The local PBS station in my area is WMEB. In February 2009 they switched over to the federally mandated digital broadcast standard. When doing so they lined up the equipment to focus most of the power in a more or less southerly direction. As a result the town of Orono Maine (my QTH) ended up in one of the deepest nulls in the transmitter's antenna pattern.
Combined with the topography and the distances involved, there is little signal available for reception. Thus, as far as my home is concerned, WMEB went off the air. One possible way get WMEB back on the air at my house, is to build an antenna that can capture more of what little signal may be available. The other possibility was to get cable or maybe "Sat-Cable". Being an antenna tinkerer I preferred the old school method of seeing what could be done using the broadcast signal.
The graphic to the right is a transect from the transmitter sight to my house. Even with an antenna height of about 128 meters the intervening hill peaks would still block the line of sight from the transmitter.
This is a problem because WMEB broadcasts on channel 9, which runs between 186 MHz to 192 MHz. At such frequencies radio waves do not have much of a "ground wave", and they don't "fill in" as well as lower frequencies. Thus anything getting in the way of the signal will more likely deflect the signal away, and it will not "flow" around the obstruction. Given the amount of rock in the way, the lack of a significant signal is no surprise. Indeed so much is in the way, any level of reception may be difficult.
Fortunately there was reason for hope. The next graphic shows another transect to a location up on the step above my location. The fellow at this location was getting intermittent DTV signals with "rabbit ears". This suggested that some amount of the signal was refracting off the crest of Kelly Hill, the tallest nearby peak. With some level signal apparently available the only remaining problem was how to get enough to my television so that it could form one of those amazing new DTV images.
There are many types of antennas that can make more of a given signal, to a greater or a lessor extent. Such antennas are said to have gain, meaning they are more sensitive than other types of antennas. The higher the gain of a given antenna usually implies a greater asymmetry to the antenna's ability to receive or transmit signals. In practice this means the antenna needs to be pointed in an appropriate direction, in order to take advantage of its greater sensitivity. In many cases a high gain antenna will perform worse than so called "rabbit ears" if it is not pointed in the correct direction.
One common commercial antenna with gain is called a log-periodic array. Such an antenna does have some gain but it can get large and unwieldy. It is considered a broad band antenna, since it can provide gain throughout the standard TV broadcast spectrum.
According to wikipedia (Television channel frequencies) the VHF band stations, allocated by the FCC, range from 54 MHz (lower edge of channel 2) to 216 MHz (upper edge of channel 13). A typical high gain antenna isn't usually expected to cope with more than a small deviation from its operating frequency. In the case of WMEB its range is some 6 MHz from 186 MHz to 192 MHz. This amounts to about 3% of the center frequency of 189 MHz. This is about as wide a span that you would want to contemplate for a high gain (greater than 10 dBi or so) antenna. Such an antenna will perform poorly outside of the 3% limit. Some folks might say 1%, some might say 5%.
All VHF frequencies are considered "line of sight", which means if
there is anything substantial between you and the transmitting antenna
you won't get much of a signal. The higher the frequency the more
pronounced this effect.
Generally speaking the more broad banded an antenna is the less gain it will have for a given size. The log-periodic antenna is no exception to this rule. Such an antenna was built but proved to have insufficient gain to change the WMEB situation. Though it did have enough broad band gain to pull in a broadcaster not seen before at my home in Orono. So the antenna seemed to be working well, it simply did not have enough gain with respect to WMEB.
Another type of gain antenna is called a Uda-Yagi. It is a narrow banded antenna and so has much higher gain. It is also much smaller than the log-periodic. The price we pay for these features is that the gain falls off rapidly for signals outside its design frequency. Thus it will not perform well across the entire television broadcast band. It will perform well for one channel though. Fortunately in the Orono area, most of the other broadcast signals have sufficient local power that we end up receiving them anyway. So we can design our Uda-Yagi for WMEB specifically, and still receive most of the other local television broadcasters.
Parts of the log-periodic antenna were scavenged to make the Uda-Yagi. In the above photo the 1/4 inch soft copper tubing elements of this Uda-Yagi design were once elements of the log periodic antenna.
The following sequence of photos illustrate the construction of the antenna.
The above is the "boom" of the array. It is a piece of 1x2 ready for marking and drilling. The length was chosen to support the desired number of directors plus one, in case I needed more gain.
