Started building a new GROUNDMOUNTED 80m 4 square

it is not because I do not believe in elevated radials or folded counterpoises (FCP) but after 23 years being a ham and 16 years working with an elevated 80m 4 square I have the chance to build a ground mounted 4 square.

Elevated systems have their pros and cons.....After all I didn't want to take care of TWO resonant structures per vertical..... sometimes the wires brake, sometimes (elevated) radials changed resonace because of change in the near sourrounding....etc.......Also elevated radials/FCP have their risks when not installing them carefully.... with ground mounted radials it is much less critical.......and: I bought the neighbors property, so more room to play around. And the verticals attached to the house made noise.... so: here we go again.... 4 Square 2.0

Spacing of the Verticals in Meter - I designed mine at 3650khz

kHz wavelength quarterwave Diagonale Diagonale/2
3500 85,71 21,43 30,30 15,15
3525 85,11 21,28 30,09 15,04
3550 84,51 21,13 29,88 14,94
3575 83,92 20,98 29,67 14,83
3600 83,33 20,83 29,46 14,73
3625 82,76 20,69 29,26 14,63
3650 82,19 20,55 29,06 14,53
3675 81,63 20,41 28,86 14,43
3700 81,08 20,27 28,67 14,33
3725 80,54 20,13 28,47 14,24
3750 80,00 20,00 28,28 14,14
3775 79,47 19,87 28,10 14,05
3800 78,95 19,74 27,91 13,96

All started with digging the holes for the bases


Antenna mounts - prototype

I welded a steel frame which is mounted to the steel post by 4 screws.
A POM (insulating material) plate will center the alloy pole and isolate it from the steel post and so from ground.
The POM plate is 19cm x 22cm with 1cm thickness. The upper insulator is made from TWO plates, the lower one is TWO plates with holes and a THIRD one without holding the complete vertikal 30cm above ground.
The alloy pole is 8cm OD with 5mm wall thickness - well, it's very heavy and NOTHING for portable operation :-)
Therefore no need for guy wires. At the end ALL verticals will be full size 1/4 wave high. Using toploading was not because of mechanical problem but trying to keep the visibility low. But the topload is more "visible" than the "fulsize"  :-)

Dieses Element verwendet noch das alte Contao 2 SRC-Format. Haben Sie die Datenbank aktualisiert?

Dx-Engineering 4 square Box

I purchased a DX-Engineering phasing unit. It is the older model with the metall cover. The circuit board and all components shall be the same as in the later "B" modell with "plastic" top cover.....

The big advantage of the DX-E unit is a.)  Omnidirectional-Position  and b.) Hot-switch prevention......


tube for the feedlines

to protect the coax from damages and water, I buried a tube. Feedline and control cable for CW/SSB switching is fed through . A ground wire is in the same trench but not into the tube. Just to ensure to have an identical ground potential at all antenna points.

Feed points with SSB / CW / left floating switching

I use a coil to switch from SSB to CW part. One can use a single relais to short out the coil  but I preferred to use TWO relais. In this case I have THREE modes:

1. direct feed (SSB)
2. coil inserted (CW)
3. vertical antenna part is NOT connected (left floating)

while mode 3. the vertical is not re-radiating RF into nearby K9AYs, Beverages or other antennas (especially into my nearby 40m 4 square)

also while mode 3.: I can measure one antenna while the other 3 are "left floating".

In this prototype I used thick solid copper is better to use copper braid in combination with vacuum relais. with unflexible connection, the glass body can brake in difficult weather (temperature) conditions. I will change this soon. btw: vacuum relais are not really needed in this case. quarterwave verticals are current fed. I just like to know that the connections are weatherproof. Hopefully the used vacuum relais are also current proof.

Toploading or fullsize ? decide yourself...

Frederik and I went to buy some alloy tubing.....I wanted to replace the 4-spoke toploading with a 6 spoke-toploading. each spoke is still 140cm long. I ordered alloy wire and will add a perimeter wire which will increase the capacity  even more. maybe even change the design again to a longer spoke version.....but maybe keep them fullsize...not sure yet....there is no mechanical need to use either version.....

The thre lower alloy poles are 15m tall in total. Plus the 6 spoke capacity hat ( 6x 140cm ) I was resonant at 3800 khz. Maybe when I add the perimeter alloywire around the 6 spokes I will be at 3750.......

