I recently found an old signal generator project I made in the very distant past. It was almost certainly built from an article in a popular electronics magazine, possibly sometime in the 1980s. Unsurprisingly, it had been left in a drawer gathering dust until it dawned on me that it could be really useful for testing audio circuits in my radio transceiver.
A quick look inside its plastic box revealed it to be a Wien Bridge design, which is known to produce a decent sine wave signal, exactly what I needed. It was also clear that I had made a pretty good job of its construction. From the battery clip, it was evident that power came from a small 9V PP3 battery.
The circuit looked fairly basic, built around two transistors on stripboard with accompanying resistors and capacitors. The problem was that the circuit diagram was long lost, and it was doubtful whether it would still work after such a long period of inactivity.
However, not to be put off, I started working out how the components connected. Determining component values wasn’t exactly easy, but fortunately some markings still remained. Where they didn’t, the parts were measured to determine their values.
After a good few hours, I had managed to untangle the wires and unusual component positions and finally recreate the circuit in KiCad. From working through the circuit, I was surprised to discover that the design had some significant features, including a frequency range from approximately 100 Hz to 1.5 MHz and output-level stabilisation using a thermistor. This controls the amount of feedback, which in turn regulates the output level.
The discovery of the thermistor was particularly interesting. Putting its part number into a web search showed that it is unfortunately now obsolete, not to mention relatively expensive. The datasheet indicates that it has a resistance of 5 kΩ at 20°C and a negative temperature coefficient.
Reading up on Wien Bridge designs suggested that it was fairly common to use an incandescent pilot lamp instead of a thermistor. For anyone building a similar circuit, using a small lamp to control the output level is definitely an option. In this circuit it wouldn’t be a direct component swap, as the lamp would need to shunt the output towards ground because of the change in temperature coefficient.
What really amazed me was that, after so many years, and despite its slightly suspect soldering and wiring, the circuit still worked. Even more impressive was that the whole design uses only two transistors, without a microcontroller or op-amp in sight.
The current drawn from the supply was only 12 mA, making the circuit highly efficient. When connected to an oscilloscope, the output displayed a crystal-clear, stable sine wave, exactly what I was hoping for.
Although the process of bringing the project back to life wasn’t straightforward, there was a real satisfaction in doing so and in rediscovering the merits of an old analogue design.
Click here to download the circuit diagram.