CORNER REFLECTOR ANTENNAS FOR
UHF RECEPTION AND TRANSMISSION

    
                          William Sheets K2MQJ                                             Rudolf F Graf KA2CWL
Often, a scanner enthusiast, ham, or UHF experimenter needs an antenna that is more effective than a discone or ground plane, but does want to erect an antenna that is overly large and difficult to tune. A yagi antenna can give excellent results over a restricted bandwidth (1% or so) but can be very difficult to tune and optimize, especially at frequencies in the UHF range. This is in part due to the difficulty of constructing a good matching network. Physical dimensions of the matching network must be kept small and a properly balanced feed must be maintained, as any radiation or pickup by the network will spoil the performance of the antenna. Also, element dimensions become small and therefore so do dimensional tolerances. Although there are a number of antenna design software packages that are not too difficult to use, things things such as the element mounting methods, boom material, and boom diameter must be considered. Much work has been done and while there are a number of good designs available, the construction and testing of an antenna of this type is rather involved and requires test equipment not available to many hobbyists. The finished antenna, if not tested, may be several DB below expectations and may not have the pattern that was sought after. Many VHF and UHF experimenters have had this disappointing experience, the authors included. Smaller yagis with 4 to 8 elements are easier but require the same care and testing if results expected in design are to be realized in the finished antenna. Parabolic dish antennas are excellent and not as critical but are large, expensive and not easy to mechanically mount so they can be rotated. Log periodic antennas work well and are wideband but are large and do not have a lot of gain (6-10 db) for their size. Actually, an antenna that has 8 to 10 db gain, is not critical to set up, and has reasonable bandwidth of 5% to 20% would be a good compromise.

      There is a class of antennas, of which the parabolic dish is a member, that consist of a dipole or other radiator and a reflecting surface. They have fairly good patterns and good front to back ratios. Short electromagnetic waves used for TV and radio communications (>400 MHz or <75 cm. wavelength) can be transmitted and received with small, easy to build reflector type antennas. A conducting, plane surface acts as a reflector for a dipole as a mirror placed behind a light bulb would for the light rays. The surface can be solid or be made of screening. The dipole is a half wave dipole fed from a balanced RF voltage source. Practically all the power radiated from the dipole toward the plane surface is reflected back toward the dipole and will add to the power radiated by the dipole if phase relations are correct. But this is an oversimplification, the exact situation being more complex and requiring a long winded mathematical analysis, taking into account the dipole spacing from the reflector, the wavelength, direction from dipole, and the size of the plane reflector, but it serves to show the principle of operation. If the reflector is bent so as to form a corner we have a corner reflector antenna. If the surface has a parabolic, spherical, or elliptical curve, we have a dish type antenna. A surface parabolic in one direction and flat 90 degrees with respect to that direction would be a cylindrical parabola, ect. All these geometrical surfaces are useful for various purposes depending on the desired antenna gain and pattern. The surfaces are generally about one or more wavelengths in extent, for a useful gain to be obtained. At 440 MHz this is 2 feet or more, proportionately less at higher frequencies. If the reflector is bent so as to form a corner we have a corner reflector antenna. If the surface has a parabolic, spherical, or elliptical curve, we have a dish type antenna. A surface parabolic in one direction and flat 90 degrees with respect to that direction would be a cylindrical parabola, ect. All these geometrical surfaces are useful for various purposes depending on the desired antenna gain and pattern. The surfaces are gen- erally about one or more wavelengths in extent, for a useful gain to be obtained. At 440 MHz this is 2 feet or more, proportionately less at higher frequencies. For the hobbyist with an average assortment of tools and some basic mechanical ability, the corner reflector antenna is probably the easiest of the aforementioned antennas to construct. Figs 1 & 2 show the details of this kind of antenna. It is somewhat foolproof, easy to set up, and will perform well. It is easy to build a simple corner reflector antenna, using hand tools and commonly available materials from your local lumber yard, home center, hardware store, or hobby shop. In fact, a number of the giant home centers that are springing up like weeds around the country may very well carry every item you may need, especially if they have a craft or hobby department. Cost should be $20 or less, or next to nothing if you can scrounge around for some leftover cutoffs and scraps of materials. The antenna will provide noticeable improvement over a discone or ground plane, have some directivity, and be well worth the time and trouble to make it. It can be used indoors as well. It will not be useful for moonbounce, weak signal SSB, DX contests or other such exotic amateur radio uses but it will be a darn good antenna for much scanner listening and routine ham use, or as a temporary, cheap antenna to use before investing in a larger yagi or other expensive setup. With a low loss feedline, an 8 to 10 db antenna will give very good results both in transmission and reception. The author has used two of these antennas for amateur TV transmission at 923 MHz, one antenna at the home station about 30 feet above ground, the other in a vehicle with a portable TV set and a receiving downconverter to allow reception of the 923 MHz amateur TV signals on VHF Ch 3. Excellent pictures were received 8 miles away and, although snowy, a picture was seen at 17 miles. A 1 Watt transmitter was used at the house. This is not bad for a 1 watt TV transmitter and simple antennas. The tests were repeated at 1289 MHz with a pair of corner reflectors designed for this frequency and similar results were obtained.

