ad: k1jek

Feeding an unbalanced antenna

Discussion in 'Antennas, Feedlines, Towers & Rotors' started by WV6U, Apr 13, 2011.

Thread Status:
Not open for further replies.
  1. WV6U

    WV6U Premium Subscriber

    After trying a few configurations over the past week, I settled for the following antenna configuration on my deck:
    - 90' loop ( 40' x 5' ) of 14AWG stranded wire along the deck's ceiling
    - At the feed point, a 1:1 current balun connects antenna to 18' coax that runs to an auto tuner

    The antenna seems to work well on 40m and 20m and usually the tuner will match to below 2:1 SWR. I have been able to make contacts along the west coast and towards east coast running barefoot.

    That said, there are obvious flaws in how the antenna is fed. While the antenna is unbalanced, it is fed with a balun. Well, I used whatever I had in my junk box but now I am looking to improve it. Of course, the easier route is to stick a remote tuner at feed point but I want to understand what's at play here. Testing with an antenna analyzer (MFJ-259), I connected a 3' coax to the balun and analyzer. At the middle of 40m, resistance shows up at 500 ohms and SWR goes to beyond 3:1 in the red zone. Readings for 20m are the same. On 80m, the lowest SWR is 2.5 with resistance hovering around 100 ohms. When same measurements are made with an 18' coax, on 40m, the resistance hovers around 50 ohms and SWR is still in the red zone. On 80m, the SWR goes to 2 and resistance shows at around 75 ohms. So clearly the 18' coax has high losses due to high SWR. Right?

    Another test I did was connecting the auto tuner directly at the feed point and then run a coax from tuner to transmitter. The difference was night and day between this set-up and the former one of running coax to tuner from balun. A station from AZ that wasn't audible at all and was below the noise floor went up above the noise floor to 57-58. Now, even when I turned off the tuner that puts it in bypass mode, the AZ station was a good copy. That essentially means the 4:1 current balun built into the tuner improves matching significantly because in bypass mode only a 4:1 current balun is in path of the feed line.

    Questions I have are:
    1. How does a balun impact the system when both ends are unbalanced?
    2. How does a step down balun impact matching for a band/frequency that is already close to 50 ohm? For example, if the antenna presents an impedance of 100 ohms at 80m but 500 ohms at 40m then a 9:1 balun would help match 40m but what happens on 80m?

    Also, if someone can point me to good online text about baluns and impedance matching that is relatively mathematics free (if there is such a text) :)

  2. KB2BLS

    KB2BLS Ham Member

    Just a suggestion to try. My 40 meter loop is in my attic and to get enough wire up to resonate on the low end of 40, it's folded back on itself in a "U" shape. Using a LDG 1000 pro tuner I tried feeding it with 50 ohm coax, twinlead, and open wire with marginal results. I tried baluns on one or both ends and wound a choke on one or both ends, with no improvement. I now feed it direct with 75 ohm coax, from the cable company, and it will load 6 meters to 40 meters, using the same antenna tuner, don't use baluns or chokes. It tunes the warc bands great.
  3. K4SAV

    K4SAV Ham Member

    A 40x5 ft loop does not look like a loop. It looks more like a folded dipole. The feedpoint impedance will be no where near 100 ohms. If you feed it in the middle of one of the long sides, it should be resonant somewhere around 10.5 MHz, with a feedpoint impedance of about 238 ohms. If you feed it in the middle of a short side, it should be resonant at about 11.1 MHz with a feedpoint impedance of about 7.5 ohms. Notice the huge difference depending on where you feed it.

    If the antenna is fed at a point where the loop is symmetrical, then the loop will be balanced. You need a good choke with this antenna. There is a possibility for large common mode currents on some frequencies.

    This antenna will not work well at 40 or 80 meters. It will look like a short circuit at those frequencies, with SWR in excess of 100 to 1. However, common mode currents (without a good choke) can modify that SWR and you may see something much lower. In that case the feedline and house wiring will be doing most of the radiating.

    None of the data you gave jives with anything obtainable with EZNEC, so I suspect there are other factors at play, like proximity to other conductors, a choke that doesn't work at the operating frequency, or hand capacity when making measurements. If fed in the center of a long side using a good choke, SWR on 3.7 MHz when measured thru 3 ft of coax should be about 970 to 1. On 7.1 MHz it should be about 167 to 1.

    Jerry, K4SAV
  4. WV6U

    WV6U Premium Subscriber

    Sorry, I did not fully describe the antenna configuration.

