The forum is read-only currently.

Combining/Diplexing Antennas to Single Kiwi



  • edited December 2019
    I'm not sure about using bicones or the OE-254 which I think was originally designed for 30-88MHz tactical communications.

    The problem is that by virtue of them being balanced antennas each element is 1/2 the electrical length that could be achieved by using an unbalanced antenna of the same height.

    In the past I have played with discone antennas on the VHF and UHF but but the problem with these is that the feed point impedance changes dramatically below the cutoff point associated with the 'disc' size and the radiation pattern tilts upwards as you go higher in frequency. One way to minimise these two problems is to invert the discone and feed the angled elements against ground and the vestigial 'disc'. This gives an improved radiation pattern and the feed point impedance excursion is much more gradual below the electrical 1/4 wave length cutoff of the angled elements.

    I had a quick play with models of various vertical 'cone' structures, and I think something like this, fed against a ground mesh or multiple short radials, may be the way to go in terms of simplicity vs performance.

    It still requires some optimisation WRT to the overall height vs LF performance and angle of tilt vs feed point impedance. It may also be possible to use just three elements and no cross connection at the top of the cone. But I think it should provide sufficient gain on the HF bands and still have a low enough value of mismatch loss to facilitate decent reception on the LF bands.

    A pre-amp and suitable amplitude / frequency equalisation would also be required to maximise the available dynamic range when used in conjunction with the KiWi, but that's another discussion in itself.

    My TC2M (terminated coaxial cage monopole) with pre-amp, equaliser and BC band notch filters still continues to be the antenna to beat, as I'm often surprised (and frustrated) when I try other KiWi's when using the TDoA function, an find that I'm frequently the only station who can actually hear the wanted signal, often by a very large margin.

    The TC2M on average provides about 5dB better Signal to Noise than my Loop on the ground (or just above it), and is about 10 to 20 dB better than a 1m diameter loop.

    These are currently the antennas I've got on test.

    My main KiWi is using a 10m per side loop (just above) the ground and a DX Engineering RPA-1 pre-amp copy.

    Details of the loop can be found on this so far unpublished (I'm still experimenting) webpage.

    I've also got my TC2M broadband vertical antenna (this is the best performer) with a similar pre-amp running on HF Vertical G8JNJ.pdf

    I've also currently got a 1m diameter loop with a similar pre-amp running on


    Martin - G8JNJ
  • edited December 2019

    That conical monopole is something I worked on 30 years ago and yes, it is a good antenna. They use(d) them on shipboard where there were many TX antenna and that small aperture as RX, isolated and split. Marconi and Harris/RF both offered them. Harris/RF also had a horizontal dual "eggbeater" version intended more for NVIS that was mounted on a 10 meter mast.

    I used the OE-254 as a starting point simply because I had one and I didn't want any more than a mast at ground level for now.

    The Marconi was a single 2 meter conical structure. I will try to get a picture of one next week.
  • Incidentally it may also be possible to slightly improve the LF end by incorporating some form of lossy broadband matching LPF filter network as proposed by Fano & Witt et al and latterly by Cuthbert. Book Supplemental Files/23rd Edition/Optimum Lossy Broadband Matching Networks for Resonant Antennas - AI1H - QEX Sep 2008.pdf


    Martin - G8JNJ
  • Hi Jim,

    Yes the Russians had a broadband vertical antenna that they nicknamed the 'Karrot' (or something similar) that looked a bit like this. (note, may flag up a unsafe but it seems OK).

    There is also this sort of variant which requires a lot more space.

    Are you perhaps thinking of the Dieckmann & Klapper MAS 140 shipboard antenna or something similar ?

    "The description is a vertical tapering fibreglass mast, about 400mm in diameter at the base tapering to the top. At the top of this mast, besides a fibreglass whip aerial was a tubular structure with hexagonal form. From this structure were six aerial wires connected to a ring which was about a metre in diameter. This ring then being tensioned with plastic insulators and bottle-screws to ring-bolts into the base of the fibreglass mast. The overall height being 14 metres." "the rated frequency range was 0.2 - 25 MHz"


    Martin - G8JNJ
  • Martin...