      The corner reflector consists of a balanced half wave dipole placed in front of a conducting surface which has been bent at an angle of 90 degrees or less. As the angle gets smaller, the gain tends to increase but the antenna tends to get larger and the dipole feed impedance becomes lower. The dipole may be constructed with thick elements to increase bandwidth and a bowtie shaped dipole can also be used, this being commonly done in a UHF TV antenna to get wider bandwidth. The dipole may also be made adjustable in length (see figures) to vary center frequency, and the spacing of the dipole from the reflector may also have adjustment provisions for optimizing the feed impedance. A folded dipole element can also be used with an appropriate balun to feed it. The dipole is parallel to the axis of the bend. Polarization is depends on the direction of the dipole axis, being in the same direction. Gain is typically 8 to 10 db (isotropic reference) for a reasonable sized antenna of 1 to 2 wavelengths, but can be made 14 to 15 db or more with a large reflector (>5 wavelengths) and a narrow (45 degree or less) angle. However this may get mechanically rather impractical below 1000 MHz. A dipole is a balanced antenna and cannot be directly fed with coaxial line. Doing so causes the outer conductor to act as part of the antenna and a large amount of signal is radiated or received by the outer conductor. For casual reception this may not matter much, but the pattern of the antenna is destroyed and generally is no longer predictable. In a directional antenna such as a yagi array or reflector type antenna, this is a disaster. It will destroy a sought after radiation pattern and really negate careful design efforts. In fact, a big problem in yagi antenna design is getting a good balun so a truly balanced feed is obtained to the driven element and no trace of radiation or reception from the feedline or matching system is evident. This is important when you are seeking 20 to 30 db rejection of signal in unwanted directions. In order to have a true dipole, balanced feed is a must. In order to derive a balanced feed for the dipole some sort of a trans former is necessary. For an unbalanced input voltage V, where V is the voltage between the coaxial line center conductor and the outer (grounded) conductor we need the same voltage but isolated and balanced with respect to ground reference, or in other words, +KV for one dipole element and -KV for the other. K is a constant which depends on the configuration of the dipole or other load. This can be done in several ways. An actual transformer can be used or the transmission line can be used,see figure 4. One way is to coil the coaxial line to form an inductor to isolate the end of the line from ground. Another is to use a ferrite bead to accomplish the same thing (See fig 4) These methods are not too practical at UHF but work well at lower frequencies (<200 MHz). A quarter wave choke sleeve can be placed over the outer conductor as shown in figure 4. The quarter wave sleeve looks like an open circuit at the corresponding frequency, and effectively isolates the dipole element from ground. Another way is to split the outer conductor lengthwise for a quarter wavelength and connect the inner conductor to the end of one segment, and to one dipole element. The other segment is connected to the opposite half of the dipole. This is known as slot feeding. This type of balun gives a 4:1 impedance transformation and can feed a folded dipole or a simple dipole. We have tried both and were able to achieve a satisfactory match (1.5 : 1 or better VSWR) by adjusting slot length and trimming dipole length, and adjusting dipole to reflector spacing. The folded dipole was ex- pected to match the split balun better but a simple dipole worked just as well. Other stray effects such as slot width, element thickness, and the fact that the feedline diameter is not negligible with respect to the dipole length probably are the reasons for this observed behavior. Dipole to reflector spacing affects the dipole impedance. See Fig 3. However, field tests show that both methods work well and the slot feed method seems a little easier to implement mechanically. It also has the advantage of DC grounding both sides of the dipole. With a slider ring around the outer conductor, the slot can be adjusted in effective length to adjust the match. The sleeve balun is a little more difficult to build, since suitable diameter hardware and insulators to fit the tubing can be hard to find and may have to be made. This is best done on a lathe and is therefore out of the question for many hobbyists having no access to a lathe or a machinist friend. While both methods give good results, unless you can find or make suitable hardware yourself, the slot method is the easier one to use. However you should get equal results in each case.

