80 Meter Vertical Array Details
I built a phased array of four in-line verticals for 80 meters. This is sometimes referred to a directional end fire array. I have learned a fair amount from this project and can pass some of my learning's on to you so that you can successfully build a phased array.
The verticals are located at VE6FI, which is a contest station just south of Morinville Alberta along Highway 2. The verticals are almost 0.25 wavelengths high and are 0.25 wavelengths apart. Each self-supporting tower (vertical) has 48 radials, each 0.25 wavelengths long in a circular pattern around the base.
Physically these verticals are mounted on a concrete base that is seven feet into the earth. The base of the towers are a lamp standard which measures 12 inches in diameter. At the thirty-foot level the lamp standard ends and the tapered aluminum sections start initially with two inch aluminum pipe and tapers down to one half inch aluminum tubing at the top. The structure was design on a spreadsheet with some software that initially originated with K5IU (Ham Radio 1988) The software allows you to use material to meet a specified wind load. The base of the towers are insulated from the concrete footing. The radials are 48 wires made up of pairs of #22 insulated wire laying on the surface.
I modeled the array on the antenna modeling software called ‘Eznec’. Modeling allows you to see the antenna patterns that can be obtained with different heights and different spacing. It is a lot easier to move elements in a model than it is to try moving them in the field. With the model you can change the phase and current magnitude at each tower and see how it affects the pattern, gain, take off angle and front to back ratios. The model will also give you the impedance you can expect at the base of each tower. With the model you can actually calculate how long each tapered section should be so that you only have to put it up once. With a sixty-eight foot structure weighing three hundred pounds you do not want to be taking it down to adjust the length! Even if you have a crane, you still do not want to do it. Modeling allows you to have a good chance of success with your antenna on the first attempt.
Matching - Matching these antenna means that you required to match the towers so that you have the correct amount of power going to the respective tower, have the correct magnitude of current into the base of each tower and that you have the correct phase for the current. at that tower. The only two things that are important in phased verticals are the current magnitude and the phase of the current at the vertical base; nothing else really matters except that you should make the structure strong enough to withstand the elements. Matching involves using Smith Charts, and you may remember, that can be a complicated adventure. Not so, as one just has to use the correct software and thus minimize your mathematical mistakes. I use software from ON4UN, which is called ‘Lowband DX’ It allows you rotate impedances, calculate the current magnitude and phase, voltage magnitude and phase and impedance at any point along the transmission lines. It is 2004 and one does not have to struggle with hand calculations using the transmission equations anymore, however, it is best to verify, by manual means and common sense, that the answers generated by the software are correct..
Some of my Learning’s:
 | One does not make the verticals exactly 0.25 electrical wavelengths long rather you make them slightly shorter so that you can make the reactive portion equal to zero. A quarter wave vertical has a theoretical impedance of 36 + j20 ohms and at resonance you want something theoretical like 36 + j0. In effect you shorten the vertical by about 5 %. This percentage depends on your average diameter of your vertical. W2PV discusses this in his book ‘Yagi Antenna Design’ |
| The vertical antenna is always physically longer than 0.25 wavelengths in order to get the electrical length equivalent to 0.25 wavelengths. In my case the actual length is 68 ft to get 61.5 feet of electrical length. This is because the vertical starts out as a large diameter at the bottom and tapers down at the top. If the vertical had a constant diameter the actual length would have been 61.5 feet. |
| When you insulate the vertical from the concrete make sure the materials that you are using are really insulators. Silicon caulking is good and so is fiberglass material. If in doubt put them in the microwave and see if they heat up. If they don’t heat up at 2 GHz you are OK at 3.8 MHz. Ordinary window caulking does not work. |
| There is a lot of debate about radials. From my readings the outcome is that you want copper radials. They do not have to be big fat copper radials as long as you have quit a few of them. Fifty radials per tower is a minimum with the point of diminishing returns being in the 100 to 120 radials per tower range. At the 50 radial per tower you can use #22 wire as the amount of current carried per radial is milliamps and not amperes even when running a kilowatt. Where radials intersect or cross from different towers you join them together with another common wire. You should terminate the radials at the base of the vertical on a copper strap or copper plate by soldering them. Radials that are buried are not any better than radials along the top of the ground. The benefit of elevated radials has not been verified or dismissed. ( A lot of discussions only) |
| The magnitude of currents into the base of the verticals follows the binomial theorem, which would be the ratio of 1 amp each if you have two verticals, 1,2,1 ratio if you have three verticals and 1,3,3,1 ratio if you have 4 verticals. |
| When verticals are 0.25 wavelengths apart you would think that you would phase them 90 degrees apart such as 0, -90, -180 and –270. I have found that the theoretical optimum for maximum gain and best front to back ratio is 0, -116, -232 and –348 degrees. |
| Phasing by putting a 0.25 electrical wavelength of coax in the circuit will not necessary move the phase by 90 degrees. If the impedance is other than 50 ohms it WILL NOT. |
| All your matching and phasing can be done with physical components of capacitors and Inductances. You just do not have to use multi feet of coaxial cable. A symmetrical phi network is a phase stretcher and will not change the impedance. An L network is the most common network to match impedance. |
| Components used in your matching and phasing do not have to be BIG. You are looking at 1 kV ratings maximum but the current through the components will be 3 to 4 amps at kW levels. |
