iowagoldlogo.jpg (5987 bytes)




























A multi-band wire antenna that performs exceptionally well even though it confounds antenna modeling software


The design of the Mystery antenna was inspired by an article written by James E. Taylor, W2OZH, in which he described a low profile collinear coaxial array. This antenna covers 80 to 6 meters with low feed point impedance and will work with most radios, with or without an antenna tuner.  It is approximately 100 feet long, can handle the legal limit, and is easy and inexpensive to build.  It’s similar to a G5RV but a much better performer especially on 20 meters.

The W5GI Mystery antenna, erected at various heights and configurations, is currently being used by thousands of amateurs throughout the world.  Feedback from users indicates that the antenna has met or exceeded all performance criteria. The “mystery” part of the antenna comes from the fact that it is difficult, if not impossible, to model and explain why the antenna works as well as it does.   The antenna is especially well suited to hams who are unable to erect towers and rotating arrays.   All that's needed is two vertical supports (trees work well) about 130 feet apart to permit installation of wire antennas at about 25 feet above ground.  

The W5GI Multi-band Mystery Antenna is a fundamentally a collinear antenna comprising three half waves in-phase on 20 meters with a half-wave 20 meter line transformer. It may sound and look like a G5RV but it is a substantially different antenna on 20 meters.  Louis Varney’s antenna, although three half waves long, was an out-of-phase aerial.  Mr. Varney had two specific reasons for selecting a 3 half waves on 20... he wanted a four-lobe radiation pattern, at least unity gain and a low feed point impedance.   The Mystery antenna, on the other hand,  presents a six-lobe pattern on 20 meters, gain broadside to the antenna, and also low feed point impedance to simplify matching the antenna to the rig.   Additionally, the Mystery antenna is designed to work at least as well, on the other HF bands as a G5RV.  In short, the Mystery antenna is a sky wire that incorporates the advantages of a 3 element collinear and the G5RV antenna.

In its standard configuration, a collinear antenna uses phase reversing stubs added at the ends of a center fed dipole. These stubs put the instantaneous RF current in the end elements in phase with that in the center element. You can make these phase reversing stubs from open wire line or coaxial cable. Normally, a shorted quarter-wave stub is used, but an open-ended half wave stub would also work. The problem is that the dangling stubs are unwieldy and or unsightly.

An article written by James E. Taylor, “COCOA-A Collinear Coaxial Array,”  published in 73 Amateur Radio, August 1989, describes a low profile collinear coaxial array.  According to Taylor, when you apply a RF voltage to the center conductor at the open end, the stub causes a voltage phase lag of 180 degrees at the adjacent coax shield. This happens because the RF is delayed by one quarter-cycle as it passes from left to right, inside the coax to the shorted (opposite) end. There’s another quarter-cycle delay as the wave passes back from right to left inside the coax and emerges on the shield at the open end. Add up the delays and you get a total time delay of one-half cycle, or 180 degrees. In essence, the coax section serves two purposes: it provides the necessary delay and provides part of the radiating element in a collinear array.

The first prototypes of the Mystery antenna used the Taylor formulas, which which called for cutting the wires to a quarter wave length using the formula 234/f(Mhz)  and the coax, using the same formula, but applying an appropriate velocity factor. The first version of my antenna worked well on 20 meters but failed as a multi-band antenna.

The second antenna was built with constructed with the coax cut to the same length as the wire.  This was done with the belief that perhaps the coax didn’t behave like coax and therefore the velocity factor wasn’t applicable.  Surprisingly, the new antenna performed exceptionally well on 20 meters, had low SWR and performed just as well on the other HF bands and 6 meters as my G5RV reference antenna.

Step-by-Step Construction

The W5GI Multi-band Mystery Antenna looks like a plain dipole (see figure1 and photo A below) and is very simple to build. 








Figure 1 - Schematic drawing of the W5GI Multi-band Mystery Antenna.  See text
for details on connection of coax sections in center of antenna legs and on length of
of twin lead stub.

