Terminated Inverted U low noise antenna

Discussion in 'Antennas, Feedlines, Towers & Rotors' started by G4HYG, Oct 10, 2016.

ad: L-HROutlet
ad: l-rl
ad: Subscribe
ad: Left-3
ad: L-MFJ
ad: FBNews-1
ad: Left-2
  1. G4HYG

    G4HYG Ham Member QRZ Page

    I've spent some time working on a low noise antenna that can also be used on transmit.

    After doing some antenna modeling and building a working prototype I made a web page for the new antenna design. I've now added an update to the web page following feedback from users and questions I've been asked about it.

    It's a simple antenna to build and offers wide band HF coverage with low noise performance with a simple earthing or grounding system.

    The web page is:


    Chris, G4HYG
  2. PA5COR

    PA5COR Ham Member QRZ Page

    Interesting to read up, since i'm most on 160 with the inverted L i use an homebrew amplified receive antenne to reduce the noise, i used the EWE antenna as well, which in my situation didn't do as i hoped.
    I tried loop antenna's amplified and non amplified, in the end i use the horizontal FD-4 with added coil and 30 feet of wire which is then also resonant at 160 as well.
    Switching between 3 antenna's mostly gives me one that will do the best on the edge of the town.

    Might give it a try to use your design to bring down the wire from the L and terminate that, to see what that does, the L is 77 feet high, 77 feet wire to the side.
    3000 feet of copper wire in/on the ground, so that will do, 17 10 feet long ground stakes used in that net.
    Thanks for sharing ;)
  3. AI3V

    AI3V Ham Member QRZ Page

    What property of your antenna allows it to receive a signal and reject noise?

    And exactly what type of noise can it discriminate against?

    K2XT and KD6RF like this.
  4. G3TXQ

    G3TXQ Ham Member QRZ Page


    I wonder if you've measured the power dissipated in the terminating resistor in order to assess the efficiency?

    I modelled it quickly in EZNEC, and with zero ground losses or balun losses I got the following load dissipation figures with 100W applied:
    160m: 99.2W
    80m: 94.3W
    40m: 64W
    30m: 58W
    20m: 63.9W
    17m: 62.8W
    15m: 49.3W
    12m: 43.7W
    10m: 48.2W

    The multiband terminated inverted-V version seems quite common; the inverted-U less so, although I seem to remember seeing it in one of my ARRL antenna books. I'll see if I can find it!

    Steve G3TXQ
    Last edited: Oct 10, 2016
    KA0HCP and KD6RF like this.
  5. KD6RF

    KD6RF XML Subscriber QRZ Page

    Chris - I'm wondering if you've measured the Signal to Noise Ratio (SNR) with your antenna of various stations and compared it to measured SNR using another standard type antenna with a similar radiation pattern???

    I can see why the noise (and signal levels) would be lower with your antenna because of it's inefficiency, but it's hard to imagine an SNR improvement beyond the few dB of directionality that your antenna provides (and as already posted on the fB group).
    K2XT likes this.
  6. K4SAV

    K4SAV Ham Member QRZ Page

    Solutions for local noise problems are often different than solutions for atmospheric noise. There are only a few ways to reduce a local noise problem. The first is to use an antenna that has a null, and center the null on the noise source. This solution usually only works for a single noise sources or noise sources from only one direction, and for signals that are not in the null. There will be a signal to noise decrease if the signal happens to be in the null of the antenna and there is atmospheric noise present.

    The second is to use an antenna that has a narrow forward lobe and center the lobe on the station. These solutions only work for a station that is in the lobe of the antenna. Signals outside the lobe will results in a signal to noise decrease.

    Another way is to move the antenna away from the noise source. This sometimes works if the noise source is very close, but this solution is very specific for each station having this problem.

    Another way that sometime works for a local noise source is to use an antenna that is cross polarized to the noise source. On the low bands most local noise sources that are from a few wavelengths away are predominately vertically polarized because the horizontal component of the noise gets attenuated by the earth. That means the antenna usually must be horizontally polarized.