| Most mica capacitors at the 1KV level do not have a current rating stamp on them. To test current capabilities of capacitors connect the capacitor in series with you dummy load, tune out the capacitance reactance with a coil and load your 1 kW amplifier into it. Calculate the current by Ohms law and run 1 kW into it for 5 minutes. Check if the capacitor under test gets warm or not. Air variables can handle a lot of current but just make sure the plates are far enough apart to handle the voltage present. |
| You cannot measure the driving impedance of a vertical – you have to calculate it. You can measure the self-impedance of a tower but make sure all the other towers are above ground when you do the measurement. Having another tower connected will throw off you measurements because of the mutual coupling. YOU HAVE to calculate the driving point impedance, which is the impedance your transmitter will see. The driving point impedance takes into account the currents in all other verticals. |
| Good measurements of resistance and reactance are a must. A MFJ 259B impedance meter is marginal. You will need a RF impedance bridge. The only measurements you require are the self impedance at the base of each tower with all the others above ground, and the coupled impedance which is the coupled impedance when one tower is grounded. With this information you can calculate the mutual impedance between all towers and the driving point impedance of each tower. |
| Building a four square vertical array is easier as you can buy the phasing harness and adjust the verticals for minimum dumped power. |
| The final adjustment on these verticals are done by maximizing the front to back ratio. I expect to achieve a front to back ratio of 35 dB on this array. Time will tell. |
| You can get more gain with phased inverted Vees but you cannot get the low take off angle of a verticals. |
| Fat verticals will have an impedance lower than the theoretical values of a thin vertical which is 36 +j20 ohms. In my case the self-resonant impedance was around the 30 +j0 at resonance. |
| Never try to match of vertical for minimum SWR by adding or subtracting radials. Put in as many radials as you can and then use a simple L match for matching your transmission line to the antenna. |
Verticals are a low maintenance antenna and can be arrange to give you gain at low take off angles in different directions.
The verticals were made from lamp poles plus aluminum pipe and tubing. This
picture shows the lamp standards as I received them and after I cleaned them up
and painted them.
The flange at the bottom is 1 inch thick and is 12 inches square. I used the lamp standards for the first 30 feet and then went to Aluminum pipe and tubing.
T
he
base of the vertical is a concrete pile which is about seven feet deep. At
the surface the pile is 20 inch in diameter but the main portion in the
ground is eight inches in diameter. There is a R bar cage in the concrete with
four 3/4 bolts protruding from the top but attached to the R bar cage.
This
pictures shows the details of the insulation of the vertical from the concrete
base and from the mounting bolts. Each bolt coming out of the concrete is insulated
from the vertical by a plastic tube and silicon material around the
bolt as well as a fiberglass washer on the top before the metal washer and nut.
Notice on the bottom is a 3/16 inch piece of fiberglass material which insulates
the pole from the concrete. The ground straps go below the fiberglass
sheet.
The grounding includes a ground ring which is at the base of each
vertical. The 48 radials are soldered to the copper strap.
The
copper straps are # 20 or so copper and is easily cut with the tin snips.
It is also easy to solder the ground radials to it when it is new and clean.
To make sure the radials were against the ground I just cut the grass short and
used wooden stakes at the end of the wire to keep it slug so it would stay close
to the ground. After a year the ground cover will cover the radials.
One still has to make sure that the lawn mower blade does not dig into the dirt
(radials) when one is cutting the grass/hay.
To make these verticals work you merely have to have
the correct amplitude of current and the correct phase at the feed point to each
vertical. In this vertical array there is no matching at the respective
vertical itself. I merely feed it with 50 ohm cable and do all the matching and
phasing and power distribution at a central box.
At the base to each vertical I have a pick up coil which is a toroid wound with
10 turns of wire that surrounds the 50 ohm cable. This is the pick up for
current and phase which is inserted at the base of each vertical. The output
signal to take by identical lengths of RG59 to the central location. Thus a
sample of the current magnitude and the current phase from the base of each
vertical is brought to the central location mid point in the vertical line up.
From each vertical a length of 50 ohm cable is run to the central location.
The lengths of this coax is not important as long as you know the length.
The matching, phasing and power distribution is all done at the central location
by 'the networks'. Networks are discrete components that are used to produce
matching sections (generally L networks) and phasing sections (generally phi
sections)
The
network are built by following the guidelines provided by K2BT in his articles
Ham Radio magazine in 1983 and by software from ON4UN (Lowband DXing). The
schematic of the networks are shown on the left. Relays K4 and K3 are reversing
relays that allows the array to be switched from end fire SE to end fire NW. In
order to switch the array broadside relay K5 is energized and the pattern then
is switched to the SW to NE.
The box shown on the left is one of two boxes housing the networks. This box houses the relays required to switch the direction of the array. There are a few variable capacitors for doing the final adjustments. The coil values are changed by spreading the turns. This box shows the network for feeding the correct phase and magnitude of current to vertical tower two and three. There is a similar network for feeding vertical tower one and two.
This picture shows the networks being adjusted at the central location of the verticals. I use a two channel scope to compare the phases and amplitudes and peak up the networks for the correct phase and currents. A four channel scope would be nice for tuning the four verticals but we have to make the best use of what we have. I tuned the networks with 100 watts from a transceiver. Once the tuning is complete I remove the sampling units from the circuit and the complete arrangements runs easily at 2 KW with a SWR in around 1.2 or lower. I measured the front to back and front to side with instruments and it was 27 db. The model shows that when all the phases and current magnitudes are on the F/B would be 39 db. (Always something to strive for)