You are here

Project Radio: first prototype

The first prototype will serve as a proof-of-concept, since I have at this point only basic experience in both electronics and woodworking.

Scanning the internet for radio projects done by others, there's a lot to be found. One thing that caught my eye is that there are quite a few FM radio projects around that use a simple radio-on-a-chip to get a head start on the design. Transistor-based ones exist too, but they looked either very simple, proof-of-concept style items, or too complex for my knowledge level. Having only little experience in building circuits from schematics, especially where inductors and high-frequency oscillators are involved, using a largely pre-made tuner seemed like a good idea to increase the chances of actually succeeding in building a working radio. Getting to the bottom of this FM stuff is left for a future project. :-)

The IC in question is the TDA7000, which is (or was, because it seems to be out of production but still available for purchase) made by Philips. It is a chip that basically contains everything you need to receive FM signals. It needs a few external components though, which you use to configure the internal circuits to your liking. The parameters that are of importance for the end result are mostly tuning (of course) and setting reception sensitivity.

Components

Broadly speaking, we need two components: a tuner, and an amplifier. The tuner part is composed by the TDA7000, and the external oscillator circuit to tune it. Additionally there are some auxiliary capacitors for configuring the internal stages of the chip.

The TDA7000 outputs about 75mV of signal, which is not enough to drive a loudspeaker, so we need an external amplifier. In this first prototype I chose to use an LM386 chip. This is a low-power amplifier that will allow us to test the radio with little effort. The LM386 can deliver about 1W, so I paired it with a small 0.75W loudspeaker.

The oscillator: 2 possibilities

The TDA7000 is tuned by an external oscillator that oscillates at the tuning frequency. The data sheet (PDF) shows an oscillator composed of an inductor, an adjustable capacitor and a few fixed capacitors. It seems that "decent" adjustable capacitors are very hard to find, or very expensive. Most example circuits use trimmers; these are OK for prototypes, but you cannot attach a decent knob on a front panel to them. Another possibility, shown in some other example circuits, is to use a varicap diode. Varicap diodes are diodes, but they act as capacitors if you connect them "the wrong way". Depending on the voltage applied to them, the capacity will be different, which opens the possibility of building an adjustable oscillator, by adjusting the voltage across the varicap diode. Controlling the voltage can be done using a normal potentiometer, which is much more practical for building a nice front panel.

Generally, the advice for good quality seems to be to use an adjustable capacitor. A stable voltage-controlled oscillator requires a very stable reference voltage, which is not easy to achieve. The example circuits with varicaps use Zener diodes to regulate the voltage, but apparently Zeners suffer from voltage drift which translates into frequency drift, which in turn translates to a radio which "loses" its tuning.

For the first prototype, I chose nevertheless to go with the varicap design. The circuit by Gerd Roethig has a voltage regulation stage in 2 parts, using a 7805 regulator and a Zener diode, which might just be stable enough to be usable. In this case, the practicality concerns outweighed the quality concerns.

References:

The Schematic

One problem that I faced was that the varicap used in the reference design by Gerd Roethig was not available from any of my parts sources. I used another one, but from the data sheets I could not easily infer how the capacity ranges compared. This was something I would have to test in practice. The first step would be to build a tunable radio; first get it to work, then get it to work well.

This is the reference schematic I used from Gerd Roethig:
Het schema

It has many of the things I want: the varicap controlled oscillator and added voltage stabilization to offset the variable voltage from the battery. For the amplifier, I used a simpler design based on the LM386 that I found on the Internet, and that basically corresponds to the reference design from the LM386 data sheet.

The Circuit Board

The circuit board will be built up on an experimental PCB with pre-etched parallel tracks, since I have no experience creating actual PCB layouts and etching them. After a lot of thinking and drawing, the following is what I came up with:

Layout op de experimenteerprint

On the top right is the tuner, on the bottom right is the amplifier. On the left are the voltage regulating components.

One problem I encountered building the prototype PCB, is that in my drawing I forgot to mirror the IC pinouts. The first time around, I had to mount the ICs on the track side of the PCB to make the pins match the connections on the PCB… As you can see, I made a template with a grid on it and numbers on the top and the side, so that I could easily translate the locations on the drawing to locations on the PCB. As a result of this mistake, I added a step to tape a second grid to the back of the drawing, and punching holes in the places where tracks would have to be interrupted. Turning around the drawing would then show the correct places to cut the tracks.

The Prototype

This is what the first prototype looks like:
De schakeling

It took a while to get sound from the speaker; having no experience with oscillators, it took a lot of trial and error to stretch the inductor just the right amount to let the oscillating frequency fall in the FM range.

The Case

In order to be able to test the prototype in practice, I slapped together a few pieces of wood to have a makeshift case. It actually took a surprising amount of work to turn the prototype above into something that was easily usable; especially the front panel took some work, because the controls and the speaker had to be mounted securely.

Prototype behuizing

Inside view with the electronics in place:

Alle stukjes op hun plaats

Final prototype, I still need to figure out good knobs:

Prototype goedgekeurd, nu nog een definitief kastje.

Observations

First of all, every step in the process above turned out to be quite the amount of work. From piecing together the schematic from parts of other schematics found on the Web, testing the amplifier part on a breadboard first, turning the schematic into a PCB layout, making the PCB, testing the PCB, and finally putting the working electronics into a usable case, each of these steps took a few hours to complete. Respect to people who do this stuff daily!

Secondly, our radio works BUT with limitations. Currently, it can only tune about half of the FM band (10 MHz). Depending on how much the inductor is stretched, the 10 MHz moves up or down the FM band, so in order to enjoy the prototype I settled on covering from 94 MHz until about 104 MHz, since this covers the largest number of stations I care about. This is something to look at more closely. Using a different varicap than used in the original schematic clearly makes it necessary to replace some of the other components with components of different values to compensate for the different capacity range of the BB204.

Thirdly, when designing the PCB from the schematic, you look at the track side. This means you should reverse the IC pinouts since you're looking at them from the bottom, and not from the top… That was a frustrating lesson to learn!

The volume control is not very precise, it goes from silent to pretty loud with only very little turning of the volume knob. It seems that the LM386 is configured for a higher gain than what is desirable. Also, the LM386 + 0.75W speaker produces fairly weak sound that does not fill a larger room very well. It works somewhat OK for our small kitchen room, but if I place the radio in the living room, the sound becomes stale. I should revisit the choice for the LM386 for subsequent prototypes, and look for something a little bit more powerful.


The first prototype can be called successful: we can tune in stations, and we have somewhat reasonable sound. However, there are many things that we can improve upon. In future installments, I'll try to cover them, and improve the radio step by step until I'm fully happy with it. Stay tuned!