    Your antenna is vertically polarized on the low bands and tilted polarization on the higher bands. That means it's unlikely to make any improvements due to cross polarization.

    Your antenna is an EWE but with dimensions a little larger than normal and it has not been optimized for best front to back. On 160 it does have about 9.1 dB front to back (less on higher bands), so some improvement in signal to noise is possible for noise sources to the rear of the antenna. A single element EWE should have about 17 dB front to back, so more optimization should produce a better S/N. For receiving from all directions multiple antennas are required so that the signal can be placed in the forward lobe of the antenna.

    Small and simple antennas with almost the same performance include the EWE, K9AY, flag, delta, pennant, and diamond. The next step up, multiple element versions of these can make a huge increase in S/N.

    I wouldn't recommend transmitting on this antenna. On 160 the gain should be about -21.5 dBi.

    Jerry, K4SAV
    KA0HCP and KD6RF like this.
  7. G4HYG

    G4HYG Ham Member QRZ Page

    Quite a few questions there.

    On antennas mounted in a suburban or urban environment most interfering noise comes from very local sources either directly radiated as an electric or magnetic field within the near field of the antenna or via common mode noise conducted from the mains earth via the feeder.

    This antenna uses the two ground rods as clean low noise reference points and the current balun as an isolator to reduce the common mode noise current trying to flow into the earth rods via the coax cable.

    I've also tried this design using a counterpoise as a substitute ground and found that it acted as an effective antenna picking up local electric field noise. I've also tried running a wire between the two ground rods and while this increases the antenna efficiency it does pick up a low level of local noise from the magnetic component of the noise. Probably adding more ground rods along the wire will reduce this. I will try this next.

    The losses in the resistor appear to be lower than calculated even with the wire between the ground rods. The terminating resistor used is a 390 ohm Arcol thick film 50W rating resistor mounted in a small diecast box. I've tested the antenna using an Icom IC-M600 marine radio on JT-65 running a measured 150W output on all bands down to 80m. From the losses calculated in EZNEC I would expect the resistor to be very hot after a CQ and several QSOs. It was barely warm to the touch. This was with the wire between the two ground rods. I'll measure the temperature rise with an infra red thermometer and replicate the same temperature rise with a DC power supply to give a more accurate indication. It will be weekend before I have the time to do that.

    Regarding signal to noise there are no definitive measurements yet as had to remove my other antenna to fit this one. When it's tested with a SDR-4+ receiver the usual background tones, wobbly carriers and other general background racket have gone. I still see other lower level noise sources in the far field of the antenna but there is nothing I can do about that apart from move to a beach.

    I've tried this antenna in two separate suburban locations and found that it works in removing the local noise. I released the article over two weeks ago and others have now tried it with most of then quickly running a wire from the end of an Inverted L antenna to ground via a resistor and they report the same effect on noise.

    It's interesting that the original motivation for this was to try and improve results on the local 160m club net especially as a few of the members were ready to stop operating on HF due to the noise. It's improved my signals on transmit and receive and brought one member who had already given up back on to 160m.

    Obviously if you operate in a nice location miles from the neighbors there are far better antennas to model and use. Unfortunately most of us are not in that position and following the improvement it's made to my operating (my noise level was previously S9 on all bands) it works for me. I offer it as an option for those in a similar situation.

    Chris G4HYG
  8. K4SAV

    K4SAV Ham Member QRZ Page

    I think you will find that a wire connecting the ground rods makes little difference except for a small change in the best value of termination resistor and whatever noise the wire picks up. I have tried that on EWE antennas and found no improvement. EZNEC also says there is no improvement.

    One way to reverse the direction of this antenna is to add relays at each end which swap the load resistor with the coax that takes the signal. That essentially turns the antenna around. Power for the relays can be run down the coax. You have to be careful with those coax lines on the ground and decouple the shield so that noise picked up on the shield doesn't get into the antenna. It also helps if you bury the coax.

