Easy fix: just do CW; your paddles won’t be affected. Just kidding!
Odd. Does your KX2 have the internal ATU? Mine always gets 1:1.0 regardless of what I’m doing. Does your EFHW use a 49:1 UnUn? I’ve tried my EFHWs as a sloper, inv-V, inv-L and normally inv-7 with choke-less feeders.
I am assuming that you are using the tuner. This may be symptomatic of an EFHW that is not tuned.
The EFHW idea is that at EFHW resonance, there is a is very little current in the counterpoise/feed-line-braid. As you move away from the EFHW resonance, in either direction, the counterpoise current increases, and you start to have RFI problems you didn’t have at resonance. While your tuner cancels the reactance, and matches the impedance seen at the BNC, it does not change the RF current in the braid/radio/headphones - that is entirely a function of the distance from the EFHW condition of the wire.
I find that changing counterpoise length or lifting the radio off the ground doesn’t change the EFHW resonant frequency (and it is a bit of a test that I have enough counterpoise/coax).
If I have my EFHW resonant at 15m , when I change to 12m and transmit, I get RFI problems in the microphone of the FX4 (well I did until I installed filtering of the PTT line)
I forgot to mention, I don’t use a counterpoise wire with my EFHWs any more. I tried one years ago - running parallel to the antenna wire & directly beneath it - and it seemed to have no noticeable effect on antenna performance running whilst running 10W of CW or cause erratic behaviour of my rig / iPhone / Casio wristwatch or my dental fillings. Or maybe I’m the counterpoise when I sit close to the antenna pole.
So less setup/pack-up faff now. IMO if a counterpoise were really necessary with EFHWs then I might as well deploy my linked dipoles instead.
At the end of the day, the antenna is the antenna, and without a choke on the coax feedline the outer shield of the coax is part of the antenna unless choked. An DC connected shield on the EFHW transformer turns the EFHW into a slightly weird OCFD with the feedline completing the OCFD (wierd because of the impedance/inductance of the EFHW transformer.) It’s also definitely true that the amount of RF will depend on the frequency and the full system resonance.
After encountering the RFI I dropped in my nanovna and watched the swr change as I changed how the feedline was run and how my body was positioned relative to the feedline. Dropping a choke onto the feedline almost completely eliminated these effects and also eliminated the RFI.
Tuning wasn’t the issue. I noticed the SWR changes once I dropped the NANOVNA when i was seeing RFI to see if I was out of resonance. The antenna was well turned but the structure would shift around as I moved. When I inserted a choke at about 1m from the antenna the resonances shifted a tiny bit but the sensitivity was dramatically reduced. Maybe I had been at a bad feedline length. I guess these results feel pretty intuitive. The feedline is, for better or worse, a part of the antenna system, and and EFHW with a counterpoise (or unchoked feedline) is more of an OCFD than a true EFHW.
Dale,
OK it may be possible to use say a short 50 ohm coax connected to a high impedance feed point and get lower (I^2)R losses but the dielectric losses will be higher due to the higher voltage and negate some if not all of the apparent advantage. How much lower would the loss be? Has anyone exploited this to measurable advantage?
73
Ron
VK3AFW.
Dale,
Your point about the length of a simply connected coax being potentially a significant part of the antenna is good. However, using a balun at the feed point won’t stop CM current if the coax runs asymmetrically wrt the radiator.
I believe that if you swapped the 50 ohm coax for a 50 ohm twin wire feed the results should be the same wrt CM. Twin feeders do radiate.
73
Ron
VK3AFW.
Ron,
It is true that common mode current can be introduced on the feedline when it is asymmetric to the antenna: that would be in addition to the common mode from connecting the outside of the coax shield to the feedpoint.
The difference is that, even when the feedline is symmetric with both sides of the antenna, the latter case forces common mode current. That doesn’t happen with balanced line.
Due to skin effect in the shield, coax actually acts like 3 separate conductors at RF: the center conductor, the inside of the shield, and the outside of the shield. The RF current is balanced in the first two (equal in magnitude and opposite in phase) because of the shielding, and any common mode current is on the outside of the shield.
So. for a straight dipole, we have one wire connected to the center conductor, and two wires (the antenna wire and the “third wire” of the coax) connected to the other side. That antenna is inherently unbalanced due to the extra wire.