Photo A  - Full view of the W5GI multi-band Mystery Antenna with all sections
shortened considerably for illustration purposes.

Builders of the Mystery antenna will need the following materials:

  • 3 wish bone insulators

  • About 70 feet of wire (14 gauge household electrical wire works well,)

  • Sufficient twin lead or open wire to make a half wave section on 20 meters. Window-type 18 gauge 300 ohm ribbon works best.  The Wireman is an excellent source for antenna wire and 300 ohm line.

  • 34 feet of RG8X mini-coax

  • An electrical connector, available from most electrical parts stores, to connect the twin lead and coax

  • Shrink tubing to cover the exposed coax joints

The antenna can be built in less than an hour when you have the above materials. When you’re ready to proceed, perform the following steps:

  1. Cut the electrical wire into four equal lengths of 17 feet.

  2. Cut the two lengths of coax to 16’6” each.

  3. Cut a 20 meter half-wave section of twin lead.  This piece needs to be adjusted by its velocity factor.  If 300 ohm window type line is used with a VF of .91, the total length will be 30 ft.  Alternatively, 450 ohm, solid 300 ohm or homemade open-wire line can be used provided the electrical length is on-half wave on 20 meters.  Actual length will vary, typically between 27 and 35 ft., depending on type and velocity factor.

  4. Trim two inches of braid from one end of both lengths of coax (Item A).

  5. Trim one inch of braid and center insulator from the opposite end of both coax sections (Item B).

  6. Build a 20-meter dipole without end insulators.  Note:  The next two steps 7 and 8 of the construction process involve connecting only the "inner" end section of the coax section to one end of the dipole; the shield is not connected to anything here.  At the other end of the coax section both the coax shield and second wire section are connected to the coax center conductor.

  7. Connect one end of the dipole to the center conductor of the coax (Item A) and cover with shrink tubing as shown in photo B below.

  8. Connect the opposite end of the coax (Item B) to braid AND quarter wave wire section, cover with shrink tubing, and connect to end insulator as shown in Photo C below.

  9. Install the twin lead through the holes of the center insulator (you may have to enlarge the holes) and solder to antenna wire as shown in photo D below.

  10. Connect the opposite side of the twin lead to the coax as shown in photo E below. Almost any type of connection will work provided the connection is stable and sealed properly.

  11. Install the antenna with the center conductor at least 25 feet high. Mine is installed in a horizontal plane; however, others have installed the ‘GI antenna as an inverted-vee and are getting excellent results. 

Photo B  - Connection of inner end of coax section (closer to center).
Note that only the center conductor is connected to the wire.

Photo - Connection of outer end of coax section (further from center).
Note that both center conductor and shield are connected to the wire.

Photo D  - Connection of twin lead to inner antenna wires at center of antenna.

Photo - Connection of twin lead to coax.  Short length of coax section is for
illustration purposes only.  All connections should be weatherproofed with
shrink-tubing, CoaxSeal, or similar.

Table 1 below depicts the typical SWR results for the W5GI multi-band antenna:

Table 1 - Measured performance of the W5GI Mystery Antenna at various frequencies. 
Columns list frequency, SWR (all as a ratio to 1), Resistance (R) in ohms,
and Reactance (X) in ohms.

On-the-Air Performance

On 20 meters, you should expect 3-6 dB gain over a dipole and a 6-lobe radiation pattern with an elongated figure 8 pattern perpendicular to the plane of the antenna. This is typical of a 3 element collinear array.  For a simple explanation of collinear arrays read "Troubleshooting Antennas and Feed lines" by Ralph Tyrrell, W1TF.  On all other bands the antenna performs like a G5RV, which is really a random length dipole on all but 20 meters. M. Walter Maxwell, in "Reflections II, Transmission Lines and Antennas",  aptly describes this phenomenon. Several users report it is possible to use the antenna on 160 meters but you will need to connect the twin lead together at the point where it connects to the coax. On 160, the antenna performs like a Marconi. Those who have used the antenna on 160 say the “GI Mystery” antenna is a quieter receiving aerial compared to other 160-meter antennas.