    The loop (40'x5') or folded dipole is fed to +ve side of a 1:1 current balun. The -ve side of the balun is connected to a welded wire fence (on wood railings) that covers three side of the deck, runs about 50' and is 4' high. The unbalanced side of the balun is connected to a tuner with a 18' RG-8X coax. Distance between top of the fence and loop (on deck ceiling) is about 6 feet. The feed point is at a corner of the loop near the deck door such that I can run shortest possible coax to the tuner. The long side of the loop (and fence) runs North-South.

    You might ask why this configuration? Trial and error. I tried:
    - Long wire and fence as counterpoise fed directly to tuner
    - Pseudo loop, that is, dipole with two L-shaped elements and fed with 1:1 current balun
    - Loop fed with 1:1 current balun
    - And, finally what I described above.

    I cannot find a good reference design for this configuration but the antenna works really well so far on 40m and 20m. On 40m, I routinely get 59 reports from up and down the west coast all the way to San Diego in the south and Portland in the north. This morning, I made a contact in Vancouver and got a 55 report. I have also been able to make contacts towards the east coast to GA (got 59) and KY, again 40m. I haven't tried 20m much mostly because I am new to 40m and excited about working this band but on 20m, I copied a station from NH at 599+ the other day. 80m loads up really easily to under 1.5 SWR with the tuner but I haven't tried working on this band much either.

    So back to the question about how is this configuration working and how does the balun handle two unbalanced sides? How can improve this antenna other than the obvious solution of putting a remote antenna tuner right at the feed point?

    Thanks for the responses.
  5. K4SAV

    K4SAV Ham Member

    Thanks for the added description. No that's not a loop or a folded dipole. It doesn't have a name, and you won't find any references to this type antenna. You could get an approximation of its performance by modeling it in EZNEC. I tried but could not get close to your numbers, which means I am still missing some of the details.

    But I can guess. Since there has to be a wire from the fence to the loop above (open or closed loop?) and the feedpoint is on that wire, that 6 ft long vertical wire will look like a top loaded vertical on 80 meters. It should be omni-directional on that band. On the higher bands it is likely to radiate mostly from the horizontal portion of the long wires, so I would expect to see mostly broadside (to the long wires) radiation on 40 meters. On the higher bands, it's likely that the best direction will be at some weird angle to the long wires.

    A current mode 1 to1 balun (another name for a choke) is made to handle unbalanced antennas. If it has high impedance at the frequencies your are operating at, it should work well for this application.

    It will be difficult to determine just how well it is working without any other antenna to compare it to, but if you are happy with it, enjoy. I can't even determine what the loss in the coax might be. My model says about 3.6 dB on 80 and 3.3 dB on 40, but my modeled impedance doesn't agree at all with your measurements, so the coax loss may be totally wrong. Your experiments with the remote tuner, also suggests that the coax loss may be very high even though it is very short. My gain numbers show about -6.9 dBi on 80 and -5.4 dBi on 40, but those could be in error also.

    It's difficult to recommend improvements for this antenna since I know that I don't have all the details, but if the loop on the ceiling is a closed loop, try breaking the loop on one side of the feedpoint so that it is open. That should offer some improvement in SWR and gain.

    If you want to supply a detailed drawing with dimensions, I will improve my model and tell you the results.

    Jerry, K4SAV
  6. WV6U

    WV6U Premium Subscriber

    Closed loop. I started with an open loop but found that closed worked better.

    Jerry, here is a diagram:

    I was wondering if a 9:1 unun would work better for 40m and 20m given that impedance is around 500 ohms on those two bands. But, I am not sure what would it do to 80m. How do I estimate loss on that 18' coax?

    Thanks! :)
  7. K8JD

    K8JD Ham Member

    Looks like a BIG capacitor antenna.
  8. WV6U

    WV6U Premium Subscriber

    Where is the inductor? The tuner? Doesn't have a big enough coil.
  9. WB2WIK

    WB2WIK Premium Subscriber

    Sounds to me like two random length conductors which aren't similar and thus yield an unbalanced load.

    The bands have been pretty good lately, so anyone can work anywhere with anything!

    But I certainly can't complain about that. Propagation is finally getting a bit better here in what ought to be Cycle 24.

    I heard 9K2MQ on 20m while I was switched to my 80m inverted vee, which is about 30 dB down from a 20m antenna.

