A partner to our 433MHz RF module article last month, we look at ways to improve the range of these modules with different antenna designs - By Daniel Koch
Antenna design is a deep, deep science. It is a study field and qualification all on its own, and while we can simplify a lot of it to a point, all simplifications are an exercise in deciding what information to include and what to reject. In the case of antenna design, there is a lot we need to include for even a basic working knowledge, and even then, there would be traps and pitfalls.
Therefore, while we might tackle antenna design and radio theory in a future Classroom, this month we have a Fundamentals covering some common 433MHz antenna designs and how to build them, with only the most basic 'big picture' theory so you can see roughly why things are designed the way they are. The idea is to improve the reception range of the 433MHz ASK modules we used last month. To that end, after we build a series of antennas, we will present a build for testing the range of these units with only one person. Normally, testing an antenna involves sophisticated equipment that belongs only to the RF design field and will, therefore, be of little use to the maker who is not regularly working with RF.
THE VERY BASICS OF ANTENNA THEORY
In the simplest terms, an antenna is an electrical conductor that takes the current fed to it by a transmitter and radiates it outward as an electromagnetic field through its surroundings; or, it is an electrical conductor that has a current induced in it by electromagnetic waves around it, and sends that current to a receiver circuit. It is really a form of induction.
All electrical circuits need to be complete for current to flow. The antenna seems like a wire sticking into the air, which confuses people. However, they are actually a capacitor, and the other side of the capacitor is the ground of the circuit. This may or may not be connected to the physical ground, or another ground, called a ground plane. Some radios, for example, have a metal plate inside the case as the ground plane.
So, the capacitor formed by the antenna and ground is charged and discharged. This is where the current flows, and it is why antennas are never DC driven: they are always AC, and that is why there is always a resonating circuit of some sort driving a transmitter or being driven by the antenna in a receiver.
It is possible to alter this very basic wire sending electromagnetic signals situation to make it more effective or more directional. We will cover some of that theory in each relevant antenna build.
ANTENNA ORIENTATION AND RADIANT PATTERNS
When dealing with antennas, you will likely find the terms 'horizontal polarisation' and 'vertical polarisation'. These refer not to the orientation of the antenna, but to the electromagnetic wave they transmit or receive. In some designs, this is the same as the physical orientation of the antenna, while in others, it is at 90° to it. So, a vertical antenna may pick up a horizontally polarised wave, and therefore be named as such. This can be confusing at times. The name refers to whether the electrical component of the electromagnetic wave is horizontal when compared to the earth, or vertical when compared to the earth. The radio signal itself is still transmitted horizontally to the ground in most cases, it is just that the electrical and magnetic components of an electromagnetic wave are at 90° to each other and so one will be vertical within the wave, and the other horizontal within the wave, even though the wave overall is travelling in the horizontal plane.
All antennas have a radiation pattern. Some are circular, while others are not. Some have radiation in both the horizontal and vertical plane. In fact, the electromagnetic waves spread a little no matter what, so for a horizontal antenna, there will be an angle at which the signal radiates that is above and below the exact plane. Some antenna designs, such as the Yagi, are naturally directional and
have a radiation pattern to match. The monopole is as close to circular as is realistically possible, while the dipole is close to circular, with some weaker or dead spots perpendicular to the antenna where the elements meet. These are all factors to consider when choosing an antenna design. You cannot, for example, choose a Yagi as a receiver if you need to receive signals from all around you, as may be the case with a set of temperature transmitters or soil moisture sensors.
In traditional British English, the term used is 'Aerial', while in traditional American English, the term is 'Antenna'. Australia's history of being influenced by both of those countries in terms of technology and media means both terms are regularly used here. In modern English around the world, thanks to globalised media, both terms are increasingly common in all English-speaking regions. There is no 'right' and 'wrong' to the issue, despite the fact that some will argue like their lives depend on it. Both terms are correct and valid.
FREQUENCY AND WAVELENGTH
There is a fundamental relationship between the frequency and wavelength of an electromagnetic signal. The frequency is the number of complete cycles per second, while the wavelength is the distance between identical points on the wave. The higher the frequency, the lower the wavelength, because electromagnetic signals travel at the same speed.