With balanced line on a straight dipole, the antenna is still balanced, and the effect of connecting the feedline on one side of the feedpoint is balanced by the same connection on the other side.
I ran a couple quick models (which aren’t perfect by any means, as I had to create the feedline using two parallel wires to see the current on it). With the feedline connected directly to a straight dipole, the twinlead was nicely balanced, but with coax feed in the worst case up to about half of the feed current can flow down the coax as common mode current (depending on many factors, of course, including the relative diameters of the two conductors). When I made the feedline asymmetric to the wires (at about a 45 degree angle under one wire), the current in the closest wire was about 70% of the current in the other wire, and the missing 30% was common mode current (well, roughly, as I just looked at the current magnitude and not the phase).
So, yes, both can be causes of common mode current. The first problem with the coax shield connected to the feedpoint can happen even with a perfectly symmetric arrangement of the feedline, and is resolved with a balun, or using twinlead as a feeder. Asymmetric feedline is a second cause of common mode current in addition to the first, and also affects balanced line.
In my experience, the first is more common, and can’t be resolved simply by rearranging the feedline. It is also highly dependent on the coax length, etc. But both can cause problems.
So I’m not trying to ignore the issues with asymmetric feedlines relative to the antenna, but I don’t think that is the most frequency cause of common mode current for most hams at HF.
Hi Dale,
I appreciate your comments.
I have no quantitative data on which is the greatest cause of CM current on feedlines. I rarely see a feedline brought away from the feed point of right angles for the whole of its drop. It’s something that is ignored these days. On the occasions I haven’t done that I’ve had grief.
I have trouble conceptionalising the relationship between line SWR and CM current as a percentage of the feedline normal current. It would be case dependant. Maybe you can model that. Of course there are a lot of variables, feedline length being the main one so worst case/best case calcs are probably enough.
I would not use a coaxial feedline with an SWR above 3:1 for reasons other than CM. Actually my pain threshold kicks in around 1.7:1.
For my multiband doublets I regularly use twin 300 ohm feeder or a HB OWL with > 10:1 SWR so I’m not against using high SWR is particular circumstances where the losses are modest.
73
Ron
VK3AFW
Hi Ron,
Part of the issue is that there is no relationship between SWR and CM. They are totally different things. The SWR (reflected power) is on the two conductors inside the coax, while common mode current is on the outside of shield.
Well, different, but not entirely independent. When you have common mode current on a coax feedline, it means that the feedline is effectively part of the antenna. So any changes to that leg of the antenna can affect the feedpoint impedance, and hence the SWR.
A friend did a lot of adjustment to get a low SWR on his attic 40m dipole. It worked great, and he made many contacts out over 3000 km. Then one day he plugged another coax into the unused port of an antenna switch in the shack, and the SWR jumped way up. That’s because he had changed that leg of the antenna, which changed the current distribution and the feedpoint impedance. He added a balun at the feedpoint, readjusted the antenna, and now he could add or remove stray cables in the shack without changing the SWR.
But he couldn’t work stations beyond 2000 km: he no longer had the low angle vertically polarized radiation from the feedline, only the higher angle radiation from the dipole.
I was testing an 80m dipole, trying to see how much adjustment was needed to shift the resonance from 3.5 MHz to 4 MHz. I discovered that folding back the ends affected the resonant frequency at the high end of the band, but at the low end. Turns out the antenna had two resonances: one due to the antenna, one due to the length of the coax. Adding a balun eliminated the response I was seeing at 3.5 MHz. Years later I ran into the same problem again (probably because I happened to be using the same coax). Being in a hurry, I just added 8 or 10m of coax to shift that resonance out of band, and completed my testing.
One day we were trying to tune a short 40m dipole made using two mobile whips. We kept lowering it to adjust the tips, then raising it back up again to measure the SWR. Then we lowered it and raised it back without making any changes, and the resonant frequency still moved. It turned out that how the extra coax was coiled on the ground had more impact on the resonant frequency than the changes we were making to the antenna.
So, common mode current can cause some very quirky symptoms, even with antennas where the SWR is 1 : 1.