    No, not the D&M... the Marconi was a simple icecream cone skeletal structure. I'll have a photo when I get to work and find the one in the archive
  • CM 4 leaf biconical with double wire conductor for upper HF
    SY len=3.0 'cone element length
    SY angle=30 'with respect to z-axis
    SY xdim=len*sin(angle) 'major yardarm
    SY zdim=len*cos(angle)
    SY feed=0.05
    SY feed_width=0.0 'minor yardarm
    SY angle2=20 'WRT yardarm
    SY extension=xdim*tan(angle2) 'Z length to tip beyond cone
    SY wire_radius = .0005 '.020"
    SY wire_spacing = .0254*2 '1"
    SY HAG=2 'Height above ground
    GW 10 30 0 0 feed xdim 0 zdim+feed wire_radius 'element long
    GW 20 30 xdim 0 zdim+feed 0 0 len+extension wire_radius 'element short
    GW 30 30 wire_spacing 0 feed xdim+wire_spacing 0 zdim+feed wire_radius 'element2 long
    GW 40 30 xdim+wire_spacing 0 zdim+feed wire_spacing 0 len+extension wire_radius 'element2 short
    GW 50 1 0 0 feed wire_spacing 0 feed wire_radius 'short1
    GW 60 1 0 0 len+extension wire_spacing 0 len+extension wire_radius 'short2
    GR 1000 4
    GX 200 001
    GW 5001 1 0 0 -feed 0 0 feed wire_radius*5 'source vertical conductor
    GM 1 0 0 0 0 0 0 HAG+len+extension
    GE 1
    LD 5 0 0 0 58000000
    GN 3 0 0 0 4 0.003
    EX 0 5002 1 0 100 0 0
    FR 0 0 0 0 20 0
  • edited December 2019
    NEC2 file from 4NEC2 but the deck should run on most of the flavors
  • edited December 2019

    I'm confused by this thread. Are there two with the same title?
    I guess I don't know what "comments" are...
  • Thanks! I have it up on 4NEC2 as well!

    @G8JNJ For LF to upper BCB how does the loop-near-ground fair to the T2CM? I have a local noise emitter from which I use the Wellbrook loop to try and null out, will be trying out the LOG to see how well I can mitigate noise that way.
  • edited January 2020

    Try it for yourself both antennas are using pretty much the same pre-amp and amplitude / frequency slope correction and BC band notches,

    Loop on the ground (just above it actually)


    I'm logging WSPR using the wsprdeamon on a selection of bands using the TC2M KiWi and reporting with the call of SWUKSDR.

    My noise graphs are here


    The TC2M is slightly more sensitive on the HF bands (approx 5dB ?), but the loop is more resistant to local noise sources. You can hear this around 1-5MHz during the working week, when the adjacent quarry is cutting stone with a variable speed drive motor. I can decrease the noise still further by reducing the height of the loop above ground, but the HF band performance falls off quite dramatically as a result. In comparison a 1m diameter loop at 3m AGL is about 10-20dB worse in terms of Signal to Noise in comparison to the TC2M on the HF bands and is still about 6-10 dB worse on the LF bands.

    I've been comparing various KiWi's around the world (usually when trying to find suitable ones for TDoA runs) and mine's usually the best, even though I say it myself.

    I can clearly hear HF air traffic and Volmet broadcasts from all around the world, the most consistent probably being the Australian marine weather broadcasts VMC & VMW, which can be heard most days on 8176KHz and 12365KHz

    The past few evenings have also been very good on the LF bands, and I've been monitoring USA 75m Amateur nets and frequently find I'm hearing them better than some of the North American based KiWi's, even the good ones :-)

    As an example, one station in an AM net the other night (21:30 UTC) on 3885KHz, had a carrier level that was peaking 60dB above my noise floor, unbelievable for an Amateur station using a dipole at 90ft, although he was running more than a 'bit' of power at the time ;-)


    Martin - G8JNJ
  • Hi Jim,

    Ah OK it's not quite as I thought it would be, but it does look like it would do the job quite well. It would be interesting to know the actual feedpoint impedance, as I suspect it has some sort of matching transformer built into the base.

    The cross connecting rings help mechanically support the structure and control unwanted resonances at the HF end, but you could get away without them for 'normal' use.

    Incidentally, I've just been comparing my TC2M noise plots against those of Clint's Northern Utah KiWi using the big TCI Model 530 Omnidirectional Log Periodic antenna, and I seem to be getting pretty much the same shape of 1dB, day / night noise variation on 10 & 15m which is encouraging :-)

    30m is similar

    But I'm not doing quite so well on the LF bands.