Antennas need to be a resonant part of the circuit. For mathematical reasons well beyond the scope of this article, antenna designs are usually in factors of 2 of the wavelength. A quarter wavelength is very common, but half wavelength and eight wavelength are possible too. This may be expressed as ¼ λ, where λ is the Greek letter Lambda and represents wavelength, or it may be written as λ/4. It is also possible to have other relationships that are not factors of 2, like ⅝ λ, but these are uncommon and very specifically engineered and calculated for specific reasons.
Calculating the wavelength of a given frequency is a good place to start when designing any antenna. To find the wavelength from the frequency, we divide the speed of light, C, by the frequency of our wave, ƒ. The speed of light used is the universal physical constant value, which is the speed of light in a vacuum, equal to 299,792,458 metres per second.
However, that's the speed of an RF electromagnetic wave in free air. In a conductor, the speed is lower and this affects antenna length. You will see a 'velocity factor' mentioned sometimes in antenna calculations, and that is the amount of C that can be expected in a conductor. For copper and aluminium, the velocity factor is around 96%, which means the current moves at around 96% of C. This will reduce the lengths of antenna elements slightly. It may mean elements need tuning by slight trimming to gain the exact resonance for a given antenna.
For our target 433.92MHz ASK modules, the full wavelength is 0.690893385876 metres, and that was still rounded to wherever we got sick of reading to! In more practical millimetres, the wavelength is 690.89mm, and most people will round that to 691 because we cannot cut wire more accurately than around 0.25mm or so. Notice that that value divided by four is the same 172.72mm we had in last month's ASK module article, which is rounded for practical use to 173mm. Note that this does not include the velocity factor, but on a quarter wave whip, 173mm becomes 166mm, and you will see this in some documents as the recommended antenna length.
Velocity factor is less of an issue when the wire connects directly to the antenna pad, which is why last month's whips were still 173mm. It becomes very much an issue when feed lines are used, the wire between the antenna connection on a transmitter or receiver, and the antenna itself.
WIRE AND CONNECTORS
Radio signals cannot just be sent down a wire. If they are, it becomes part of the radiating part of the conductor and affects tuning. Instead, antennas must be connected with shielded wire. In addition, the impedance of that wire, which is the AC relationship between the core and shield due to inductance and capacitance (and some other factors), must be carefully considered. While there are many factors that go into choosing impedance, we need to worry about none of them: The industry has standardised on 50Ω. Just don't buy 75Ω impedance coaxial cable, as this is for TV signals.
The same can be said of connectors - they need to be carefully considered if they are to effectively and cleanly pass the radio signal without affecting the antenna tuning. Because of this, it is best to stick to industry standard connectors. While often more expensive than using, say, a plastic RCA plug, they are far more likely to work. The plug and socket section also has its own impedance consideration, as well as contact quality, so do not be tempted to just split the coax into core and shield, and use a speaker spring terminal set to connect the different antennas to the test modules.
The best cable available on the retail market for 433MHz antennas is RG58 coax. It is a good balance between performance and cost, and is more capable than it needs to be for the frequency and data rate of these ASK modules. It is a little bulky for some uses, but the tradeoff is that it is far easier to physically work with when terminating than the tiny RG316 coax often used for SMA connectors.
Having said all that, sometimes twin-core wire is acceptable. In many dipole antennas, for example, one core connects to one side and the other to the remaining side, and because one is positive and one is negative during the cycle, the current in the twin core is always 180° out of phase between the cores and induction cancels out.
However, impedance still matters and it is hard today to find twin-core of the right impedance. In days gone by when it was more commonly used, it was more readily available. So, while it is theoretically very possible to use a twin-core wire with some antennas, speaker wire will not be up to the task. For most makers, coax is still the go-to.
As far as plugs and sockets are concerned, there are also several options the market has decided for us. The one to avoid is the UHF plugs and sockets, more accurately called a PL259 plug and SO239 socket. These date from World War 2 and in fact are only rated to 300MHz, with performance guaranteed only below 100MHz. Despite this, they are commonly seen on UHF CB radios, even those from reputable manufacturers, likely as a legacy from the industry standards from the 27MHz VHF days. They are also very bulky.