    I love propagation!
  10. K4SAV

    K4SAV Ham Member

    Thanks for the diagram. I wasn't off too much. I still can't get anything close to what you are measuring, unless you are giving me the magnitude of the impedance and not the resistance. With the bottom of the deck at 4 ft, here is the EZNEC data. All the impedance numbers and SWR are for 3 ft of coax to the antenna. The coax loss and gain numbers are for 18 ft of coax to the antenna. Tuner loss is not included.

    3.5 MHz, Z = 3.2 - j137, Z mag = 137, SWR = 67, coax loss = 3.0 dB, gain = -6.5 dBi
    4.0 MHz, Z = 4.9 - j59, Zmag = 59, SWR = 19.4, coax loss = 1.5 dB, gain = -5.2 dBi
    7.0 MHz, Z = 101- j927, Z mag = 923, SWR = 149, coax loss = 4.8 dB, gain = -5.9 dBi
    7.3 MHz, Z = 57- j681, Zmag = 583, SWR = 136, coax loss = 4.8 dB, gain = -5.7 dBi
    14.0 MHz, Z = 470+ j174, Zmag = 501, SWR = 10.7, coax loss = 1.2 dB, gain = 2.4 dBi
    14.35 MHz, Z = 398 - j516, Zmag = 651, SWR = 21.1, coax loss = 2.3 dB, gain = 1.0 dBi

    As you can see the coax loss is very high, especially for only 18 ft of coax. That is because of the very high SWR on the coax, plus the gain is low because the antenna is very low to the ground. Adding a 4 to 1 balun doesn't help (even if the balun happened to operate properly at those impedances).

    It's a weird antenna. You would have to be very lucky for it to perform well over a wide frequency range. I tried several configurations using roughly the same wires in the same space but got nothing better. For comparison, a 40 meter half wave dipole at ceiling height should beat it by about 11 dB on 40 meters. An 80 meter dipole at the same height would beat it by about 13.7 dB. I realize you don't have room for either of those antennas. (Do you? How did you implement a long wire?) Those are comparisons of maximum gain numbers for the two antennas and doesn't include the fact that they don't both have maximum gains at the same angles, so the comparison is incomplete (and in the case of 80 meters somewhat misleading for low angle signals). A 20 meter dipole on the ceiling would beat this antenna by about 3.5 to 4.8 dB on 20 meters, and you do have room for that.

    Jerry, K4SAV
  11. WV6U

    WV6U Premium Subscriber

    Thanks for staying with me so far and running the model through EZNEC. The analyzer is MFJ-259 and to be honest, I am not sure how to interpret the readings. The manual says "The resistance meter section of the MFJ-259 will provide resistance readings for the load connected to the ANTENNA jack. The resistance meter will not give an accurate reading of a reactive load. For example, if an antenna is resonant at 7.1 MHz and you attempt to measure the resistance at 7.3 MHz, the resistance reading will not be accurate."

    One thing is certain that the tuner loads up all three bands - 80, 40 and 20, to under 2.5 SWR. The tuner's range is limited to 6 to 1600 ohms so SWR cannot be great than 32. Right? Else, the tuner would fail much less actually feed any real power to the antenna to radiate.

    Well, if I keep the antenna, then I think I am definitely going with a remote tuner to get best efficiency from the feeding system.

    :) True and true. This morning I made a contact on 40m in Vancouver, BC and near Portland, OR. In the evening, on 20m, I reached GA and then worked two contacts on 40m - in CO and Southern CA. Following Occam's razor, the simplest explanation is one WB2WIK provides, that is, it's just a piece of random wire radiating poorly but helped by good band conditions.

    For the long wire, I was able to use roof of the apartment below mine that extends in front of my deck. But I cannot use that roof for any permanent fixtures. I will post pictures in the morning that will help visualize the physical space and dimensions. Helped by your EZNEC modelling, the simplest test I can do is put up a 20m half wave dipole, prune it to lowest SWR and compare it to the "weird" antenna. If the dipole wins, we know WB2WIK was right, else I will need to buy better test equipment to figure out what's going on here :)
    Last edited: Apr 14, 2011
  12. K4SAV

    K4SAV Ham Member

    OK. That means you are giving me the magnitude of the impedance. If you measure the antenna at its resonance point, then that number will represent the resistance. At any other frequency that reading will be the magnitude of the reactance. That makes my EZNEC data agree with your measurements.

