You are here

Second prototype: electronics

The first prototype of the radio worked, but it left a few things to be desired:

  • The sound was weak in all but the smallest rooms.
  • The tuning range covered only about half of the FM band.
  • The case was bare-bones, quickly made from cheap, relatively rough wood.

In the second prototype, we try to address these issues to some extent.

Output power

The LM386 used in the first prototype delivers about 1W of power. Paired with a cheap, generic 0.75W speaker, the sound of the radio was pretty basic: good enough for a prototype, but not for a "daily use" radio. I decided to stay with a chip-based amplifier for now, and I started to look around for a slightly more powerful amplifier chip. I'm not sure how I arrived at the result, but I settled on the TBA820 as the chip to use, paired with a more decent 10cm diameter 2W speaker. Probably this was the result of a combination of chip characteristics, simplicity of the circuit, and easy availability.

Tuning range

Some experimentation was due to find out why only half of the FM band could be tuned be the first prototype. My initial thought was that the capacity range of the BB204 varicap was smaller than the range of the BB305, which the original schematic called for. The original tuner included a voltage stabilization circuit, which is designed such that the voltage across the varicap can be varied by (slightly less than) 2V. Maybe this circuit should be modified for a different voltage range?

After playing with the voltage range that we can apply to the BB204, it turned out that the BB204 could cover the whole band, but that we would have to modify the voltage range spanned by the tuning potentiometer.

Sensitivity

One of the external capacitors for the TDA7000 determines the sensitivity of the tuner; this is the ease with which it locks on to a signal when it detects one. The example circuit provided by Philips in the data sheet uses a 10nF capacitor for this. This seems to make the tuner fairly selective though; only very strong stations are easy to tune. For the second prototype, based on the Practical Everyday Electronics article (see "First Prototype" installment), I decided to test with a bigger value. The article suggests a 68nF capacitor, but notes that the sound might become flatter as a consequence of the high value of the capacitor; the article's setup uses treble and bass controls to offset this. I'm not using a treble/bass circuit, so I used a 47nF capacitor as a compromise.

Tone control

The main side-effect of increasing the sensitivity of the tuner would be that the received signal would be duller. I wanted to experiment with tone control anyway, so I decided to try to incorporate a passive Baxandall type tone circuit. This would allow me to correct the decrease in sound richness if it would really be perceptible. It quickly turned out that my tone circuit didn't quite work, so it didn't take long before I bypassed it on the PCB to allow me to concentrate on other improvements.

The schematic and the PCB

This is the PCB layout transferred to the actual PCB. I use a black felt-tip pen to draw the locations of the components on the component side of the PCB; this is to make adding the components easier. Fixing the ink with a fixing spray (easy to find in any art shop) is needed on the PCBs I use; without fixing, the ink is easily erased even when it is dry.

2e prototype (1)

This is the fully populated PCB with externals:

Componenten op hun plaats

The case

This time I used MDF to make the case, since MDF is very easy to work with, and can be painted to a nice finish. Here the case is under construction: the walls have been made, and are being glued together:
Prototype #2: kastje in opbouw

Case and electronics:
Alle onderdelen tot nu toe

A complication was the production of a nice front plate. The front plate of the original prototype was made of 3 mm thick MDF; this worked, but it was not easy to nicely mount the speaker to it. For the second prototype, I decided to keep the 3 mm MDF as the "real" faceplate, but it would be backed by a 6 mm thick second layer, in which the speaker and the potentiometers would be mounted. The front plate the only needed holes, and there would be no screws or nuts on the outside of the case.

Other stuff

Initially, the first prototype worked using a single LR9 9V battery. It became clear pretty quickly that this was not enough: it worked at first, but after some time the tuner started losing its tuned frequently. As time went on, this happened more and more, until it became impossible to listen to the radio because the tuned station would fade into static already after a minute or so. Upgrading from the LR9 battery to a set of 6 1.5V AA batteries solved that problem for the first prototype. 2 months later, the radio still works perfectly fine, with no or very little tendency to drift.

For the second prototype, I thought of two ways to improve the design. First of all, since we are now using a 2W amplifier instead of a 1W one, we would need even more power. I decided not to go with the 6 x 1.5V battery pack, but rather to upgrade to an 8 x 1.5V battery pack for good measure. This allows us to use rechargeable batteries, while keeping the more or less the same operating voltage (8 x 1.2V = 9.6V). Secondly, it would be useful to get an indication to show when the battery voltage is dropping too much to guarantee proper operation of the tuner. The design for the second prototype includes a small 2-transistor circuit that will turn on a LED if the operating voltage delivered by the batteries drops too much. Of course, this circuit uses some power itself, but the visual indication will allow us to recharge the batteries without becoming frustrated by a drifting tuner first.

Observations

  • Above a certain (not so loud) volume, a lot of crackling becomes evident. After some thought, I'm now thinking that this may be caused by the amplifier sitting very close to the tuning oscillator circuit on the PCB, which might be causing cross-talk (is that the right word?) The next version of the PCB should have the oscillator and the amplifier further apart, to check whether this improves the crackling situation.
  • For the tuning potentiometer, I thought it would be nice to use a slider instead of a regular potentiometer. However, for fine tuning the slider is not as good as the regular potentiometer since it is harder to move the slider by tiny amounts. For the next version of the radio, I will revert to using a turning potentiometer over a slider.