Remember: when building a 4 square set the resonace frequency of each single vertical 100khz lower then the desired operating frequency of the 4 square..... so one should try to get two (switchable) QRGs.....    3450khz will yield an optimum in the CW portion (3550) and the second (in my case) will be set to 3670 which will later result in a operating QRG of 3770....

one can build the antennas too long and shorten them with inserting a capacitor, or build them "too short" and add a coil or linear loading device to switch from SSB to CW.  With tube elements you may also set your center frequency to 3550 which will later be an array center frequency of 3650khz. this is too short for CW and too long for SSB. then you have no optimum because DX is mostly around 3790 or 3520.....Remember that you would also have to change the length of the feedlines to make it optimum....I decided to cut my coax for 3650.

Radials and Bus wires

A vertical needs a lot of radials. The optimum is 120 radials per vertical. 60 is very close to the optimum. The worse the ground conductivity is, the more radials are needed. The more radials you have, the more the lossy ground will be shielded. The radials collect the return currents. they are "the other half of the dipole"....

in a four square the radials would overlap. For that reason you install BUSWIRES where you connect the radials where the cross. they shall not overlap.

I got this wire from a junk yard. some have already lugs mounted at the end. I will use them as bus wires..... good idea to look at a junk yard. You save a lot of money.

there is a mistake in the buswire picture....the left and right buswire is missing...sorry....

Idea for installing radials: cut the grass as short as you can. lay  down the radials and secure them with hooks. use old clothhangers bend them in u-form, nail down the a year or two they are "in the grass" so you cann cut the gras with your lawn mower..... be careful there are no kinks the blade can is a mess when a wire gets into the rotating blade of a lawn mower... :-)  trust me.

Cutting the four 1/4 wave feedlines

Use 75 Ohm Feedline instead of 50 Ohm feedline.
Cables with 0,66 velocity factor will be too short ! You need e.g. 0,81 or better velocity factor, see list below.
Possible cable types:

PRG11 CU  ( FOAM!!  those with no foam dielektrikum will mostly have 0,66 VF, be sure you order the right one!)
you will need: e.g. PRG11 CU FOAM


Sources to order:

in Germany:
DL6LAU (VP6DX)  they offer plug & play cable (see their web-shop), incl. PL/N plugs and ferrite beads, tested/measured. See their protocol, excellent work! 4 pictures below.   cable with plugs, plug & play, cut to specified length but not measured,  I guess no ferrite beads available  only 100m spools gibt es PRG11CU von Belden, das entspricht RG11 hat aber VK 0,81  only 100m spools

in USA:   incl. ferrite beads !!!! plug & play cable, tested, measured

Dimensions for Coaxcable and radiators

      velocity   factor of the coax cable going from center to each vertical  all dimensions   in METER
kHz wavelength quarterwave 0,66 0,8 0,81 0,82 0,83 0,84 0,85 0,86 0,87 0,88
3500 85,71 21,43 14,14 17,14 17,36 17,57 17,79 18,00 18,21 18,43 18,64 18,86
3525 85,11 21,28 14,04 17,02 17,23 17,45 17,66 17,87 18,09 18,30 18,51 18,72
3550 84,51 21,13 13,94 16,90 17,11 17,32 17,54 17,75 17,96 18,17 18,38 18,59
3575 83,92 20,98 13,85 16,78 16,99 17,20 17,41 17,62 17,83 18,04 18,25 18,46
3600 83,33 20,83 13,75 16,67 16,88 17,08 17,29 17,50 17,71 17,92 18,13 18,33
3625 82,76 20,69 13,66 16,55 16,76 16,97 17,17 17,38 17,59 17,79 18,00 18,21
3650 82,19 20,55 13,56 16,44 16,64 16,85 17,05 17,26 17,47 17,67 17,88 18,08
3675 81,63 20,41 13,47 16,33 16,53 16,73 16,94 17,14 17,35 17,55 17,76 17,96
3700 81,08 20,27 13,38 16,22 16,42 16,62 16,82 17,03 17,23 17,43 17,64 17,84
3725 80,54 20,13 13,29 16,11 16,31 16,51 16,71 16,91 17,11 17,32 17,52 17,72
3750 80,00 20,00 13,20 16,00 16,20 16,40 16,60 16,80 17,00 17,20 17,40 17,60
3775 79,47 19,87 13,11 15,89 16,09 16,29 16,49 16,69 16,89 17,09 17,28 17,48
3800 78,95 19,74 13,03 15,79 15,99 16,18 16,38 16,58 16,78 16,97 17,17 17,37

OLD! elevated system no longer exists! 4 Square info by DL2OBO


First of all, I am not one of the big antenna experts. I read a lot, some things I understand, some not. Understand this page like my "note pad". I write things down here so I can come back and read again. This is an essence of what I have learnt over the last 20 years.Be carefull with the details, there might be mistakes. If you are an antenna expert, please: Let me know if there is something I have to correct.

