now actually at the end of your counterpoise and lowers as it approaches your feedpoint in proportion to the wavelength. For the low resistances you observed this would just mean your feedpoint now aligns closer to a current maximum (the point you would look to feed a OCFD).
To truly measure the resistance at the end of an EF antenna should you not remove all counterpoises and limit feedline length just like the original proposition of an antenna wire fed in directly?
I am a relatively new ham and could be well off with my thoughts.
Thank you for your always interesting and thought provoking blog.
73
LikeLike
Thanks for the question Jamie. Let’s first define what we mean by EFHW, EFRW and OCFD.
EFHW: A radiator one half-wavelength long fed at the high impedance point at one end.
EFRW: A radiator that is not resonant on any band of interest fed at one end.
OCFD: A half-wavelength dipole fed at a point part way along its length (e.g. 30%).
An EFHW has a very high impedance at its feedpoint so the return current is very small but not zero. A counterpoise is still required but it can be as short as 0.05 wavelengths. If there is no counterpoise wire, current will flow back along the braid of the coax. If there is no coax the radio and/or the operator become the counterpoise. Even though adding a counterpoise wire makes the EFHW look like an OCFD, the feedpoint is so close to the end that it makes little difference. For practical purposes, at QRP levels, there is so little current in whatever forms the counterpoise that it doesn’t make any real impact and is sometimes omitted. But at QRO levels lack of a counterpoise could potentially damage the equipment or the operator.
An EFRW differs from an EFHW in that its length is specifically not a half-wavelength – or multiple – on any band of interest. The feedpoint impedance is lower than an EFHW so the return current is quite high and requires a substantial counterpoise. Many ops use a simple 17ft wire laid on the ground as a counterpoise. I have never understood why 17 feet is used because it is detuned by the ground and becomes a random wire. Although it “works” it may not be efficient – especially on the lower bands.
LikeLiked by 1 person
Hi John
Thank you for your reply to my previous comment. Your response is as always informative and on point.
If I may ask a follow up question, but first a bit of background as to where my head is. I came to this post after reading your post on the magic conversion of your Long Tall Sally to a Rybakov by changing the counterpoise/radials. Moving from a tuned half wave to random by changing the length of your counterpoise.
I have a 10m squid (crappie) pole that I am looking to use for an expedient outdoor POTA multiband antenna. Reading around Rybakov and other random wire verticals, the general comment is “and add a couterpoise of whatever length you need”.
Where my head is stuck is that a counterpoise is potentially going to turn your random wire into a half wavelength dipole if you choose a length of counterpoise when added to your radiating wire that together equal a half wavelength. My rough calculations have your Long Tall Sally as an off centre fed dipole when using tuned radials, not feeding in the centre requires the 4:1 transformer as I understand it?
Sorry for the rambling. Long story short, I am trying to work out how careful I need to be with the choice of counterpoise length for my random (not a half wave) length vertical and does this change significantly if I go for elevated radials over an on the ground counterpoise.
Feel free to ignore my rambling, but thank you for your blog.
Jamie.
73
LikeLike
I have successfully used a Rybakov for POTA on many occasions, but I use four 13ft ground radials as a counterpoise. This seems to work well from 40m to 10m and avoids any problems with the length of a single counterpoise wire. More radials would be preferred (I have tried up to 8) but, when putting up a field antenna for an hour of operating in a public space, expedience often takes priority over efficiency. The 13ft tripod antenna I have written about recently also works well with the same four 13ft ground radials, but since the feedpoint is about 36 inches above ground, the radials are raised above ground for the first four feet. The 13ft tripod antenna is effectively a half height Rybakov and easily tunes 30m to 10m with four 13ft partially raised radials. I will be writing more about this in an upcoming post. My suggestion: try four 13ft ground radials and see if it works for you.
LikeLike
Interesting topic. The thinking seems to be: “random” wire of a certain length, 9:1 unun, choke, 17 ft counterpoise and that’s it sorted! Not for me. I used 71 ft of wire, a 4:1 unun, a 1:1 unun as a choke and a counterpoise of about 30 feet lying untidily on the ground. The random wire sloped down from 20 feet high to 8 feet, with a high wall no more than 2 to 3 feet from the bottom 30 feet of the antenna. It took a lot of experimenting to get it right. But it covers all bands, from 80 to 10, including all the warc bands.
Location, location, location….and patience!!
LikeLiked by 1 person
One of the common issues with end-fed pseudo-random wire antennas is that many of the parameters aren’t specified. There can be a big difference in performance, for example, between using ferrite vs. powdered iron cores in the transformer, and it may actually step the impedance UP on 80m instead of acting like a traditional transformer.
Let me clarify one thing that DelectablyMellow brought up: the object of a “random” wire length is to avoid having an end-fed half-wave (or multiple thereof) wire on some band, as that typically will have a very high impedance, and may stress some tuners. That doesn’t mean that the wire + counterpoises / radials can’t be resonant. The only reason to avoid using a quarter wave wire on 20m is that it will be a half wave on 10m, otherwise it should still work. (Although if the transformer steps down the impedance by a factor of 4 or 9, it may make the tuner work harder to match it.)
To add a bit of history, using unun transformers to match random wire antennas is a relatively recent fad – it seems to have become popular this century, possibly based on a Japanese article on using such a transformer for a vertical made from aluminum tubing (which wouldn’t have as high of an impedance when it is 1/2 wavelength. Yes, end-fed (truly random) wires have been used for a long time, including for clandestine stations during WWII, but the wires were plugged directly into the transmitter, and tuned up using the output tuning network of a tube transmitter. I’ve used several such end-fed wires over the years, but not with a transformer – just connecting it directly to the output of a manual tuner. (In practice, I prefer to do the same thing with an end-fed half-wave and use it on all bands: a simple L network with one coil and one capacitor will easily match the high impedance on multiple bands.)
As far as radial / counterpoise lengths are concerned, they aren’t too critical. When elevated, they typically are more effective when they are 1/4 wavelength. When lying on the ground, 1/8 wavelength may be better, but, in general, more short wires are better than a few longer ones, up to a point. That affects the ground loss resistance, which is more important with radiators up to about 3/8 wavelength or so. Longer wires have a higher input resistance, and a few extra ohms of resistance has less impacr on efficiency.
For many operators, the primary concern is whether they can match the antenna with the tuner they have, rather than performance or efficiency. In that case, a wide variety of combinations will work, depending on what tuner you are using.
LikeLike
Your knowledgeable input appreciated as always Dale. Thanks.
LikeLike