A while back I build some muscle sensing gear.  I used this during a multi-media work with a dancer, who attached the sensor probes to her muscles.  We used two of these, running long wires back to my modular synthesizer.  Fast forward to today, we are planning to revisit this work, but the dancer wanted to see if we could get rid of the wires.  And so I did!  With a good bit of help form the DIY forum on muffwiggler.com I was able to design a pair of simple circuits to accomplish the job.

Technical Design

I considered several different design approaches.  The first part was deciding on the wireless protocol.  I went with a tiny, inexpensive 433Mhz transmitter and receiver pair, widely available for under ten dollars.  Here’s a stock photo of the transmitter.  Three pins carry power, ground, and data input.  The data is a single bit, on or off.  A small spring antenna was supplied, which I soldered into the ANT hole (see schematic detail).

I had to design the transmitter and receiver interfaces.  One idea was to encode 4 bits, using an A/D converter, and on the receiving side decode them and use a D/A.   But a much simpler pulse width modulation approach was used.  Here are the schematics.  (Click images for larger size.)

The principle is simple:  Generate a pulse wave and make the pulse width proportional to the output of the muscle sensing device.  Transmit the pulses.  On the receiving side use a low pass filter to transform the pulses into a voltage level, proportional to the pulse width.  The output of the LPF is then proportional to the muscle sensor output on the transmitting side.

Transmitter

Powered by two 3V lithium cells in series, the transmitter consists of a TLC556 dual timer chip together with a scaling and offset circuit in between the muscle sensor and the pulse width control input.  Section one of the 556 runs at a constant 1Khz frequency.  The second section of the 556 runs at a lower frequency with the width of the pulses controlled by a voltage on the CONT input.  This idea is straight from the TLC556 application note.  The muscle sensing board is the same one used in the previous project:  MyoWare Muscle Sensor AT-04-001.  It puts out a smooth, rectified and integrated EMG signal in the range of its power supply.  I determined experimentally that the best range for controlling the pulse width was between +2 to +4 VDC.  The EMG signal from the MyoWare is divided by one third and a small offset is added, adjustable by a trim pot.  This kept the output on the receiving end in a stable range.

I attempted to capture the schematic of the transmitter board, in the box on my drawing.  It uses an R433 transmitter chip and a few surface mounted components.  The circuit does not resemble any of those I found with the R433 chip data sheet.

The photo below shows the transmitter housed in an eyeglasses case.  The battery board holds the power switch.  The MyoWare board (red) takes input from a TRS cable which the EMG probe plugs into.  The case even had a small loop I could use to make a strain relief for the cable.  My circuit is built on a bit of perf board.  The transmitter board is plugged into a socket for easy replacement, if needed.  The three boards are stuck to the case with double-sided sticky foam.

Receiver

The receiver circuit is mounted behind a 6hp blank panel.  It runs off the +5 volt supply in the Eurorack case.  I used a red banana jack for the antenna connection.  No adjustments are needed on the receiving side.

Update December 19, 2023

Here’s the battery upgrade.

The original batteries were horribly deficient in supplying enough current. The upgrade consisted of putting two Westinghouse ER14250 1/2AA Size 3.6V 1200mAh batteries in parallel together with a DC to DC voltage up-converter to +5 volts. These were adequate to last for the length of a live performance.

Video Demonstration