      The construction of a corner reflector antenna for 900 MHz will now be discussed. For other frequencies, the dimensions can be scaled from those given using table 1 as a guide. At the lower frequencies, the antenna is larger so some compromises as to reflector size may have to be made in order to keep within practical mechanical size, weight and structural stability. A good idea from a performance standpoint is to keep the reflector as large as you can, up to a few wavelengths. This is obviously more difficult at 400 MHz than say 1300 MHz. However, as size increases the extra gain may not be worth the mechanical difficulties and cost. Also, wind loading must be taken into account as this type of antenna presents a large projected area. Designs for lower frequencies use rods spaced about 0.05 wavelength apart to simulate a reflector surface, and these rods are usually 1 wavelength or more long. This reduces wind loading. This method will not be discussed further.

The antenna shown in fig 1 is fairly simple to construct with just hand tools. Figures 5, 6, and 7 illustrate the mechanical details of construction. First, cut a piece of 5/8 OD brass tubing with a .015 wall thickness, which is available at well stocked hobby shops, to a length of about a half of a wavelength at the desired operating frequency, plus another 1 to 2 inches to allow for the connector and mounting flange as shown. Make sure the ends are squarely cut. If you have one, use a tubing or pipe cutter. A small cutter can be obtained at a hobby shop. Cut a piece of 1/4 inch tubing to a length about 1/8 inch shorter than the 5/8 tube. These tubes will form the dipole feed assembly and balun. Cut two slits lengthwise along one end of the 5/8 tubing, on opposite sides of tubing. Use a hacksaw or small slitting saw. The slits should be 1/4 wavelength at the operating frequency. The slit width is not very critical. Next, drill holes for dipole elements as shown. Use a drill the same diameter as the dipole elements for a snug fit. Next, cut a length of 1/16 ot 1/8 inch brass rod about 1.05 wavelengths long at the desired center frequency. Shape the folded dipole as shown in the figure. Try to make the dipole as accurately as you can. We used 3/32" brass welding rod. You can also use #10 solid copper house wire with insulation removed if convienient. In addition, a simple dipole can be made and converted to a folded dipole by adding a jumper made from #12 wire or brass rod as shown in the fig 8. Carefully clean up and deburr cut edges. File the ends of the dipole elements to remove burrs and sharp projections so as not to stick yourself. Next, solder one end of the 1/4 inch tubing flush to the center pin of a type N connector, concentric as shown. If necessary. build up the diameter of the center pin with some bare copper wire as shown, for a snug fit and to ensure concentricity. Place the 5/8 tubing over the 1/4 inch tube, and check to see that the the end of the center conductor is flush with or is slightly shorter than the outer tube. Trim center conductor as needed. If needed, clean connector flange with fine steel wool for later soldering. If desired, add a shorting ring to adjust slot length as shown. This can be done later with a strip of brass or copper if preferable.