On the retail market, look for SMA connectors, N connectors, and FME connectors. These all work happily with 433MHz, and most will work above 2.4GHz. These are all available in both line plug and line socket, and panel socket too, although not always from all retailers. You may have to shop around to get exactly what you want. Also, note that SMA is a standard connector, but many WiFi systems use Reverse SMA. In standard SMA, the centre pin connector has the internal threaded nut, while the centre socket connector is surrounded by external thread.
In Reverse SMA, also known as Reverse Polarity (RP) SMA even though the polarity of the electrical signal is not reversed, the connector with the centre pin has the external thread, while the half with the centre socket has the internally threaded nut. Search the internet for 'SMA vs RP-SMA' for images, but we did not have any reverse SMA connectors to photograph so we cannot show you a comparison.
DESIGNS: WIRE ANTENNA
Last month, we used a straight wire cut to ¼ λ. This is a perfectly viable antenna for short range use, but we do not consider it to fully be a monopole as we describe next. With the wire antenna, there is no specific relationship to earth or a ground plane. They are considered monopoles by some sources, but many more sources consider them to be an arbitrary design and not a true monopole. Wire antennas can be quarter or half wavelength and still be practical, and some cheap modules come with a ⅛ λ wire antenna on them. We made wire antennas at 173mm and 166mm just by baring the end of a piece of coaxial cable to the right length.
DESIGNS: COIL LOADED ANTENNA
An improvement on the straight wire is the coil loaded antenna. This is made from the same solid-core hookup wire, but involves winding a coil in the middle. This improves performance over the basic long wire, partly because the coil forms an inductor that helps block unwanted (off-frequency) signals and improves selectivity. We promised this article was not a deep dive, so let's stop the theory there and just build it.
We cut a piece of hookup wire approximately 300 mm long. This is longer than we need but leaves working room. We left 25mm on one end, then started winding coils around a 2.5mm former. We used a 2.5mm drill bit for ours. There needs to be sixteen coils. After the coils, we straighten out the remainder of the wire.
Loosening the coils just enough to let them move, we removed the mandrel (the technical name for any former used to wrap a coil around). You can tape the coil to hold it together, or glue something inside like a matchstick. Make sure the item is non-metallic and non-conductive. Trim the ends to 17mm on one side of the coil (the 2.5cm end) and 53mm on the other side. Make sure the 17mm starts at the point of attachment, so leave room for soldering through PCB holes and such.
Either of the designs above, and a straight rod or whip aerial as well, can be made into a monopole antenna with the addition of a ground plane. The ground plane is a conductor of some sort that is connected to the circuit ground, and is placed perpendicular to the antenna element. The ground plane should radiate at least ¼ λ from the centre of the active element. The relationship between λ and antenna length alters the radiant pattern a lot in a monopole antenna. We decided to construct ¼, ½, and ⅝ wavelength antennas. These are 173mm, 345.5mm, and 431.8mm respectively. In each case, the ground plane was the same so we made use of banana plugs and a socket to change active elements.
The active elements are constructed from 1.6mm welding filler rod for rigidity. These have a fine copper electroplated coating over steel and so can be soldered to with a very hot iron. The ground plane does not have to be continuous but the more, the better. Ground planes can also be formed by radiating spokes like a bike wheel of varying number. However, as their purpose is as a reflector as much as a capacitive counterpart, solid may work better. As such, we cut a disc from aluminium foil and used that. It is glued to cardboard but in more long-term, less experimental arrangements, it could be either foil glued to plywood or made from blank PCB or metal sheet.
The banana socket was glued into an insulating spacer. The coax was passed through the foil ground plane via a hole in the middle, and up the spacer, Enough shield was folded back over the jacket to pass down the spacer and under the ground plane. A piece of copper conductive foil tape (with conductive adhesive) was used to connect to the foil on the top surface and travel through the centre hole to the underside. Then, the spacer was glued in place and the shield was attached to the copper tape underneath with more copper tape.
Finally, the active elements were trimmed to length so that the overall radiating length included the banana plugs.
Note that this is an extreme version of the monopole. Car FM antennas are an example of everyday monopoles where the car roof or body acts as the ground plane. However, for 433MHz use, you could experiment by having as few as three or four radial wires instead of the foil sheet used here. In addition, this one has an impedance of around 37Ω, which will not work great with 50Ω coax without some help. You can explore different impedance matching techniques like adding capacitors or inductors, or making chokes from the feed wire itself, but these are another topic. You can research them, or you can build the more practical version.