The current number of directors is 6. This seems to be enough for more or less consistent reception of the WMEB signal, with the current antenna height. Unfortunately the antenna does not do so well for the CBS affiliate WABI. On rare occasions, when there is a "lift" on, the station will be received but most of the time the receiver does not have enough signal to work with.
According to the FCC WABI has two channels allocated.
A batten is clamped to the boom to make marking off the element spacing easier. I also used a metric system tape measure. The metric system of measurement is much simpler to use than fractional inches and feet. I learned it in grade school in the early 70's and never looked back. Happily many of the tape measures available these days have both systems on them. This one was picked up in Sweden for me by my wife, so its all metric.
In the next photo the reflector and the driven element have been installed.
Many of the elements are the same length so a block was used as a fence from which to measure the lengths. Then the desired length was marked on the bench with respect to the block. The aluminum wire was then rolled out, straightened and cut to length. It is important to make sure the elements are straight. The straighter the better but as you can see in the photos, I didn't work too hard at making the wire or the tubing straight.
Preferably the elements would be perpendicular in both the Y and Z axes. In the photos you can see the second to last director is clearly dipping down and out of the Z plane. Again the better you can get the alignment of the elements, the more likely the antenna will perform well. As can be seen in the photo the alignment of this test antenna is quite "loose".
As crude as the element alignment is, the antenna should still work fairly well.
Note that this antenna is mounted to a wooden dowel. The dowel is then connected to the steel antenna mast. Horizontally polarized antennas are relatively immune to metal objects appearing in their near field region, if the metal object is discrete enough, and if it is orthogonal to the antenna's electric field plane.
However, mounting a piece of wood to a steel tube, without damaging the wood, can get tricky. Also, the wood mast has been used for testing such antennas in both horizontal and vertical orientation. Thus it was an easy way to implement a test configuration, without committing to a more complicated arrangement, prior to proving the efficacy of the antenna.
With the antenna in the air, any signal on it then needs to get to the television. In the test configuration, I used 300 ohm twin-lead. I used twin-lead because it was all that I had available at the time. The use of twin-lead made things a bit more complicated. The complication arises due to the nature of RF (Radio Frequency) signals. Television signals are RF signals, thus good RF practice should be used when dealing with television signals. Having said that, there is often a certain degree of latitude one can take on such issues, and still expect reasonable results. In this way good RF practice can be used as a guide, rather than a strict set of rules. One RF practice rule that cannot be fudged very far is the need for the correct type of cable (also referred to as feed line), and how it is connected to its source and sink. In the case of a television antenna the source is the antenna, and the sink is the television itself.
All RF devices have a specific impedance. A television, as well as an antenna, are no exception. An RF impedance can be thought of as a resistance to the flow of an RF signal. The theory developed to help us work with RF signals is a good deal more complicated than this simple idea but for our purposes, just think of the impedance as a resistance.
The next thing you need to keep in mind is that in order to make the most of a given signal you need to match the impedances. Every source and every sink has an impedance. Interestingly every cable also has an impedance. In order to make the most out of a system of sources, sinks, and cables, they must all have matched impedances. This can get quite complicated. However, in our case, we only have one source, one sink, and one type of cable. The number of cables is less important as long as they are all of the same type. There is some latitude here but blending different types of cable can also be complicated, we will use some of this complexity later, but it is best to use only one type of cable in order to get the signal from the antenna to the television.
The driven element of the antenna is the squashed loop shape above the single element reflector in the photo below.
Although referred to as "the driven element", in this application no source is used to apply power to this element of the antenna. Whether the antenna is used in a transmitting or receiving mode, the impedance of the antenna does not change. In the simulation of this design, the impedance of the antenna was 105 to 109 ohms. This is a higher impedance than often used in RF work. Typically 50 ohms or 75 ohms is used. It is also a bit lower than the 300 ohms of twin-lead. Most modern televisions, cable TV providers, and many satellite "cable" systems use 75 ohm systems. This antenna was designed for use with twin-lead so an impedance was chosen closer to the twin-lead impedance, so as to make other parts of the system easier to construct.