As for the theory of operation, it remains a mystery. At least three “experts” tried computer modeling the antenna. All three rendered completely different findings.

You will enjoy building a W5GI Multi-band Mystery Antenna!  Many hams has done so and find it to have been a fun project and an excellent performer.                        


(1) Information on this page has been taken from an article published in the July, 2003 issue of CQ magazine.  You can download a copy of the article in Adobe Acrobat format by clicking HERE.

(2) W5GI will build an antenna for a nominal fee. Discount prices start at $65.00, plus shipping, for the W5Gi multi-bander. Mono band antennas cost more because a 4:1 balun is used.  

(4) Dimensions for the mono-band antenna:




Inside wire


Outside wire

Overall length




23' 10"

23' 4"

23' 6"


141 ft 4 inches




17' 2"

16' 8"

16' 10"


101 ft 4 inches




13' 7"

13' 1"

13' 3"


79 ft 10 inches




11' 9"

11' 3"

11' 5"


68 ft 10 inches




10' 1"

9' 7"

9' 9"


58 ft 10 inches




8' 11"

8' 5"

8' 7"


51 ft 10 inches




7' 10 "

7' 4"

7' 6"


45 ft 4 inches










  • The above dimensions are for a dipole hung in the horizontal plane.  They were calculated by using the formula 234/freq (MHz) plus additional length for attaching to connectors/insulators.

  • If the antenna is to be installed an Inverted V, increase all lengths by 5%.

  • Any of the above antennas can easily be used as multi band antennas by eliminating  the 4:1 balun and using open wire/twin lead directly to an antenna tuner

(4) Dimensions for the multi-band antenna:

Inside wire


Outside wire

Overall length

17' 2"

16' 8"

16' 10"


101 ft 4"

This antenna uses a twin lead matching stub instead of a 4:1 balun.

  • Use only 300 ribbon line for the matching stub.  Start with 34 ft 7", trim as necessary to obtain lowest SWR.  

  • Mono-banders with either a voltage or current (preferred) 4:1 balun.

  • This antenna exhibit significant gain only on 20 meters. On all other bands the antenna performs like a G5RV.

Coaxial Balun


This Balun, (adapter from Balanced Line to Unbalanced Line and vice versa), use one section of ¼ wavelength and one of ¾ wavelength in coaxial cable.
This device requires that the electrical length of both sections include the ¼ wavelength coaxial transformer.
The unbalanced impedance value has exactly the same value of the coaxial cable. For this reasons the name is just 1:1 balun but the purpose of this device is to match any impedance value.
The length difference between the ¼ wavelength and the ¾ wavelength sections provides the necessary 180° degrees electrical phase-shift, as required, for ex. from the open dipole radiator.
Since the narrow bandwidth, this balun is well suited for the monobander antenna only, therefore is particularly indicated to couple the radiator of the VHF/UHF Yagi-Uda antenna, but is possible to use successfully in the HF also.
Remember to take account of the electrical length of the coaxial cable, the speed factor (ex. 0,659 for the RG58, but it depends on the own cable speed factor). So please, use your own particular cable specifications.
Instead of the 1:1 Balun, is possible to couple the 50 or 75 Ohm unbalanced Line or with any unbalanced line to any balanced Impedance value. To do it, is enough to replace the two ¼ wavelength pair's transformer with two sections, always of coaxial line, at calculated value. This is better understood from the figure given like immediate explanation as the 1:1 Balun example.

As shown in the balun image, the 1:1 balun is done by the

[ ¼ W L * V c ] coax cable section of Z c impedance;
and the
[ ¾ W L * V c ] coax cable section of Z c impedance;
which is refolded three times in order to obtain the same physical length of the shortest section, this is not the electrical requirement, then it is for practical layout convenience only.

WL is the Wave Length; WL = ( constant light speed / frequency )
Vc is the Velocity constant or propagation speed factor in the coax cable; i.e. the electrical length.