DESIGNS: BETTER MONOPOLE
A better monopole uses the spoke version, and changes the angle of the ground plane. Without going into the science (because it's a rabbit hole), changing the angle between the vertical radiating element and the ground plane from 90° to 135° makes our impedance close to 50Ω. That's an exact match for our coax and the design is arguably easier to build, too!
We cut five lengths of 1.6mm copper-clad welding filler rod to 180mm long. The final cut length needs to be around 166mm to be a ¼ wavelength with velocity factor included, but we can trim them after construction. Two double-sided lug terminals were used to mount the four members of the ground plane. The two terminals are soldered on top of one another first.
Following this, the four elements of the ground plane can be soldered on. Do this quickly at high temperature to avoid melting the solder holding the lugs together.
Then, a piece of coax was stripped and the shield twisted to form a core. The centre was passed through the holes in the lugs, and the shield soldered onto the other side of one lug. The centre core was stripped close to the lugs, and the vertical radiator soldered on. A skewer was glued on to keep the structure.
Finally, after all was said and done, the lugs were bent using a 135° guide marked on plain paper with a protractor and cut out.
The dipole is much easier to construct. Instead of a ground plane, it has two active elements, end to end. There is an insulating gap in the middle, and the feed from the RF module is connected to the inner end of one element, while the ground connection is connected to the other. These can also be made in quarter and half wavelengths, and probably others, too, but we stuck to quarter and half. We again used 1.6mm welding filler rod, soldering the elements to the coax and then, once cool, glued them to a piece of cardboard for structure. For permanent use, there are some great ideas floating around the internet for mounting the elements, but they don’t justify the effort until an experiment like this has established that the dipole is for you.
The dipole works because, as the signal is AC, current is flowing positively along one conductor and negatively along the other. However, as they are mounted from the centre facing out, the current ends up moving in the same direction each time. However, because of some capacitance and reactance issues, dipoles work best at a just under half a wavelength, not a quarter. That is the overall length and each element is therefore under a quarter wavelength. The total length of the elements including the gap between them should be just under half a wavelength. Our elements were 162mm long with an 8mm gap, factoring in the velocity factor of 0.96.
In addition, the dipole mounts so that the axis of the elements is vertical, for horizontal radiation.
The J-pole, which is more accurately called a J antenna or J aerial because the shape has nothing to do with the polar operation, is a sub-type of dipole antenna. It is more compact and has different radial pattern characteristics, although not by a huge amount. The top section of the antenna is a half wavelength radiator, while the bottom section is used for impedance matching. Antennas work better if the antenna matches the impedance of the feed line coming in, and without getting into the science of that (we'll save that for a future Classroom), having two quarter wave sections facing each other like this allows the impedance of the antenna to be matched to the feed line by choosing the feed point along the two quarter wavelength sections.
We started by cutting the long and short sections of the antenna conductors from 1.6mm welding filler rod, a little oversized. Then, we cut the cross piece. The two vertical pieces need to be 16mm apart measured from the inside edges, not the centres, so we used a scrap of 16mm MDF to hold them while we soldered the cross piece at the end. Then, the join was cleaned up by trimming it flush, and sanding off sharp points. At high voltages, sharp points cause charge to fly off and affect the performance of the antenna. At our 5V operation, that is not an issue, but cleaning up and sanding sharp points drastically reduces the amount of blood you will lose and the number of Band-Aids needed.
After soldering the pieces together, they were trimmed to the correct length and the feed line was attached. The solid core was soldered to the long side and the shield was soldered to the short side, 17mm from the bottom. Then, two turns of coax were wrapped around a 40mm mandrel (a piece of PVC pipe in our case), then taped after the mandrel was removed. This choke is essential in this J pole and the Slim Jim covered next.
The antenna is mounted by cable tying it to a piece of 20mm PVC pipe. There can be no metal in the immediate vicinity of the antenna, so a metal pole cannot be used.
DESIGNS: SLIM JIM
The Slim Jim is a variant of the J pole but with some different characteristics that make it suitable for certain situations. There is strong and often heated debate online as to which performs better out of the J pole and the Slim Jim. We're not getting into that debate but from what we can see, the evidence is very much in the middle on aggregate. Instead, we like the slim jim because of some of its physical properties.