    No, the SWR the tuner sees will be reduced by the feedline loss. I see that I made an error on the SWR numbers. I intended to give you the SWR looking into that 3 ft of coax, but what I gave you was the SWR looking into the antenna (no coax used). You are using 18 ft of coax when connected to the tuner. That will lower the SWR the tuner sees, due to the coax loss. Example: At 7 MHz looking into the antenna, the SWR is about 149. The SWR looking into the input to the 18 ft coax (what the tuner sees) is 28.2. As far as what the tuner can handle, it's somewhat of a mystery to me. They usually specify the load range in terms of resistance, and don't say anything about reactance, so I really don't know what it will do.

    You didn't give me the height of the antenna before, and I made a guess. It's higher than I guessed. That means the gain numbers should be a little better than I quoted. The impedance and SWR may be a little different also.

    The dipole should beat your weird antenna. It seems that your test equipment may be giving you good data. The EZNEC calculations agree reasonably close to your measurements except for the SWR on 80 meters (now that I know the number represents magnitude of impedance and not resistance). I should have looked up the MFJ-259 analyzer up to see what it gives for an output.

    I would have answered your post earlier, but I was out fishing all day today. It was a very nice day to be out on the water.

    Jerry, K4SAV
  13. WV6U

    WV6U Premium Subscriber

    Sorry for the delayed response, Jerry and thanks for the info. I wanted to put up the 20m dipole for comparison and also do some more measurements.

    First, a couple of pictures that might help visualize the physical space.

    And, here are measurements I took of the antenna with open and then closed loop. Doesn't seem to be much difference between the two.

    I also put up a 20m dipole (as seen in the first photo). Unfortunately, when fed with a 1:1 balun and coax (50'), the losses seem too high. And fed with ladder line (22'), the tuner won't load it up (figure that!!). I am not sure how good an idea is ladder line on such a busy deck. It is almost impossible to have the ladder line not be near something object or much worse, another cable that will change it's balance. With a coax, maybe a 4:1 balun would do a better job? How do I measure impedance at feed point?

    For the "weird" antenna, I relocated the desk closer to the door such that instead of the 18' coax, I now have only 6' of coax from feed point to tuner. How much loss would I save by moving the tuner to the feed point now?

  14. K4SAV

    K4SAV Ham Member

    You still haven't told me the height of this antenna. I originally assumed you were in a single family house. I can see from the photos that you are in a high rise building and also on the top or the side of a large hill. Another guess, maybe 50 ft from the ground? It also seems likely that the floor and roof of the buildings have a lot of concrete and steel. There could be steel immediately above the antenna. That should have a significant effect on the propagation and antenna impedance. There also appear to be an ethernet cable of unknown length only inches away from the loop (which could very easily account for your SWR measurements on 80 meters). Yep, it's a weird loop and almost impossible to analyze accurately with EZNEC because of all those unknown objects. Given all those conditions, don't take my impedance numbers or gain values very seriously. There is a very good chance they are in error for your antenna.

    Typically a raised ground, like an antenna on a roof, may have local ground loss depending on exactly what is under it, but being up high the antenna will have some low angle lobes with good gain which wouldn't be there if the antenna was lower. In addition to that, sloping ground can offer some additional low angle gain sometimes (it depends on exactly what the terrain is). So I suspect that the antenna has a lot more gain than what I calculated. The impedance, SWR, and coax loss could be significantly different too. Whatever it is, by reducing the length to 6 ft you cut the losses by 2/3.

    EZNEC can't handle sloping ground. It can handle a two level ground system. The results will only be as good as the model, and if the model omits a lot of things (like in this case), all bets are off.

    I see your measurements show 500+ ohms, not 500 ohms, and the SWR for most measurements are off the scale of the analyzer. The only useful information in that table is the resonances close to 3.6, 13.45, and 22.15 MHz. At least my model does show the resonances at 13.45 and 22.15, but it doesn't show the resonance at 3.6 MHz. It does show a resonance at 4.5 MHz but with a high SWR. Still I don't trust the answers.

    The 20 meter dipole inside the loop, and adjacent to the ethernet cable, and under the roof steel (likely) will also be affected by those objects. The only way to verify this will be to take data and compare it to predicted results for a dipole in the clear. Either way it might operate OK. You will just have to experiment with it.

    Jerry, K4SAV
  15. G0GQK

    G0GQK Ham Member

    This is intriguing. I've been reading the mail and trying to imagine the situation and at the end you say you are going to make a 20 metre dipole to make a comparison. I thought perhaps you were unable to have a dipole. Why not cut to the chase and make a dipole as long as you can, say 135 ft, drop the ends to make it fit, feed it with open wire feeder into a 4:1 balun ?
    I did notice that it was more successful using the 4:1 balun, like I always say, it will be, but current thinking is that a 1:1 is the answer. I wonder how much power you are actually radiating ? You could be operating QRP !
  16. WV6U

    WV6U Premium Subscriber

    Ah, I see. Sorry if I did not cover this already but yes, it is a four story apartment building on the side of a hill. I am on the second floor and that is about 12 feet above street level on one side of the hill. QTH is at about 605' from sea level. The deck is about another 12' high.