       Slip a 1/2 inch copper pipe coupling (use the kind that has no stop) over the outer conductor. The commonly used 1/2 inch copper water pipe is 5/8 OD so a smooth slip fit should result. If OK, remove and drill a hole at one end to pass a #4 screw. A # 33 drill is large enough, but we used a #28 drill to allow extra clearance. Polish with fine steel wool to a bright finish. Next solder a 2 X 2 plate made of copper, brass, or surplus PC board G-10 material to the coupling as shown in figure. Make sure coupling is perpendicular to the plate. Using rosin core solder, solder a #4 brass nut to the coupling as shown. Use a stainless steel #4 screw X 1/4 inch long to hold the brass nut in place while soldering, as solder will not stick to stainless. Remove the screw after joint cools. Next, slip the flange assembly over the 5/8 tubing, the slit end opposite the nut as shown. Clean both ends with fine steel wool to facilitate soldering in next step.
Insert the center conductor and connector assembly into the outer conductor placing connector flange flush against outer conductor. Check to see that predrilled holes for dipoles are aligned. You can insert the dipole elements and fasten them in place with tape, etc to check for correct align- ment. Making sure tubing is concentric, solder connector flange to 5/8 tubing all around the outside. Use as little solder as you can for a neat job, and if possible use a hot 100 watt iron with a 1/4 inch tip. A few wooden blocks drilled with 5/8 dia holes will be found useful for holding parts during assembly and soldering operations. Clean all flux residues with alcohol. CAUTION - FLAMMABLE. Do this outdoors away from flame or sparks. Make sure no shorts are present between outer and inner conductors (temporarily remove dipole elements). Next, insert folded dipole as shown and solder. Note that the outer and inner conductor are shorted together by one dipole element. This is normal. Make sure they are aligned as shown. Measuring from center of center conductor, the dimensions of each side of the folded dipole should be equal and symmetrical.

      You now have a half wave slot fed dipole. Connect to a receiver and check to see if it works as a receiving antenna. It should work as well as your whip antenna or better. Try orienting vertically and horizontally for best signal reception, as signals are generally vertical or horizontally polarized If you have the equipment, use an RF source and a SWR or power meter to check the VSWR. It should be 2.0 or better. The dipole elements can be trimmed for lowest VSWR later. This will be affected by reflector spacing, slot length, and also presence of nearby reflecting objects. If you cannot do this test do not worry. It is nice but unnecessary for receive only purposes. If the antenna appears dead check for shorts from burrs, solder drops, steel wool fragments, or other foreign material.

     Almost any reasonable material can be used to make the reflector. We used .019" perforated aluminum sheet sold in hardware stores for making grilles and covers for radiators, etc, but solid sheet aluminum, copper, wire mesh, or screening can be used. All you need is a conducting surface. Plywood or even heavy cardboard covered with aluminum foil can also be used, if weatherproofing is not needed, or for experimental or temporary use. Hardware cloth is also useful but hard to handle. Referring to Fig 1 the reflector is made by bending a 1 x 2 foot sheet of material as shown. The exact reflector dimen- sions are not critical, larger being better. This necessitates two 45 degree bends but is easier to use from a mechanical standpoint. A piece of scrap wood is used to support the reflector and to mount the dipole in the center of the reflector. The wood can also be used for as a surface for mounting a bracket to hold the completed antenna to a mast or other support. Bracing can be added if desired as shown in the photo. Use wood, plastic, fiberglass, or other nonconducting material. Conductive materials in front of dipole will cause detuning and pattern distortion. If thin sheet metal is used alone for the reflector it is wise to cut the metal 1 inch larger in width and length and use this extra material to form a folded edge around the perimeter to mechanically stiffen the reflector surface. A block of wood can be used to form the bends, as the material is easily worked by hand. After the reflector is formed, cut a hole as shown and fasten the flange on the dipole assembly to the reflector as shown. The dipole should be parallel to the bend in the reflector. Initially set the dipole about 0.3 wavelengths from the reflector. Install a 4-40 brass screw in the nut previously soldered to the flange assembly to lock the dipole in place. Final adjustment can be made later by setting the dipole position for lowest VSWR, with an RF source and reflected power meter or SWR bridge. For receive only applications, no further adjustments are needed. You could try peaking the adjustments on a weak signal if you are fussy, but you will find that they are very broad and have little noticeable effect. Next, mount the antenna in its final location. Make sure to mount the antenna for correct polarization. Polarization is same as dipole (eg. vertical for vertical polarization). Vertical polarization is generally used for amateur and commercial two way FM, but horizontal is used for SSB amateur work and some amateur TV. UHF TV is generally horizontal or circular. As a compromise, the antenna could be mounted at 45 degrees to vertical. You will find that the antenna has pronounced directivity, maximum pickup occurring along a line bisecting the reflector angle, in the direction the dipole faces. The pattern is clean and well defined. Pickup towards the sides or rear is much less. Therefore, face the antenna in the direction of desired reception. Two or more of these antennas can be used if multidirect- ional reception is desired, or a rotator can be used. Bandwidth depends on the VSWR desired, but an antenna as this should work well over a range of 10 percent or so. An antenna made for 900 MHz will easily work well from 800 to 1000 MHz, and reception at 450 MHz and 1280 MHz will be adequate, but not optimum. Bandwidth can be increased by using triangular shaped dipole elements but we have not tried this as of yet. You should find this antenna easy to make and quite effective and may even wish to build several for different frequency ranges. Simply scale the dipole element length and reflector size as needed.