The Slim Jim for 433.92MHz has an overall length of 505mm. The half wave side is 332mm, the quarter wave stub is 166mm, and the gap between them is 7mm. The feed point is for the J pole at 17mm. The difference with the Slim Jim is that the spacing between the sides is not critical. This makes it useful for permanent applications.
We made ours with a length of 20mm pressure pipe that we had on hand but electrical conduit works even better, because it is softer and less brittle. We started by drilling three 8mm holes around the tube, so that the outer two were opposite one another and the other in the middle, at 90° to the others. These were for the feed point wires, and to manipulate the coax cable. Smaller holes were drilled for the conductor, to match the dimensions above. A length of solid-core hookup wire was fed through the top two holes, down the far side, and through the lower two holes. The ends were brought together and the cable manipulated until the ends overlapped where the gap would be. Hot melt glue was then used to secure it in place.
The coax was stripped so that the braid was twisted into a core, and the centre core was exposed. These were fed through the centre 8mm hole, and pliers used to bring the centre core through the hole on the long, unbroken side, and the shield to the side with the gap. These were soldered on at the 17mm mark above the bottom of the conductor.
Finally, the quarter and half wave sides were measured from the lower and upper ends, respectively, and the size of the gap verified, before they were trimmed to length and glued down. The coax was wrapped twice around a 40mm mandrel and taped, for the choke. The coax exits the bottom of the pipe, and the whole assembly can be mounted to another piece of pipe as a mast. The bottom of the tube should be flooded with silicone or glue to provide strain relief for the solder joints.
The Yagi is a development of the dipole, but it has a higher gain at the expense of being quite directional. In very simple terms, it is a dipole with reflector elements that shape the radiant pattern from reasonably round, to very pointed. This is great for point-to-point systems like long-range data transfer between, say, a water tank on a property, and the main work area. However, it is no good where multiple nodes are involved unless they are on the same line. Having said that, it is perfect for direction finding, like tracking a beacon.
A Yagi antenna has a driven element that is active, where the coax connects and the signal is transmitted or received; a reflector at the back; and one or more directing elements in front. All but the driven element are electrically isolated and passive. Some tapering of the directors is often found to assist in directionality.
For our Yagi antenna, we have a reflector of 343mm long; a first director of 318mm long; a second and third director each of 305mm long; and a fourth and final director of 280mm long. These measurements are only for one of many designs. Some use 1/8th wavelength as a basis, while the directing element number varies. The driven element is 330mm long, but that includes a small gap for joining the cable and isolating the halves. We cut two pieces 170mm long, to be trimmed later. All the other pieces were marked halfway, for centering later. The middle of the wood from the next step was marked, too.
The overall length of this Yagi is significant. Many experimenters have achieved smaller, and even compact, designs. However, this one has a good reputation and has been made by quite a few people. We need a piece of wood 700mm long to mount our elements. 10mm from the front, we glued down the fourth director element, the 280mm long one. Then, 175mm from the front, we glued the third director, one of the two 305mm ones. 328mm from the front, we glued the third director, the other 305mm one. 480mm from the front, we glued the fourth director, the 318mm one. Then, 556mm from the front, we marked a spot for the driven elements. Finally, 620mm from the very front, we glued down the reflector, the 343mm piece.
The driven elements were soldered first, so the heat did not affect the glue later. We soldered the core of a piece of coax to one, and the shield to the other. Then, they were glued down with a 6mm gap between them. The overall length was measured, and each was trimmed so that the element including the gap was 330mm long. The coax was also glued down to hold it in place.
There is a variation you can try for the yagi, which uses an unbroken element similar to the J-pole. The top is 330mm long, and then the element is bent over on itself with a spacing of 6 to 7mm, with a length of 165mm. The centre core of the coax joins to the end of the smaller length, and the shield to the middle of the longer section. This is mounted so that the longer piece is in a horizontal line with the others on top of the wood (just like the two element version), which means drilling a hole in the wood for the shorter bit of the element, and the coax. This version may be better built on a piece of electrical trunking. We did not try it but it does get good reports.