I’d like to do some studies of how far the feedline needs to drop to reduce the common mode pickup due to asymmetry with the antenna wires to a “reasonable” level. The case I modeled was where the feedline angled down at 45 degrees right under one wire, while more often the coax hangs down fairly straight for some distance before bending off one way or the other. (The length of coax laying on the ground may make a difference, too.) One of these days I need to build a common mode current meter (there are several designs) for that purpose (and others).
The more I have experimented with antennas, especially dipoles without a balun, the more of these puzzling situations I’ve been able to explain as due to common mode current. I now have built a couple of my dipole kits with a balun at the feedpoint for more repeatable results. But for a lightweight portable antenna that will survive being dropped on the rocks when the mast falls over, I still usually grab the ones without the balun, as ferrite tends to be heavy and fragile. And usually it works well enough.
73,
Dale WB6BYU
Dale,
Interesting story about your friends attic antenna. Many years agoI tidied up a friend’s antenna and it did not work as well as before.
MFJ sold a multiband wire antenna that used a section of the feedline as a nominally vertical radiator on 15 m.
You make a good point re fragility of baluns. Yet it makes sense that you use one to interface coax to a dipole centre.
These days I use a near end fed 20 wire with a 36:1 transformer in a plastic box. It has survived approaching 50 activations and covers 40 m and 20 m without an atu or end adjustment. The small extra weight penalty is repaid in having a reliable radiator that can take some abuse. For when I need 80 m contacts I use an 84 ft wire with 42 ft counterpoise, no feedline and ATU or a ZS6BKW multiband centre fed doublet and ATU.
There are many ways to radiate a signal.
I’m still thinking about a run of coax having less loss than when matched.
73
Ron
VK3AFw
Ron
Over the years, having written a few lines of code to model coax losses, I found this paper quite helpful.
YMMV
73 Dave
Hi Ron,
Yes, that’s a quirky result. I like using it as an example because it stretches our common understanding of coax loss.
Here’s an example from my article on measuring coax loss:
I wanted to measure the loss of 15m of RG-174 coax at 1 MHz. The common method is to measure the return loss with the far end open or shorted, and divide that by two. That accounts for the coax loss up the cable to the far end, and then back again, with 100% reflection. We used that method professionally for measuring the losses of striplines in test fixtures. (You can do this also by measuring the SWR instead, and converting it to return loss.)
With the far end open, the return loss was 0.5 dB. That would indicate a coax loss of 0.25 dB.
With the far end shorted, the return loss was 3.7 dB, for a cable loss of 1.85 dB.
That’s for the same piece of cable, at the same frequency. And, no, that’s not just measurement error: I get similar results from a reputable cable loss calculator (although the measurements are pushing the accuracy of the equipment).
At 1 MHz, virtually all the losses in the cable are due to the AC resistance of the conductors. In the short circuit case, current is maximum at end of the cable, and there is no corresponding current minimum because the cable is too short. Thus, the average current is greater, causing higher losses through the conductor resistance. When terminated with an open circuit, current is minimum at the end of the cable, and there is not corresponding maximum, so the average current is lower, for less loss. The Matched Line Loss is based on the current in the matched case, which is constant along the cable.
In this case, averaging the two readings gives a matched line loss of just over 1 dB, or 2.1 dB / 30m. The datasheet for Belden 8216 shows 1.9 dB / 30m, which is in pretty good agreement, given that I used that particular piece of coax when wandering around Aus as VK2DJW back in 1980, and am still using it with various portable antennas.
So that’s an example where, for a short cable (less than 1/8 wavelength), the loss could be less than the Matched Line Loss when terminated in a high impedance.
That’s not to say that I would want to operate with such a feedline, as there would be additional losses in the needed impedance matching, and other feedline choices would have lower losses. (Twinlead would be more efficient for a high impedance antenna.)
By the way, AC6LA’s TLDetails program gives both conductor and dielectric losses for each type of feedline, if you want to see how they vary with frequency.
73,
Dale WB6BYU
Hi Dale and Dave,
Thanks again. This one of those cases where size matters. Yes I see the losses in a cable connected to a high Z load are less than the matched loss for short line lengths used.at long wavelengths.
So how to capitalise on this for SOTA operation?
73
Ron
VK3AFW