The mathematical simplified formulae for the calculation is simply:

Z c = s q r t ( Z o * Z i ) { Z c is the Impedance value in the coaxial cable }
Z o = ( Z c * Z c ) / Z i { Z o is the resistive impedance at antenna balanced value }
Z i = ( Z c * Z c ) / Z o { Z i is the final unbalanced value seen at the Rx/Tx coaxial line }


Practical Example of 1:1 Balun, ready to use on 144 MHz made by RG58 (for use in QRP only):



Homemade HF Antenna Balun




A balun is a device that is used at the feedpoint of a balanced antenna when an unbalanced feedline is desired to feed the antenna. Balun is a contraction for BALanced to UNbalanced. A common example of where a balun would be desired is at the feedpoint of a dipole antenna when a coaxial transmission line is used. If a balun is not used it is possible for common mode currents to be present on the feedline. The effect of this could be undesireable if the directional properties of the balanced antenna are to be maintained. Since the feedline usually leads into the shack RF could be present in the shack to create RFI as well as the possibility of receiving excessive amounts of RFI from indoor noise sources. It is often found that a balun is not necessary and everything works just fine feeding the balanced antenna directly with coax cable. When this is possible it may be found that the feedline is an odd multiple of 1/4 wavelength. In this case the transmitter end of the feedline is usually grounded and up from this point on the coax 1/4 wavelength or a multiple thereof will appear as a high impedance. When this high impedance point occurs at the feedpoint chances of common mode currents are low. Rather then take any chances it is often recommended to use a balun.

There are several different kinds of baluns. Some provide a 1:1 impedance ratio while others can provide 1:1.5, 1:4, and many other impedance ratios. A 1:4 ratio balun would come in handy if you were feeding a folded dipole (200 Ohms) with 50 Ohms coax. For a 1:1 ratio a balun can be constructed using the feedline itself by simply winding about five turns of the feedline around a 2" diameter piece of PVC. I preferred a 1:1 ratio balun that I could easily move from one antenna to another by simply unscrewing the coax.

My balun uses AWG 12 enameled wire trifilar wound on a 6" X 1/2" piece of ferrite rod. 7 turns are are tightly wound around the electrical tape covered rod. The free ends of the windings are connected as shown below in the schematic. The whole balun is installed in a 10" piece of 1-1/2" schedule 40 PVC pipe. A SO-239 coaxial connector is installed in the bottom end cap with #4 stainless steel hardware. An eyebolt is installed in the top end cap. The antenna post consist of #10 stainless steel hardware mounted on opposite sides near the top of the PVC pipe.


My Balun Schematic


First I drilled all of the necessary holes, including a drain hole in the bottom end cap, and then painted all of the PVC pieces with olive drab paint the protect from the elements. Next the balun was connected to the SO-239 connector and then the pipe was slid over the balun and cemented in place with PVC cement. At this point the balun was connected to the antenna binding posts. Then the top end cap was installed with PVC cement. I tested the balun by attaching a 50 Ohms termination to the antenna posts and my MFJ-259B via coax to the coax connector on the bottom. The 50 Ohms resistive impedance was reflected back through the balun with little reactance throughout the HF spectrum. Since the design was based upon a tried and true design I am confident that it performs as expected as far as choking off currents.

I found this balun really easy to build and should easily handle a large amount of RF power as long as the SWR of the antenna remains low. A purchased balun may only cost a little more then my homemade version but I had the parts on hand and it was fun to build.


Internal View Bottom View Top View Completed Balun








www.klimaco.com main site lot's of links and info.

www.klimaco.net  our new R&D site ufo's ghost paranormal mind kontrol and more.... enter if you dare!

www.klimaco.org hot links express   cool  links links and more links

www.iowagold.com the main Iowa gold info page where to find gold in the USA

www.iowagold.net the Iowa gold R&D pages

www.iowagold.org links and cool gold pages plenty of pix and Iowa detector group

www.poustusa.com lot's of cool stuff here