m47_ElectronicsFF-4
Photo: Hep Svadja

The XR2206 is a slightly obscure audio chip that has surprising potential. It generates triangle waves and pure sine waves, and allows you to control them in unexpectedly imaginative ways. You could use it as the heart of a sound synthesizer.

I’ve been having a lot of fun playing with the XR2206 since it was suggested to me by a reader of my books named Jeremy Frank, who acquired some basic knowledge from Make: Electronics, but now seems to know more about some topics than I do.

Frequency and Amplification

The pin functions are shown in Figure A, and a basic test schematic is in Figure B. Don’t be put off by the apparent complexity. Each section functions separately and is easy to understand.

Figure A
Figure A: The basic pin functions.
Figure B
Figure B: A test schematic. You can leave inputs A, B, C, and D unconnected while testing the main functions of the chip.

For power, the XR2206 accepts 10VDC to 26VDC. The simplest option is to use the kind of 12V AC-to-DC adapter that you can find for less than $10 on eBay.

Pin 2 of the XR2206 is the audio output, which requires amplification. Pass it through a 33µF coupling capacitor and a 1M series resistor to the input of an LM386 single-chip amplifier, which has limited power, but is very simple. See Figure C below.

Figure C
Figure C: Add this ultra-simple amplifier circuit to hear the sounds.

A resistor-capacitor combination sets the audio frequency — similarly to a 555 timer, but with greater simplicity. A timing capacitor goes between pins 5 and 6, and a resistance goes between pin 7 and negative ground. Acceptable capacitor values range from 0.001µF to 100µF, while resistance should be between 1K and 2M.

Frequency is calculated using a very simple formula: f = 1,000 / (R × C), where R is in kilohms and C is in microfarads. I chose a 33nF ceramic capacitor and 500K trimmer. Because I used the dual inputs on pins 7 and 8 (more about them later), the frequency is doubled, ranging from about 60Hz to 7kHz. This is close to the full audible spectrum.

The resistor network applied to pin 3 allows you to boost the output from the chip. You can ignore pins 15 and 16, which enable you to fine-tune the wave symmetry, but are not needed for our purposes. Before I explain the other pin functions, I suggest you leave inputs A, B, C, and D unconnected, put all the switches in their up positions, and get acquainted with the basic functions.

Making Music

Switch 4, connecting pins 13 and 14 through a 220-ohm resistor, selects sine-wave output (Figure D) or triangle-wave output (Figure E). Sine waves have a natural quality — the kind of mellow tone you get by blowing across the top of a bottle. Adding higher frequencies to a basic sine wave can emulate various music instruments. A triangle wave (also known as saw-tooth) is richer in harmonics, and sounds more artificial.

Figure D: With the two 500K trimmers set to equal values, you get a symmetrical waveform. When switch 2 is down, it's a sine wave.
Figure D: With the two 500K trimmers set to equal values, you get a symmetrical waveform. When switch 2 is down, it’s a sine wave.
Figure D: When switch 2 is up, it's a triangle wave.
Figure D: When switch 2 is up, it’s a triangle wave.

When switch 3 is up, the XR2206 automatically uses the resistances on pin 7 for the rising time and pin 8 for the falling time of each audio cycle. This allows you to create asymmetrical sound waves. Figures F and G show some that I created. If you have an oscilloscope, you can view these signals yourself.

Figure F: Turning the left-hand trimmer all the way up minimizes the rise time, chopping the sine wave.
Figure F: Turning the left-hand trimmer all the way up minimizes the rise time, chopping the sine wave.
Figure G: The same treatment as in the previous figure, applied to a triangle wave, creates a saw-tooth wave which sounds shrill and buzzy.
Figure G: The same treatment as in the previous figure, applied to a triangle wave, creates a saw-tooth wave which sounds shrill and buzzy.

Now for some of the interesting features. Pin 9 is a digital input. In a logic-high state (2VDC or more) it selects the resistance attached to pin 7 to control frequency. In a logic-low state (1VDC or less) it selects the resistance attached to pin 8. If you feed a series of pulses to pin 9 — from a 555 timer, for instance — the XR2206 will flip between two different frequencies (Figure H). If you do this very quickly, unique sounds are possible.

Figure H
Figure H: Move switch 5 down, and the 7555 timer output makes the XR2206 alternate between two frequencies set by each of the 500K trimmers.

Because a 555 tends to create voltage spikes, I used an Intersil 7555, which has the same pin functions but a cleaner signal. If you’ve read Make: Electronics, the circuit shown in Figure I will look familiar as a basic frequency generator, with some additions. The 100µF capacitor stops frequencies from the XR2206 from trespassing into the 7555. The SPDT switch selects either a 4.7nF capacitor to generate low frequencies ranging from 150Hz to 10kHz, or a 1µF capacitor for 1Hz to 60Hz. The SPST switch rounds the square-wave output with a 1µF bypass capacitor.