My 4 Square Summary!




I bought the field marked with 108 in 2011. I started with the 4 square on 109/3 in 1996. So two elements were right on the boarder and the other to on the wall of the building.So one can errect a 4 square within a footprint of 20m x 20m. (plus a bit for the FCP radials ( 5m going outwards)


1. A (classical) 4 square consists of 4 quarterwave long verticals, spaced a quarter wave in a four corner arrangement.
300.000/3750khz = 80m wavelength / 4 = 20m = quarterwave
diagonal/Across = a^2+b^2=c^2 -> sqrt(a^2+b^2) = 28,28m across or 14,14 m radius

2. All elements are fed at the same time, (with different phase angles)

3. it fires a) through the corners ( 1 leading element + 1 lagging element + 2 elements combined as "a fat element" in the middle or b) "broadside ( 2x2 arrangement, 2 pairs of phased verticals)

4."classical" phase angles are -90 degrees (lagging), zero degrees (the two elementin the middle) and +90 degrees (leading). Hybrid - this is was Comtek offers.

5. for better/best performance you may build your own phasing network to adjust phasing angles to a maximum forward gain or front to back or a combination. att: you may need a more complex switching unit
because the beamwidth is now narrower than 90 degrees, remember, that more components can cause more failures. One who has an excellent setup with great success is Paul. Look at PA0GMWs 4 square, which is in my opinion the best developed one in EU (individual phase angles, plus terriffic QTH (salt march land, good ground, flat terrain)

6. The network/hybrid is located in the center of the square. I use 3*lambda/4 RG59 75 Ohm coax so I can bury the coax and have the Hybrid in the shack, that is easier to maintain, check, .....
I don't feel I have too much additional losses because of the longer Coax. One can choose better coax (fewer losses) than RG59. I had plenty of it and it was cheap, so I used it. Maybe one day I will change to better cable or use quarter wave long coax and put the Hybrid in the center of the square.

7. The feedlines to the verticals
You can use 50 Ohm or 75 Ohm cables for the feedlines. As far as I understand, it is better to use 75 Ohm cables, but 50 Ohm cables will also work. At another site was said, use 50 Ohm when ground radials, use 75Ohm when elevated radials are used. You need a current balun to prevent the coax from radiating when you are using elevated feed points!!! Do not try without current balluns! You can order them at or and some other shops. I got mine here in DL from dx.wire. They also have telescoping fibreglass masts for the 4 square and ropes,......

There are 4 feedlines to the 4 verticals. (all fed array) with a length of 90 degrees / quarterwave.
ATTENTION you need this in ELECTRICAL length, not mechanical! So do not forget the VELOCITY FACTOR of the Coax. 300.000 / 3750khz = 80m / 4 = 20m * VF(velocity factor) = 16,60m with Aircell7 (0,83 VF)
RG213, RG58 and RG59 mostly have 0.66 VF. so the cables are 20m * 0,66 = 13,20m long. oooops, that is TOO short. you will not reach the verticals (see fig. 1 where the radius is 14,50m!)

You need another coax e.g.: Aircell7, compare:

RG213 or RG55 or RG59 20m * 0,66 = 13,20m long TOOOOOO short!!!
Aircell 7 20m * 0,83 = 16,60m long enough!!!!


If you have no chance to get Aircell7 or want to use the very cheap RG59 or RG58, you may use
3*quarterwave lambda cables which is 300.000/3750=80/4=20*3= 60m * 0,66velocity factor = 39,60m long cables (you need 4 of them, one for each vertical)

The best choice (quality) would be Aircell7 rather than using a very long coax (3 wavelength RG58/RG59/RG213). If you want the last fraction of a db, go for a cable with 0,8x velocity factor.
For some reasons you might want to use "longer" cable. In my case, I can hide the cable better and my garden ( where I have guests of my hotel) looks a little nicer.

For cutting the right length, please calculate the cable length PLUS about 2m more and then measure properly.
However, when I calculated and then measured the cable I always came very close to the calculated length.
Most important is, that they have the same ( electrical ) length. Normally, if you take all cable from the same spool, they are also the same mechanical length.
Do not use cables from different manufacturers or model/make. Everything you build (vertical, feedlines, loading structures, ) should look the same.


