Complete and ready for portable VHF FM listening.

This kit was sold by Jaycar (cat. no. KJ-8238) as part of their "Short Circuits" series of projects. "Short Circuits" was a set of three books, along the lines of the old Dick Smith "Funway Into Electronics" series, intended to teach electronics by building various simple projects. The projects were designed by Silicon Chip. This TDA7000 FM receiver is described in Volume 2, project 21, under the heading of "Personal FM Radio".  As with the DSE Funway kits, one was required to buy the book to get the instructions, since they were not supplied with the individual kits. In later years, the books were downloadable from the Jaycar site. Volume 2 is available here.

• Details and history of the TDA7000

The Design.
The receiver is a headphone only design, and intended to be mounted inside a HB-6032 plastic case. This makes a convenient handheld unit, which, with the location of the tuning control, is quite ergonomic to use. It will also fit a shirt pocket. A headphone socket and volume control is provided at one end of the case, and a trailing wire aerial at the other end. No mechanical on/off switch is used, with this function instead performed by removal or insertion of the headphone plug. The power supply is a 9V battery. Since the TDA7000 works down to 4.5V, this receiver is an ideal way to use up 9V batteries which no longer work in voltage critical applications like smoke detectors or digital multimeters.
It has been well thought out, and is a very elegant design.

For the most part, the TDA7000 part of the circuit is as per the Philips application notes. There is a minor error regarding the tuning capacitor earthing.

The TDA7000 has been described in detail here. Briefly, it is a superhet FM receiver IC for operation up to 110 MHz. What makes it unique is the 70 kHz IF, which allows conventional audio type op-amp filter circuits to provide the gain and selectivity. Thus, coils or ceramic filters are eliminated. Only one coil is actually required, which is that of the local oscillator. A second (optional) coil provides broad RF input filtering. Because of the 75 kHz deviation of the FM signal, it can be imagined that distortion would occur with a 70 kHz IF. To avoid this is a 'Frequency Locked Loop', which compresses the deviation to 15 kHz, by modulating the local oscillator with the detected audio signal. The TDA7000 also features a muting or squelch circuit, and an artificial noise source to make the receiver still sound alive while tuned off station. These two features can be enabled or disabled as required.

Simplified block diagram of the TDA7000.

A standard plastic type of tuning condenser intended for MW superhet receivers is used, with a local oscillator section of about 60pF and an aerial section of about 160pF. Many other TDA7000 designs do likewise, with the lower value local oscillator section used for tuning the TDA7000's local oscillator. Since 60pF is too high for tuning the VHF FM band, it is padded out with a series capacitor. Interestingly, the Philips application notes show the higher value aerial section used instead, with a series 56pF capacitor. The higher the value of the variable capacitor and the smaller the series padding capacitor, the more non-linear is the tuning action. Stations at the low end of the band are spread apart while those at the high frequency end are cramped together.

This design is unconventional in that the two capacitor sections in series. This should present a more linear tuning action, but the problem is that the shaft and moving plates are not earthed. Hand capacitance could be problematic with the shaft live at RF. Additionally, the 9V battery is right up against the tuning condenser, and could also have a capacitive effect with the live moving plates.

The oscillator coil is adjustable with an F29 ferrite slug, as was done with the Silicon Chip, November 1992, and Electronics Australia, June 1988 designs. As with the EA design, the RF input coil is actually an inductance etched on the PCB, and I suspect was the source of inspiration here.

As with most TDA7000 receiver designs, the ubiquitous LM386 provides audio amplification. Although the receiver is intended for stereo headphones, there is no reason it cannot drive an external speaker instead.

Another undesirable aspect of the design is that the headphones are wired in series, so that the drivers are out of phase with each other. This can give an unnatural effect, since one driver is delayed by 180 degrees relative to the other. With loudspeakers, the effect is that of largely cancelling the bass response.
The reason for this method of connection is due to the power switching. The PN200 transistor switches the 9V supply. This is a PNP type, so when base current flows to the negative rail via the 6.8k resistor, the transistor switches on. In this instance, the 6.8k resistor is connected through the right channel of the headphones to negative. When the headphones are unplugged, no base current flows, and the transistor switches off, disconnecting the 9V supply. The 22uF ensures that any audio signal is bypassed. Although I'm sure it was never intended, a mono plug can also be inserted, and in this case, the 6.8k will be connected directly to negative.

Normally, for mono use of stereo headphones, the left and right channel are connected in parallel to the audio source, and the sleeve connection of the socket is earthed. Obviously, this would mean the supply would always be switched on, if this was done here.

A perhaps minor thing to note is the audio load resistor connected to pin 2 of the TDA7000. The data states this should be 47k for a 9V supply, and 22k for 4.5V. Seeing as this set will be used with run down 9V batteries, I did not bother investigating what the practical difference is between the two resistor values. The 47k across the volume control should be omitted for 9V operation to be correct, however.

Building the Kit.
I built up the kit as per the instructions, with the idea of modifying it, if my suspicions about its design 'limitations' turned out to be problematic. My initial impression was not good. The RF performance was not as it should be, with signals cutting in and out due to the squelch. I am in a good signal area for VHF reception, and this was not what a good TDA7000 receiver should be like.

Kit as originally constructed. Note the wire aerial.

1. RF Performance.
Even though the squelch was enabled by default of the design, the sensitivity should have been a lot better. The hand capacitance effect of the live tuning condenser shaft didn't turn out to be problematic, mainly because of the isolation provided by the plastic knob. But, I still didn't like it, and it is not good design. Because of the way the PCB has been designed, it was better to short out the local oscillator section and use the aerial section to tune the receiver. This was because an undesirable length of wire would have to be run to the local oscillator section. Had the tuning condenser been mounted 90 degrees turned to the left, it would have been easy to use the local oscillator section. I found that the ideal padding capacitance was 27pF, so this capacitor was taken from its former position in parallel with the tuning capacitor, and installed in place of the 68pF.