Figure I: A 7555 timer can create pulses and frequencies to control the XR2206 audio chip.
Figure I: A 7555 timer can create pulses and frequencies to control the XR2206 audio chip.

The complete breadboarded circuit is shown in Figure J, with component values in Figure K. I used E-switch EG1218 SPDT slide switches, because they take up minimal space and are very affordable.

Figure J
Figure J: Breadboarded circuit including 7555 control timer and LM386 amplifier.
Figure K
Figure K: Values of components on the breadboard.

Use any combination of switches 1, 3, 4, and 5 to apply the 7555 output to the various inputs of the XR2206 (Figures L and M). Run the 7555 slowly, and it can pulse the sound or create a tremolo or vibrato effect if you apply it to pin 1, which modulates the amplitude of the sound (Figure N).

Figure L
Figure L: Applying the 7555 timer output to pin 7 of the XR2206 chip produces unpredictable results.
Figure M
Figure M: Weird outputs are possible if two or more of switches 1, 3, 4, and 5 are combined in their down positions.
Figure N
Figure N: Switch 1 allows you to vary the amplitude very rapidly.

Going Further

If you want to go further, substitute another XR2206 for the 7555 timer, so that your control pulses will have more subtlety than a simple square wave.

To create a real keyboard-controlled synthesizer, buy the cheapest old keyboard you can find, open it up, and look for the contacts that close when keys are pressed. Each key must apply a different resistance to pin 7 of the XR2206, to generate a different note. You’ll need as many trimmers as there are keys, and to adjust the trimmers you can use an electronic guitar tuner.

Keyboard from Walmart, just under $10, in box. The out-of-box experience ... rather similar to the in-box experience.
Keyboard from Walmart, just under $10, in box. The out-of-box experience … rather similar to the in-box experience.
I was happy to discover that the case is held together with screws. All of them must be unscrewed.
I was happy to discover that the case is held together with screws. All of them must be unscrewed.
Underneath the thin circuit board are little graphite button switches that are activated by each key press, to make a connection between two traces on the circuit board. Theoretically, you could use these buttons by running a pair of wires to each pair of traces on the circuit board — but it would be difficult.
Underneath the thin circuit board are little graphite button switches that are activated by each key press, to make a connection between two traces on the circuit board. Theoretically, you could use these buttons by running a pair of wires to each pair of traces on the circuit board — but it would be difficult.
Removing the thin circuit board reveals rubber buttons, such as you would find inside a computer keyboard. These activate the traces underneath.
Removing the thin circuit board reveals rubber buttons, such as you would find inside a computer keyboard. These activate the traces underneath.
When the rubber buttons are taken away, we have the circuit board. It communicates with the synth sound board via a 16-conductor cable. I would need a carefully programmed microcontroller to poll the conductors. That would be too much trouble for me.
When the rubber buttons are taken away, we have the circuit board. It communicates with the synth sound board via a 16-conductor cable. I would need a carefully programmed microcontroller to poll the conductors. That would be too much trouble for me.
The answer is to buy a bunch of pushbuttons like this. I found 100 on eBay for only $9.99!
The answer is to buy a bunch of pushbuttons like this. I found 100 on eBay for only $9.99!
The underside of each key is accessible through a hole in a plastic strip. Now that the circuit board has gone, some plywood reinforcement will be necessary. The holes are not equally spaced, so I did a rubbing on paper with a soft pencil. Then I pricked through the center of each circle onto a strip of ¼" plywood.
The underside of each key is accessible through a hole in a plastic strip. Now that the circuit board has gone, some plywood reinforcement will be necessary. The holes are not equally spaced, so I did a rubbing on paper with a soft pencil. Then I pricked through the center of each circle onto a strip of ¼” plywood.
Having made my marks, it was time to drill them with a ¼" bit.
Having made my marks, it was time to drill them with a ¼” bit.
The pushbuttons fit side by side with the plungers through the holes, where they will be activated by pressing the keys of the keyboard. Gluing the pushbuttons in place requires care, to avoid getting epoxy glue on the sliding plungers. But I ended up with a separate switch activated by each key.
The pushbuttons fit side by side with the plungers through the holes, where they will be activated by pressing the keys of the keyboard. Gluing the pushbuttons in place requires care, to avoid getting epoxy glue on the sliding plungers. But I ended up with a separate switch activated by each key.

Lastly, here’s a fun experiment. You can run the XR2206 at a radio frequency, such as 70kHz, using a 0.001µF timing capacitor and a 1K resistor in series with a 5K trimmer between pin 7 and negative ground. (Leave pins 8 and 9 unconnected.) Pass the signal from a microphone through an LM358 op-amp into pin 1, which enables amplitude modulation. When my friend Jeremy Frank did this, he managed to convert the oscillator chip into a tiny radio transmitter.

What other interesting things are possible? Your first step should be to read the datasheet for the XR2206, which is very informative. I have a feeling we’ve only just scratched the surface with this versatile audio generator.