Shortly after the successful demonstration of this antenna, a couple of lengths of RG-6 cable became available. After suffering some DSL service interruptions, I switched over to Internet service, provided by my local cable TV vendor. The legacy cable TV install, had many feet of "out of spec" RG-6. All the old cable was removed, and replaced. The cable guy was nice enough to let me have the old cable. The designation RG-6 is the engineering name of a specific type of coaxial feed line. The RG-6 cable has a characteristic impedance of 75 ohms. One advantage of coaxial cable is that it is not effected by the proximity of other metal objects. This is not true for twin-lead. Routing twin-lead close to other metal objects, copper water pipe or metal TV antenna masts, can mess up its impedance. Coaxial cable, like RG-6, is not effected in this way. Be that as it may, even at 105 or so ohms there was a mismatch between the antenna and the feed line (twin-lead). This mismatch was corrected by building what is commonly called a transmission line transformer.
Here is a view of the transmission line matching section.
This simple construction is used to transform the feed point impedance of 105 ohms or so, to the twin-lead impedance of 300 ohms. Doing so maximizes the amount of signal transferred from the antenna to the feed line.
The somewhat out of focus picture below shows where the transmission line transformer is used.
The order of the elements is important. The copper pipe driven element is connected to the twin-lead section. Then comes an RG-58 coax section (of 50 ohm character), which would then be connected to the twin-lead feed line as a continuous run to the television. In this photo the twin-lead run to the television has been disconnected, in preparation for replacement with RG-6 cable. The RG-6 can't be used directly with this transformer.
In this next photo the RG-6 has been attached to the antenna using a 75 ohm to 300 ohm transformer, sold by Radio Shack. This is a bit convolved but it serves the purpose. It would be better to match the "driven element" directly to the RG-6 but I had everything built already, and the transformer was laying around unused.
There are a number of ways to match the feed line to the antenna.
Getting back to the transformer... This is not a weather proof part but it can be taped up to survive for some period of time. With these parts arranged in this manner, they should not interfere with the antenna radiation pastern, but also should maximize the amount of signal getting onto the RG-6 cable.
A redesign of the driven element could eliminate the matching section, and the transformer. For the time being the added complexity of the transformer is tolerated, since it allows this test antenna to be connected to either twin-lead or RG-6 cable.
above plot is the azimuth plot from EZNEC. The azimuth plot can be thought of as looking at the radiation pattern from above the antenna. This azimuthal plot is a prediction of how the antenna will perform with the antenna pointing to the right. The pattern will be largely the same if the antenna were pointing in any direction in the plane of the circle. It just so happens that the x direction (left/right in the plot) is common to the axis of the antenna, with the maximum gain in the positive (to the right of center) x direction. The concentric circles in the plot mark the levels of attenuation from the outermost circle which is the reference maximum gain. Thus the best gain will occur at "0dB", with respect to the "forward" direction of the antenna. At all other angles to the antenna's axis there will be some degree of attenuation. The least gain will occur at the perpendiculars of the antenna axis.
The same antenna can be simulated at alternative frequencies. When this antenna is simulated at the WABI center frequency of 213 MHz some interesting results appear.
In the graphic an unexpected result is shown. The antenna's front to back ratio has reversed. This means that the greater gain direction is opposite to the design direction. The greatest gain is found going out the back of the antenna. Also the forward gain direction now acts as an attenuator. The direction which once had 11 dBi of gain, now has -4.45 dBi of loss. Not good if you are trying to receive WABI.
The reversal of the front to back ratio means that the antenna, at WABI's frequency, will perform better if it is pointed in the opposite direction. From my location all the major TV stations are in the same azimuth. This means that WABI may well be received if the antenna is pointed toward the NE. The backward gain is 5.35 dBi. That's a whole lot better than -4.45 dBi.
Unfortunately reversing the direction of the antenna did not enable the reception of WABI.
Oddly aiming the antenna so that it pointed more or less 45 degrees West of the correct direction (SW) did manage to pull in WABI but only occasionally. Not all the time.
This suggests that there may be some refraction/reflection off an object which increased the amount of available signal. Most likely into the back of the antenna, when the front is facing West.
I have not included dimensions of this particular version of the antenna, since it was the first experimental version. Among some of its shortcomings is the need for a 300 ohm to 75 ohm transformer. It would be better if the driven element were designed to be connected to standard RG-6 (75 ohm) TV coax directly, so a redesign is called for.
Here is a picture of the test antenna installed in its test configuration. In the picture the feed line is 300 ohm twin lead.