    Adding counterpoise wires can also pick up signals and destroy the antenna pattern if the wires are not perpendicular to the antenna. The only reason for doing this is if your ground is so bad that a ground rod has a very high impedance. A normal ground rod is sufficient for this antenna.

    If you have a local noise source in the near field of the antenna, it will be very difficult to get rid of it by relying on antenna pattern. Moving the antenna might help or it might make it worse. The near field response of the antenna is not the same as its far field pattern. It's best to try and eliminate the noise source if it's that close.

    Its good that you have a common mode choke on the feedline because these low gain antennas are very sensitive to common mode noise. If you need more rejection, adding another ground rod and connecting only the coax shield to that rod, then the choke, then to the antenna ground rod. This is the standard way I decouple long coax lines from low gain receiving antennas. The feedlines to my receiving antennas are about 150 and 635 ft. The extra ground rod also helps protect the circuits from induced lightning currents, which can be very large on long feedlines.

    The amount of power dissipated in the load also includes the impedance of the ground rod because that is in series with the load resistor. There is also some power lost in the ground rod on the coax end the antenna. If this antenna has improved your signal on 160 meters then either you had an extremely poor antenna to start, or there is something wrong with the connections on this one. The gain of this antenna on 160 should be about -21.5 dBi (making some assumptions about the impedance of the ground rods), and that gain is at a fairly low elevation angle.

    Sometimes for local communications cross polarization can cause loss of signal.

    Jerry, K4SAV
  9. G4HYG

    G4HYG Ham Member QRZ Page

    Thanks for the reply, Jerry.

    At present I have the antenna running SW to NE fed at the SW end so on the higher bands the two main lobes go NW and E to the US and Europe from here in the UK. The single lobe on 160 and 80m goes SW. I'll look at a possible switching arrangement. It's easy enough to swap around the balun and resistor for a test.

    I like the tip about grounding the coax shield before the balun. I would suggest that the ground rod for that is placed at least a metre or two away from the antenna ground rod so that the RF noise current has space to dissipate into the ground. If you don't mind I'll add that to the next update on the web page attributed to you of course. It's on the list of things to try with this antenna.

    I agree with your comments about the gain on 160m, I recall 4NEC2 came up with -18 dBi so it's not exactly the most efficient antenna out there.For most hams trying to put a 160m antenna in a small space gains are probably similar or even less when losses in loading coils and tiny ground systems are taken into account. For the local across town 160m net it's fine especially with the low noise on receive.

    Chris G4HYG
  10. SP3L

    SP3L Ham Member QRZ Page


    you mentioned that you used 4nec2 for modeling the antenna. I had downloaded your model from your webpage and ran the "Far Field pattern" plot. That's what I got for 3.6 MHz:


    Radiation efficiency: 1.866%.

    But I am afraid this is not the end of bad news. I ran the same plot but this time I ticked the box "Run Average Gain Test". And this is what I got:


    The 4nec2 marked the result of the test in red (-1.5 dB). This means that the model is completely not credible. The reason for that is: one of the wires in your model (tag no. 2) has radius 0.002 m while all the others: 0.02 m. NEC2 engine does not like such huge difference of radius in the connected wires. When I changed the radius of this wire to 0.02, the AGT went down to -0.1 dB what is a very good result.

    However, as far as I know the NEC2 engine does not work very well when you put wires very close to the ground. Due to that, the author of the 4nec2 added a possibility to use MiniNEC ground in modeling. With this type of ground you can connect a wire to ground (Z=0) and get a valid simulation results.

    I checked the antenna efficiency for two models:
    G4HYG_Terminated_Inverted_U_corr1.nec (radius corrected)
    G4HYG_Terminated_Inverted_U_corr2.nec (radius corrected, bottom horizontal wires removed, vertical wires connected to MiniNEC ground)
    And here are the results:
    The efficiency is reduced due to losses in the resistor and losses in the ground.
    It does not look good for the low bands. It looks better for higher bands but if you simulated the radiation patterns you noticed that this antenna radiates in different directions in different bands.

    I am attaching the models.
    Jacek, SP3L

    Attached Files:

Share This Page