    Martin - G8JNJ
  • edited January 2020

    That conical radiator structure sits atop a ~1M high pedestal base which like the radiator is made like a battleship, literally. That base contains the matching network. If I get time, I'll see if I can find the network.

  • Hi Jim,

    I'm sure I've seen the side mounted eggbeater in a US Navy antenna catalogue, but I've not been able to find it again.

    However I did find this inverted cone RX antenna as fitted to a Kingston Class Canadian Navy coastal defence vessel, but there is no clue as to who produced it.

    Are there any ID plates etc. on the antenna you have seen ?


    Martin - G8JNJ
  • I'll have to check with my shipboard antenna cronie and find out
  • The con ical I was describing is a Marconi R1900
  • Martin,
    I have studied the noise plots briefly on the Grafana site and have a few remarks.
    - I think you are far too optimistic about the efficiency of the 'loop-on-ground' antenna at higher frequencies (>10Mhz). To me , it is very clear, that NEC2 engine calculates not useful gain figures for horizontal antennas close to ground.
    NEC2 calculates about -8dbi for the 4x10m loop at 14 and 21Mhz. I think the real figure might be 10-15db lower at least! That explains, why your judgement of the noise plots of your system seem to be better at high frequencies rather than at low frequencies...
    - The reality is however, that the loop on ground is a very poor performer (on receive) on the high frequencies, due to the very high (but in detail unknown) ground losses. It gets better on lower frequencies, as the signal strength generally rises (signal and noise) with lower frequencies.
    - Note that NEC2 engine shows no difference in gain figure, if you change the height of the horizontal loop from 15cm to the double (30cm). This is not at all the reality! As you know, with the real antenna the output voltage drastically rises (and most likely also the pattern may change somewhat) even with only small changes in height close to ground.
    - I made a test on 10m with a beacon about 100km from here. While my horizontal dipole at 10m height as well as my vertical dipole (resonant on 10m) and the V-beam hears the beacon very loud (about 15db above the noisefloor) the Beverage antenna pointed directly on the beacon, doesn't hear it as good (at noise level). Although NEC2 predict about 9dbi gain on 10m with the beverage at 2m height. The reality is however, that the beverage has at least additional 15db ground loss, not included in the NEC2 calculation, which I since some time include in the noise graphs as additional loss! For the same reason of additional ground losses, I couldn't hear the beacon at all with the loop on ground antenna (using 20db pre-amp at the feedpoint of the LOG-antenna).
    - So, the reason, why the Loop on ground antenna is so quit on 10m is, that it is rather insensitive. Real weak signals cannot be picked up.
    - Why don't you install a well known resonant antenna for real comparison at your same location? Say, a resonant horizontal dipole at half wavelength at clear sourrounding. Such a dipole would have approx 8dbi gain - incl ground gain - and you would really compare apples to apples. Is doesn't make sense at all to compare to other kiwi's, because the (local) noise situation at the other kiwi's location is unknown. I believe the reason for your nice receiving performance has very little to do with your current antennas. The reason for the nice results is, because your qth is rather quit and your antennas are not sensitive enough to hear the (little remaining) noise. That's why I want to encourage to finally build an antenna with gain, so you can really take advantage of your quit location. The reception on the high bands will be far better than today.
    - To hear the galactic noise on the upper frequencies (say 10m), you would need a real net-gain antenna (barefoot before pre-amp!) INCLUDING a quit rural location. Even a resonant dipole at 10m height would need a serious pre-amp. As you know with a large mono-band Yagi turning it in a quit location often clearly shows noise level differences, if the band is just a tiny bit open in some directions.... (difficult to test, as the band is totally dead and all you hear will be man-made noise from distance villages)

    - As Glenn already mentioned, we need to seperate the antenna strategy by frequency: Below 10Mhz we may use a broadband voltage probe antenna, if we can keep it out of strong near fields, and manage the other difficulties such as common-mode, IMD,etc... But above 10Mhz, the higher the frequency the higher the gain must be, to really hear the bottom of the band-noise. In a 9kHz Bandwidth (Kiwi-AMwide) the antenna power at the kiwi-input must be no higher than -110dbm with a resonant dipole in free sourrounding at 10m height (+8dbi antenna). You must then subtract the loss of your feedsystem, matching system/common-mode chokes, etc.... so it is probably impossible without a pre-amp as the S-meter figure might drop another 10db on 10m.... In case you use a pre-amp, you need to install it at the feedpoint of the antenna and subtract the NF of the pre-amp from the gain! On 10m it is easily 3-5db if not more as usually such pre-amps are not UHF designs!
    - the more gain the antenna has by itself (=directivity) the easier it is to detect bottom of bandnoise on 10m.
    - At my qth it is impossible to reach the galactic noise floor with a sensitive system - by far too noisy location - during the week, the noise is at least 10-15db too high (city location according to ITU).

