Radio is certainly one of humanity’s most interesting achievements. With it, we are able to wirelessly communicate over very long distances. That’s why I picked up this really old General Electric pocket radio. It only operates through AM modulation, and as you will see in the video, it is a very simple design. That’s what sparked my interest in creating an AM transmitter. How difficult would it be to send an audio signal, like music or one of my videos, to this radio wirelessly? And what would it’s quality be like? Well, in this video, I’ll show you if its really feasible to make your own AM transmitter, how far it can transmit, and how good the quality is.
Before we can make the transmitter, its best to start with the thing we are transmitting to. Based on the description, we can already know that it is a transistor radio. And before I can even turn it on, we’ve already ran into our first challenge: the battery is unlike any I’ve ever seen. Unless I’m missing something, I’d say that this kind of battery is not being made anymore. Anyways, its voltage is 4.5 volts. So, I connected my power supply up to the radio to make sure that it actually works. And it does, but it has horrible sound quality. With that, I decided to open the radio up. It only took four screws to get to the inside. The bottom was attached to the speaker, which was operated with a transformer, which interesting to say the least. I then turned my attention to the circuit board at the top, and it took another few screws to free it. Looking at the anatomy of the board, we can see just how simple it was. There are only capacitors, inductors, resistors, and transistors. Now, I can’t tell for sure what the IC in the middle is, but I have fair confidence that it is a group of transistors, since I tested the diode junctions with my multimeter. There’s also a large coil with an iron core, which forms the antenna. If you remember my first radio repair video, you’ll remember something similiar. Unlike FM, AM doesn’t typically use one of those telescoping antennas.
Anyways, we should take a quick moment to explain how this AM radio is able to take the radio waves and turn it into those radio talk stations that we are familiar with. This website, made by Peter Vis, has some good information on this topic, and I’ll put a link in the description. Although the design isn’t exactly the same, this schematic should provide you a rough idea of what’s going on the radio. Let me illustrate an overview what is happening. Let’s say that we tune the radio to one megahertz. That means an LC tank circuit’s resonance was set to one megahertz. This is because capacitors and inductors have inverse reactive values, so every frequency will be blocked except for the point where they come together and let our intended frequency through. Ok, now we’ve selected our frequency, but what in the wave carries the audio signal? Well, in AM, its the amplitude. Basically, the signal oscillates at the carrier frequency, which is one megahertz in our case, but the amplitude of the signal is based on the audio signal. So, when the audio signal is low, the carrier wave is a small amplitude, but when the audio signal is higher, the amplitude is higher. So, the reciever has to reverse this and regain the audio signal, which we won’t look at in this video. We are focused on the transmitter side, but stay tuned for a future AM receiver video.
Anyways, for anybody curious on how I fixed the radio, I realized that it had no power supply noise rejection, so my switiching supply was just too noisy. So I decided to just keep it battery powered for simplicity. I used a nine volt battery along with an LM317 circuit put together in mid air. Anyways, after gluing it in place, the radio was as good as new. One thing I’d like to add is that I appreciate the design of this thing, since modern electronics have an annoying habit of being incredibly difficult to open without breaking anything. Anyways, let’s get into the design of our transmitter.
Like I described, we need two signals to transmit, the carrier signal and the audio signal. The audio signal is easy, it will come from my phone’s audio port. We just need to generate the carrier wave. And in the spirit of keeping it old school, I will keep it mostly discreet, although I will use one IC later on in the process. Anyways, some of you may have already thought of oscillators that could generate our carrier wave, such as the 555 timer. Well, we are going to use what’s called a colpitts oscillator. This is because LC oscillators are better at running at higher frequencies, like we are going to do for our one megahertz signal. Just like a reciever tunes an LC tank to filter out other signals, we will use it to generate our signal. If I connect an inductor and a capacitor in parallel, then they will pass voltage between each other, forming a perfectly tuned sine wave. The frequency can be modeled by the equation: 1/2pisqrt(LC). The only problem is that in real life, there are parasitic resistances, so it eventually runs out of momentum, so we will need to setup a circuit to refuel it. And we can use a transistor as an amplifier to refill the LC tank. So, here is the complete colpitts oscillator. The reason why I split the capacitor up into two different parts with ground in the middle, is because we should give a voltage feedback, with respect to ground, from the LC tank, so that the transistor can refill it at the correct time. The LC tank is still calculated the same way, just calculate the capacitance in series.