     For outdoor use, it would be a good idea to cover the dipole and slotted feedline assembly with a plastic cover to keep out water and insects. A food container from the local dollar store can be adapted for this. Use a clear plastic container that is microwave safe. If it does not heat up in a microwave oven, it probably has low RF losses and will not affect the antenna. The container can be slit and placed over the dipole and slot (slit facing down). The lid can be used to cover the open end of the tubing. Clear silicone seal can be used to seal edges and joints against leakage. Clear materials are preferred since pigments such as metal dust or carbon can be lossy for RF. You can test for this. Place a material sample in a microwave oven for a few minutes. The material will heat up if lossy. If it stays cool it is OK. Make sure to leave a small hole in the bottom to allow escape of condensation. The construction of a cover assembly is left to the ingenuity and discretion of the builder.


                         List of Materials*

     1 ea     5/8 OD X 19/32 ID Brass tubing (.015 nom wall)  
     1 ea     1/4 OD Brass rod or tubing, ID not critical.
     1 ea     1/16 to 1/8 brass rod, or 3/32 brass welding rod 15" long
     1 ea     Type N, Flange mount, type UG58A/U, or BNC Flange Mt UG290A/U
              RF connector, preferably silver plated
     1 ea     2-3 sq ft material for reflector(.019 perf aluminum recommended)            
     1 ea     4-40 brass hex nut
     1 ea     4-40 X 1/2 brass screw
     1 ea     4-40 X 1/4" stainless steel screw
     1 ea     .032 Brass or copper plate,  2" X 2"
                 or double sided g-10 matl, 2" x 2" X .062
     1 ea     1/2" stopless copper pipe coupling, sweat type
     1 ea     Wood, 1" X 2" X 12" as required.
     8 ea     #6 X 1/2" sheet metal or wood screws
     Also     Solder, 60/40, both resin core and solid, fine steel wool,
              hardware as needed, misc wood blocks for jigging purposes,
              suitable plastic container for optional cover.


  *           Tubing is sold in 1 ft lengths. Hobby shops that sell model airplane
              supplies and large hardware and craft suppliers may also carry this
              material. The amount of materials are shown to build one
              antenna for 900 MHz. Antennas for lower frequencies are larger
              and will require correspondingly more materials. Brass tubing
              for inner and outer conductors can also be any other reasonable
              dimensions as long as they have approximately a 2.3 to 1 ratio 
              of ID (outer) to OD (center). See table 2 for other suitable 
              sizes and resultant impedances. (45 to 55 ohms will be OK)

      Note: This article is the unedited version of the one published in Electronics Now magazine, Aug 1996

           We suggest obtaining a printed copy of the article if more detailed artwork is desired.

           This article is for informative purposes only and intended for experimental applications, and no warranties of any kind are given or implied
           with respect to final performance or fitness for any applications whatever.  

                                   W. Sheets  K2MQJ                                  Orig   6/26/95
                                    Rev A   7/06/95
                                    Rev B  10/05/95
                                    Rev C  08/06/02