April 2014. When I'm building transceivers etc, I often need to tailor the audio response on transmit and receive. For this an audio signal generator is required to make sure you get the right frequency response. My main RF signal generator will produce audio tones from 1 Hz up to 20 KHz, but there's a bit of faffing around to do and the level is fixed. What I needed was a quick, knob-tuned sine generator covering about 100 Hz to perhaps 10KHz in one range with a variable output level.
I have a basic DDS signal generator which will provide an output from 10 Hz to 6 MHz, but I want to keep that for something more demanding. I considered building a Wien Bridge oscillator or maybe a phase-shift one, but these need ganged pots and they really need some kind of auto-level control to stop them clipping or stalling. They also tend to need band-switching to get from one end of the audio range to the other, which can be very inconvenient.
What about a heterodyne type? These are a bit more complex than a simple audio oscillator, but they can produce nice sine waves and they tend to have a constant output level right across the range.
Basically a heterodyne generator consists of two radio frequency oscillators running at slightly different frequencies. The two are mixed together and the output from the mixer is filtered, leaving just the 'beat frequency' between the two.
Crystal oscillators are best for this, as they are very stable but can be 'pulled' enough to produce useful audio beats.
I've got a decent collection of crystals, and I found two identical 13 MHz ones. The first oscillator is a single transistor type, the only special feature being an inductor in series with the crystal and a varicap for altering the frequency over a few KHz. Tuning is done by a standard potentiometer. The second oscillator is fixed frequency, using the internal oscillator of an NE602 mixer chip. The mixer takes the signals from both oscillators and produces sum and difference outputs. The sum (13 MHz variable plus 13 MHz fixed) is unwanted and is suppressed in the post-mixer filter. The difference product (13 MHz variable minus 13 MHz fixed) is in the audio range and this is the one we want.
I had a cunning plan to pull the second oscillator in the opposite direction to the first one to double the tuning range, but as I managed to get an 8 KHz range by tuning just one oscillator I left this complication out. As for calibration, I'll just make an analogue scale. I don't need ultra-accuracy for this thing.
After the filter (which is just a single capacitor) there is an amplifier, a level pot and an emitter follower which buffers the audio signal and provides a low-ish impedance output.
Below is a shot of the board under construction and below that is my scope showing the nice sine wave output.