YOU NEED FERRITE CORES on the feedlines at the 4 verticals to let the radial and vertical resonate and NOT the feedline! (current balun)

8. The Verticals

there are many different types of antennas forms you can use. Classical and best is a fullsize vertical which is 20m tall (for 80m / 3750 khz). It can be a freestanding or guyed aluminium pole or a wire attached to a tree. you may use sloping verticals hung from a tower (the more vertical the wires are, the better the performance). you can use loaded verticals, the must not be fullsize! It depends on the loading form how effective the system will work. Try to use top loading, like an inverted L. If you have to use coils, don't worry about "lossy" base loading too much. If the verticals are not shorter than say 10 or 12 m and the Q is higher than 300 you will not mention any big difference compared to a fullsize vertical. Please read DF6SJ Gerd Janzens book "Monopolantennen und Vertikalantennen). it is in German language but the diagramms are easy to will soon learn that antennas with well chosen compromaises must not be too bad. At least the are worth a try!

he claims, that a single(80m Band) base loaded (8,8uH coil) 15m tall vertical with 4 only 15m long radials (elevated 4m above ground) is 95% effecient (only 5% power lost into heat in the base and radial coil) -0,14 gain dBi and 23degree take of angle.
isn't that worth a try ? 15m tall ? just a few radials ( if you have enough land: USE GROUND RADIALS PLEASE!!!) The losses come mostly from the base loading coil not from the radial loading coil. Use toploading and enjoy even more.

A 15,6m tall vertical just needs 6 x 1m long top radials (top loading) to be resonant on 80m. 0,14 dbi Gain at 22deg take of angle
(4m elevated above normal ground) If you connect the outer tips of the top loading you can even make the vertical a bit shorter.

If you want the last fraction of a db and can not use fullsize elements, use Inverted L elements or even better T elements because then the high angle radiation of the horizontal wire will be cancelled out.
"Fat" aluminium poles are better (broadbander) than thinner copper wire verticals. But they are more expensive and takes longer to construct.
You can NOT use grounded towers with omega or gamma match. This will not work! (or at least you will get mad before you get the phase angles right,)
When you build your vertical, try to build them all the same dimensions! Same length, same loading, same coils etc! Later you can fine tune the resonance frequency with the length of the radials. Also CW/SSB band part switching is done by inserting a biece of wire (stub) or coil in the RADIAL. (much easier to construct and maintain)

Whenever I could use ground radials, I would use ground radials. Plenty of them. The problem with elevated radials is, that they can interfer/interact with other antennas / radials around, which is not the case with ground radials. If you go on dxpedition, have a QTH on the beach: 1 radial is enough, 2 are better because of NO horizontal radiation.

9. The Radials
Maximum performance: use 120 quarterwave long radials, under EACH vertical. This is what is tested and calculated the "best" ground. Use broad bus wires to connect radials, where they cross each other.
A 4 square will work also with less (ground) radials. Remember: the better the ground the fewer radials are neccessary to get maximum performance. Some say 60 radials per element is very good. if you want it perfect, install 120. There is no reason to make more than 120 per vertical.



for temporary use or where ground radials are not possible, one may try with elevated radials. The simplest way is 1 radial per vertical installed "outwards" from the array. Raised 2-6m (80m array). ON4UN has installed his 4 square ABOVE the radial field of his 160m vertical. When you live near the sea or can install your 4 sqaure right at the beach you will not need more than just a single elevated radial or better TWO of them, so the radials will not radiate (unwanted horizontal high angel radiation)


Elevated radials must be resonant. Ground radials must NOT be resonant.
You may even use loaded/shortened elevated radials! Use let's say 4 short radials ( some say, up to two-thirds is no big difference to fullsize) per vertical and use ONE coil (with high Q) per vertical to get them to resonate at 3650MHz.

Fullsize elevated quarterwave radials should be avoided. Use e.g.4 x 1/8 wavelength radials per verical and load them with a common high Q (> 300) coil. So you need 4 coils for your 4 square in total. ( not 1 coil per radial)

How shall I build my elevated radials ?

They shall not interact with other radials --- symmetrically, so they cancel the high radiation out
They must be loaded the most effective way if not fullsize --- good coil, or endloading (like a laying L or T). DF6SJ states in his excellent book "monopol und vertikalantennen" that if using a common loading coil with Q>300 you will not notice a difference compared to fullsize element/radials. But use a common coil for all radials per each vertical (= 4 coils) not a single coil per each radial (16 coils)