So, based on this, I started building the circuit on the breadboard, using a 100 uH inductor and a combined capacitance of 200 pF by using 330 pF and 680 pF capacitors. And, according to the equation and the simulation, the circuit should oscillate at 1 MHz. But for the life of me, I just could not get it to work. That’s when I realized that the inductor was just way too large for the frequency we were running at. To make it work in real life, we should use a smaller inductor the higher the frequency that we use. And since I didn’t want to wait another couple of weeks to release this video, I decided that I should make my own inductor. You can follow along too, the process is easier than you may think. I used this 24 AWG enameled copper wire. For 1 MHz, I decided to aim roughly for 10 uH. So, we can use the general inductance formula to calculate for how many turns we need to get 10 uH. The equation is L = N^2uA/l. L is the inductance in henries, u is the permeability of the core, N is the number of turns, A is the area of the coil, and l is the length of the coil in meters. For u, we can use 1.256 * 10^-6 for an air core. Since coils are usually circular, we can use the equation pi * radius^2 for our area. We already know our target inductance. For length, we can use a bit of algebra manipulation to make the calculations easier. My wire is 0.3 mm in diameter, and if we assume that our turns are roughly exactly next to each other, then we can say that the length of the coil is the number of turns multiplied by the wire’s diameter. Coincidentally, I found that the battery from the old radio actually was a good size for the coil, and I measured it to be 16.5 mm. So, plugging in everything, and estimating about 30 turns, I found that we would get about 10 uH.
So, I got started. I wound a few turns of the wire tightly around the battery, and then taped it on. I then wound the rest of the 30 turns and taped it together. After a bit of hot glue on the outside, I was able to remove the coil from the battery and take off the tape. Now, I don’t have an inductance meter currently, so I just decided to jump straight into testing. And after selecting the capacitors to be both 4.7nF, the circuit worked. And by some miracle, the output frequency was basically exactly one megahertz. I was lucky with this one, so you may have to play around with the individual values of the capacitors to get the frequency you need. Anyways, I decided to modify the circuit to the common-base colpitts oscillator. The circuit works in the same fashion, the components are just rearranged so that there is better stability. And that marks the end of the carrier frequency generation, we can now move into modulating our audio signal into it like described earlier.
Since the oscillator is sensitive to external stimulus, I fed the output into a simple emitter follower. Now we can modulate it without worrying about loading the oscillator. Remember from earlier when I said that the amplitude of the output wave is based on the audio signal? Well, we can actually enhance our emitter follower to achieve that purpose. Basically, we want to reduce the supply voltage of the emitter when the audio signal is at its valley, and increase the supply voltage when the audio signal is at its peak. And there are two ways to do that: the first is to modify the actual voltage source, which is a bit tricky. The other is to raise it from ground. And to do that, we can place a PNP transistor in between the ground and the emitter resistor. And this will modify the virtual supply voltage of the emitter follower. And running my function generator shows us that the output signal is actually successfully modulated. I attached an antenna to test it, and we can actually hear our sine wave on that little radio from earlier.
But a sine wave isn’t the most exciting thing, so I started working on making this compatable with a phone’s audio jack. Since the output voltage of an audio jack is usually something like 1 volt peak to peak, I needed to amplify its voltage before feeding it into the modulator. So I used an op-amp as an inverting amplifier here. I know, it breaks the otherwise completely discrete design, but this isn’t the important part. Anyways, with the circuit working, it was time to solder it and make it permanent. And after a few hours of soldering and debugging the radio works. Although its quality isn’t the greatest. It was at this point that I realize a tiny improvement that could be made. So I placed one 22k resistor before the modulator, and as you can clearly hear, it really improves the audio quality. And after soldering that, the project is actually complete this time. The transmitter only has a range of a couple of feet, and there is a lot of noise in the audio signal, not to mention the interference from neighboring radio stations. Here are a couple of audio samples. First is one of my videos… This is some music… Well, that’s to be expected from a low-power DIY AM transmitter. Further experimenting showed that flicking my light switches and turning on the desk fan actually interfered with the transmission. That’s why FM is preferred over AM these days, since AM is very susceptable to external noise.
Anyways, that’s it for this video. I hope you enjoyed it and learned something new. If you did, please consider subscribing so that you can see my other videos. Have a good one!