    Yes, the joy of trying to put up an antenna in an apartment especially a rental. You have almost no control of objects that can interfere with the antenna. Did I tell you there is a thick power line about 40 feet away from the antenna and runs parallel to the antenna's long sides? :) It is a power line that runs along the street on the other side of the apartment.

    You see the ethernet cable because it is white but tacked to the ethernet cable is a black cable TV coax.

    As far as I can tell, right above the antenna is only a wooden deck but I don't know much about what's inside the apartment walls and roof.

    I was looking at a RF current probe and found some references to measuring various parameters of an antenna using an oscilloscope, RF current probe and signal generator. How well would this work? Radio/Experimentation/RFProbe.htm

    I don't trust that 80m reading much either. For one, I have only heard one station on 80m so far and made no contacts on 80m. Two, noise that sounds like RF from a power source pretty much crowds 80m.

    For the analyzer, I used to have a miniVNA analyzer that showed a lot more information that his big black metal box from MFJ. I sold the miniVNA couple of year back when I got out of HAM radio. Now, I am kicking myself for it. Still on the lookout for a more sophisticated analyzer.

    The 20m dipole, connected to antenna analyzer via a 1:1 current balun and 3' coax shows lowest SWR of 2.5 at about 14.1 Mhz with a resistance of 30 ohms. But I am not sure yet about how to make measurements on a dipole with equipment that is primarily designed for 50 ohms loads like this MFJ analyzer.

    All said, I am determined to solider on and see how much I can squeeze out of this space and antenna :) Thanks Jerry.
  17. WV6U

    WV6U Premium Subscriber

    Thanks for patiently reading through the thread. Yes, a 20m half-wave dipole is do-able since I have about 40ft of deck space but I wanted to build an antenna to get on to 80/40m. 135' won't fit because there is no space to drop the extra 95' anywhere :) Unlike a lot of other places where apartment blocks have space between them, in San Francisco, most apartment blocks and houses are right next to each other with no space in between. So three sides of the deck don't have access to the street below directly and on the fourth side it does not have a drop down to street either. Instead, the deck opens on the roof of the apartment below (see photo above). Fixtures or antennas sticking out of the apartment unit are explicitly prohibited and I don't want to antagonize the apartment managers who have been very nice to us.
  18. W5DXP

    W5DXP Ham Member

    A 1:1 choke, like a W2DU balun, can be a BALUN or UNUN and/or choke depending upon the system in which it is installed. The primary purpose of such a 1:1 choke, is to reduce common mode current on the outside braid of the coax. That is usually a good thing whether the system is balanced or unbalanced. The purpose is to equalize the currents flowing in the two elements of the antenna whether those elements are balanced (dipole) or unbalanced (monopole+counterpoise).
  19. WV6U

    WV6U Premium Subscriber

    Thanks for clarifying the 1:1 balun use, Cecil.

    Two more questions:
    1. The way antenna is configured now, the loop is along the ceiling. Feed point sits about 6 feet below the ceiling and just above the railing to keep wire from the wire fence as short as possible (it is about 6 inches) going to the balun. This means there is six feet of wire from both sides of the loop that drops down to the feed point. Does it make sense to move the feed point to the loop's level? Of course, that would mean a 6' wire would have to run from the fence to the feed point.

    2. What other measurements can I make do get some understanding about workings of this antenna? A RF current meter perhaps?

  20. WB2WIK

    WB2WIK Premium Subscriber

    Best "apartment antenna" I ever had was a tunable loop.

    I hung it from the ceiling over the balcony, using a string, so it was surrounded by nothing but "air" for a distance.

    Worked the world with that. The MFJ "Super Hi-Q" loops work very, very well for such an application; although extremely narrow band (you have to re-tune using the control box every few kHz) they are amazingly effective.

    My first contact from my first apartment in L.A. was with a station in the Philippines who advised I was nearly the strongest signal on the band at the time (and there was a lot of local activity), on 17m CW. I was blown away. Of course, condx do "most" of the work but it's impressive when someone tells you you're stronger than your own neighbors are.
Thread Status:
Not open for further replies.

Share This Page