A slight error in the Jaycar circuit shows the tuning condenser earthed to the negative supply, but the Philips application notes show it earthed to the positive supply. Wondering if this was part of the problem, I investigated closely and found that in actual fact, the PCB was correct, and the tuning condenser is earthed to the supply rail, as it should be.

2. PCB Design.
One aspect of using the TDA7000 often overlooked is that the PCB design is very important. While the negative supply rail is at RF earth, and it shouldn't make a difference whether the tuning condenser is earthed to that or the positive rail, the TDA7000 can be critical in this regard. When I see people build up TDA7000 circuits on solderless breadboards, or various matrix or Veroboard constructions, it is a disappointment knowing that such constructions won't work to this IC's best ability. Experience has shown that the closer to the Philips PCB design the circuit is, the better it performs.

With the tuning capacitor shaft now at RF earth, the performance was still no better. Keeping in mind the PCB design, which while good, was still not quite exactly as per Philips design, I remembered the supply bypass capacitor which should be as close to pin 5 as possible. Best results can be had by connecting this across the supply under the IC. To this effect, I soldered a surface mount 0.1uF across the supply tracks. This still didn't provide the performance I was looking for.

Next was the tuned circuit itself. I didn't like the amount of track inductance between the supply rail where the oscillator coil is actually connected, and the earthy side of the tuning condenser, also at supply rail potential. It was quite a round about path despite the two connections actually being very close to each other. It only took a short link from the coil positive connection to the tuning condenser earth. This changed everything, with sensitivity much as expected. The surprising thing is this connection could have easily been part of the PCB. To give an idea of how much the original PCB layout affected performance, once the missing link was in place, the local oscillator coil had to be tuned down by about 4 MHz.

Mods done to the PCB to obtain correct RF performance.

3. Muting.
The TDA7000 muting can be a bit aggressive, and for a portable receiver with an inferior aerial, my preference is to disable it. This is simply done by connecting a 10k resistor across the 0.15uF capacitor at pin 1.

4. Parallel Headphones & the Power Switch.
My intention was to connect the left and right channels together and modify the switching circuit. To do this entailed cutting the PCB adjacent to the R (ring) terminal of the socket and connecting it to the T (tip). Then the S (sleeve) was to be connected to earth. I tested this, and it worked as expected. The switching is where things became problematic. My idea was to connect the 6.8k resistor to the now paralleled tip and ring connections. The circuit would work as before, except when the phones were unplugged. In this instance, current would now flow through the 220uF in reverse, since the LM386 output pin was now at earth potential. The way around this would have been to replace the 220uF with a non-polarised type. Unfortunately the largest non-polarised capacitor I had that could fit the PCB was 47uF, which was too low. The cost and inconvenience of ordering a capacitor just for this killed off that idea.

Mechanical switches were considered, but slide or toggle switches always get bumped into the 'on' position when least expected, discharging the battery. The ideal would have been to use a switch pot for the volume control. None of the small size required were conveniently available, and there is certainly no room for anything larger than the existing potentiometer. Ultimately, it was decided to revert back to the original set up. The anti-phase connection of the headphone drivers was actually tolerable, although not 'correct'.

5. Headphone Lead as Aerial.
No one wants a wire dangling out the back of a small portable radio. Aerials should be self contained. As it happens, there is already a wire we can use instead; that being the headphone cable. The Elektor TDA7000 receiver uses the headphones this way, as do the small auto-scan TDA7088T receivers, and my 6 transistor super-regenerative receiver.
For the Jaycar receiver, implementation is quite easy. Two likely looking RF chokes were found and connected between the T and R connections of the headphone socket, and their respective audio and earth connections The value is not critical, but about 5 to 25uH would be typical. The chokes I found were the tiniest toroids I had ever seen, and they measured about 22uH. Having done this, the headphone lead can now float at RF above the circuit earth, while the low inductance of the chokes does not impede the audio signal. The sleeve connection of the headphone socket was then connected to the aerial input, at the junction of the 39pF and 47pF capacitors. By default, these capacitors also isolate any DC on the headphone socket, so the aerial input does not interfere with the switching. The 6.8k is high enough not to load down the RF signal, so a third choke is not required for the sleeve connection.
Results are quite good, and no worse than the dangling 75cm length of wire. A limitation of headphone cable or other wire aerials for VHF is that they're often not in the ideal position, so some experimentation is required to obtain best reception. A 75cm telescopic aerial would be better, but there is definitely no room to include one!
At this point I was happy, and it was working as a typical TDA7000 receiver of this type should.

Showing the modifications to the PCB. The 10k disables the squelch. Two links are used to earth the tuning condenser moving plates to minimise inductance.

Fitting the Case together.
I had a lot of difficulty getting the lid on the case because of the battery. Ideally, the PCB would be lower in the case, but the side of the tuning knob would then be too close to the other side of the case. The way the PCB is mounted requires four small spacers so that the tuning condenser knob clears the bottom of the case.

Taken from the book, this shows how the PCB is mounted in the case. Problem is the height of the battery prevented the top half fitting correctly.

I thought about some other 6 to 9V battery that would fit, but nothing convenient came to mind. Ultimately, I had to cut plastic from the top half of the case with a  Dremel. This was messy and time consuming, and care had to be taken not to remove too much. Nevertheless, the required result was achieved.

Completed receiver. Note the plastic removed from the lid so the battery could fit. Red wire connects headphone socket to aerial input.

Before building the other kits, I will select the spacers first so the battery fits, and then shorten the tuning condenser shaft to fit if necessary.

.