  • Ulli
    Interesting observations, thanks for posting. I'd add from my experience that in attempting to measure galactic noise at ~20 MHz , as the University of New Mexico reported, that horizontal polarization of a dipole is down at > 6 dB from vertical due to ground losses. I would suggest a vertical dipole rather than horizontal along with LNA and normal precautions against CM ingress.
  • Interesting observation Glenn about polatization.
    At least you need the sensitivity to hear it at all. I still believe that elevation angle may play a role too, so I do not know, why polarization should make so much difference.

    When I lived in the Netherlands in the past, at our contest station we had a 5/8wavelength vertical (homemade J-ant) on top of a 90feet tower for 10m. Another tower carried a 7ele long boom mono-band Yagi at 90feet also. Both antennas where really sensitive on 10m.
    During contests we used the high vertical to check, if the band opens in any direction to work the rare multi's.
    With all pre-amps on, we tried to estimate, if bandnoise is just coming up a tiny bit higher than at the usual completely dead times while slightly tuning on 10m up and down. Only if the bandnoise was little higher as usual, there was a chance to catch these rare multi's on 10m which are difficult to hear. The omnidirectional pattern and extremely low elevation angle of the high vertical helped.
    However the bandnoise itself was still usually higher on the long boom yagi.
    Certainly both antennas could hear the bottom of the bandnoise...
    (PI4COM contest stations 15years ago...)

  • edited July 2020
    Update - SV1AFN constructed a diplexer for running two antennas on my Kiwi, based on his HF Multicoupler. The cut-over frequency is 12 MHz. My intent is to run the low end with the Welbrook loop, and the high end is slated to try a variety of antennas with an LNA, including Glenn's Biconical, which is ready in the garage. More soon.

  • I'm using a diplexer to combine a Wellbrook loop with a 27m horizontal terminated folded dipole (TFD). The crossover frequency is about 1.7 MHz. (I found the TFD marginally better on 1.8 MHz.) The TFD has notches for the 4, 6 and 7 MHz broadcast bands. This feeds an Elad ASA16 distribution amplifier which feeds the Kiwis.

    When doing TDoA, I often find that Martin's RX is the only other one in the UK that can hear what I'm hearing. (It's 50/50 which is best.)

    The main factor affecting performance is neighbour QRM which varies enormously.

    I attempted to upload a circuit diagram and a photo, but lost my fight with this forum's software and gave up.

  • Neighbor QRM is also my contention - without the Wellbrook's null, I'd have bands in low to mid HF washed out.

    My setup on the low-side (0-12 MHz) is a Wellbrook Loop, with a DistillAM BCB 'knock-down' filter. High side is a 25 foot elevated wire dipole with a converted LNA (DX Engineering RPA-1A) at the feedpoint and a 30 MHz low-pass on the diplexer-side input.

    With this, I'm getting better CB performance but some additional strong, narrow QRM from something likely coupled from or into the barn's AC lines. I also see occasional AM BCB getting imaged into the 18-19 MHz range, fading in and out - I've seen this before when I've run the LNA before.
  • From an earlier message by Ulli, "I think you are far too optimistic about the efficiency of the 'loop-on-ground' antenna at higher frequencies (>10Mhz). "

    My 120ft (36.6 meter) Loop-on-Ground (without pre-amp) is working surprisingly well. Part of the success is, I think, due to the very poor ground conductivity at my location (Atlanta suburbs, SE USA).

    Here is a recent screen shot of Chinese DRM received on 17770 kHz with the LoG:

  • Loop-on-ground, on poorly conductive ground, is effectively a low to moderate altitude horizontal loop..

    It is weird at times to think that the radio altitude of an antenna may be a few tens or hundreds of meters higher than the actual ground level, even if the antenna is actually at ground level..
Sign In or Register to comment.