It is necessary to match the impedance of the feed coax to the antenna in all designs. This may mean some tweaking or modification of designs, or the addition of capacitors or inductors. The specifics of this are for a future Classroom, as it's a topic on its own, but simple things like altering the angle of the ground pane can sometimes be enough. That is why we went with the second Monopole design. In a dipole, modifying the length of the elements often helps. With the J-pole and Slim Jim, changing the height of the feed point changes the impedance, and that is one of the big advantages to these designs for a maker untrained in the dark art of antenna design and construction.
When connecting wire to the antenna, care must be taken when choosing the exact point at which the wire attaches. In some designs, like the Slim Jim, this is prescribed. The reason is the balance of the different active elements. However, in other designs, mistakes are easier to make. Think of a straight-wire monopole, for example. If you cut the antenna to the right length, but then solder a wire to it that adds another 10mm before the shielding, then this adds to the active length and therefore affects the tuning. The opposite is true, too, if the point of attachment is somewhere along the monopole rather than at the end, then the resonating section of the antenna is smaller than calculated.
The other aspect of wiring to antennas is fitting the plug for the other end of the coaxial cable. While each type has its own specific method, they all share some aspects in common and have some common points to watch for. The first is that the inner core must be terminated to the centre contact of the plug or socket. Some do this with a separately crimped or soldered pin, while others do it with a fixed pin or socket, which has spring contacts on the back. In either case, the jacket and shield need to be removed from the coax to leave the correct amount of centre core and foam spacer exposed. This is really easy to get wrong and may take experimentation.
Pushing the centre core into spring contacts also takes care, as the core must be straight enough and the cut end clean enough to go into the contacts rather than deviate to the side and bunch up without contact. The amount of the (usually) foam or PVC spacer that needs to be left between the shield and exposed care varies between plug designs.
Also differing between connector types is how the shield is treated. For some, a spacer tool is used to open a cavity between the jacket and shield, with the shield either opened against the jacket or left against the inner insulation core. Then, a ring is inserted or a projection on the connector is slid into the space, ahead of crimping the outer ring. This may be a separate piece or integral to the connector body. Other coaxial plugs are designed so that the shield folds back over the outer jacket.
The body of the connector goes over this and is then crimped on. There are threaded connectors too, which are designed to be used without tools. In any of these cases, the critical points are to make good contact between the shield and the connector body, and to avoid stray strands of shield touching the centre core. This would short-circuit the conductor to the shield.
For each connector type, there are good tutorials to be found online. However, there are also poor ones, so if one makes no sense or you feel something is missing, keep looking.
We are only making and using one antenna at a time for each unit when we test these antennas. We are using both transmitter and receiver in our system but never at the same time. This means we can connect the antenna to both the receiver's antenna pin, and the transmitter's. Normally this would be bad news. In a transceiver, which both transmits and receives, dedicated circuitry and software handles communication in the single antenna. In the system we have here, two antennas would be needed if the system was going to both transmit and receive data at random. However, ours is a very structured system. On the mobile unit, the transmitter is used first, then only after a significant delay will any data come back.
With the transmitter's antenna pin connected to the receiver's connector pin through the common antenna connection, the mobile unit's receiver will receive the transmitter's signal. However, because of our timing structure, the Arduino is not reading the receiver's data pin while the transmitter is running. Therefore, it will not confuse itself. Likewise, on the base station, the normal mode is receiving. Only after receipt of a signal and the given delay will the transmitter be used, so the Arduino is no longer listening to the receiver at this point.
So, if you are ever building a system that uses both a transmitter module and receiver module together at the same end of the system, and data will be random so that the receiver should be monitored even while the transmitter is being used, two antennas are needed. However, unless your code is carefully written, the Arduino will not be capable of both at once anyway, as transmitting the packets often uses blocking delays. You would have to write a code that used timers without delay. That's perfectly possible, but for most maker-level transceiver applications, one antenna is fine. In the greater majority of ASK uses we can think of, the system is single-ended with transmitter at one end and receiver at the other, possibly plural in each case. This renders the issue irrelevant but it is worth keeping in mind.
Making your own 433MHz antennas is fiddly but rewarding. You can make designs that you cannot buy commercially made (without searching for hours through specialised and trade suppliers), and you can save a lot of money on the designs that you can buy. In addition, you can tailor designs to suit your needs. There is a lot of tuning and modifying needed for optimal performance, and we have only covered some of the designs around. However, this information should give you a solid grounding with which to expand the range of your 433MHz ASK projects.