Problem:
  After using a piezo tap pedal live, it became apparent that I wasn't always going to be able to concentrate on keeping my foot on the flat little pedal, play, and sing all at the same time. I did it, but a couple of songs were monkeyed up because I was trying to find the silly pedal that, despite my best efforts with no-slip and silicone pads, still managed to move around on the cornmeal floor (or whatever it was!). Once I was sitting which is generally easier, but I was wearing a long skirt and I couldn't see where it was, so I hiked my skirt up mid-song in a desperate  effort to find it and got cat calls from the audience (blush!)

I made a new pedal with edges shaped like a shoe sole (shown in photo left), so my foot is always "in" it. It works good and has a separate sensor for toe tapping, but still my foot moves off of it and I don't feel it unless I'm barefoot. Gena, my partner, told me I should make it strap on like a ski, or maybe a sandal, which is a good idea. But really, I might as well make it a shoe after all that!

A Real Solution:
  A while back I saw that some guy had made a pair of MIDI shoes that had piezo's built right in, and a wire coming out of each shoe to plug into a V-Drum set. That gave me the solution to this. A well placed piezo embedded in the heel (and maybe on e in the toe) would mean I could be anywhere on stage and it'd work! Right away the wire must go. With all of the tiny & affordable ($3) wireless boards out there, wires are a real no-go with me.

 Of course this greatly complicates the whole thing. A battery is needed, a transmitter, a switch so the battery doesn't die, a socket to plug in a charger to charge the battery, an LED to show it's on, and a dedicated receiver to decode that a "tap" has actually happened. (Built inside the Show-In-A-Box case.

 I bought some used ( but nice ) shoes that have a bit of a clog height, about an inch, and tore it apart. To be light weight, I know shoes aren't solid rubber, they are hollow with supports, usually triangular in shape, throughout. I was right. I don't want to weaken the integrity oft he shoe, especially being it's for stompin', so the boards must fit into the deep triangles.

 A 318Mhz TX with Rx'er is definitly tiny enough. I found a flash PIC in my stuff that only has 8 legs, but fits the description of what I need here. I've ordered 7.6 V worth of LiPo Batteries already but they haven't arrived yet so I guess I'll have to wait.

 What did arrive was a curiosity item I purchased about a month ago. An MPU-6050 accelerometer / gyro chip. I've played around with accelerometer chips years ago, but this gyro is new to me. Back in the day the accel. sent out tiny x/y/z voltages that required the amps to the ADC's to be right next to the chip in order to get a stable usable reading.
 But since the advent of tablets and phones and cameras using them for image stabilizing they have really evolved. And thankfully the price is 10% of what it was only 6 years ago!

            The original "no frills" idea...

The new diagram to the left describes the new layout. (No I KNOW I'm an engineer! ) No project is complete without an RGB LED, and I had to sacrifice for it as well.

The Circuit:
 The MPU6050 board has it's own 3.3V regulator so 5 volts is fed to it from the 5V regulator. The TX board is also 5 volts so is on the same bus.
 The Clk & Dat on the I2C bus must have resistors. The Clk can be a 3.9k & 10k to Gnd on low side to get 3.3v. The Dat must be bi-directional so is just a 3.9k. The PIC12F675 will provide a weak pullup for rx'ing data from the MPU.

The "3" LEDs are driven by 3 ports, 2 of which are used as inputs as well. All 6 pins are GPIO.
Green: I have the INT output from the MPU marked as amplified, but using a diode on the LED will allow a weak pull-up to do it's job.
Red: Shares a pin (5) with a serial data input from the charge plug. (Must be a stereo 1/8") The zener will stop the surge of power when plugging in to charge.
*If the pin is high on power-up, will switch into programming mode to change settings via 9600 baud.

 Pin 2 sends pulse data to the Radio TX'er. This will be in PSK non-standard, works good if signal is good. This TX'er is one way of course, but the "bootstrap antenna" (hahaha sorry!) should give a few feet of range.

The batteries will only charge when power  switch off.

 Thoughts on usage:
For different modes, the control must be on shoe-pedal itself:
Hold foot vertical and tap heel to turn on/off tap function: led goes Green or flashing(off) *turned off by next..
Hold foot sideways and tap on side to switch in/out Note mode (like tapping toe on a circle of music keys!):Led flashes Red
Hold foot toe-down and tap toe for another mode....

Update:
 The radio transmitter/receiver will still be the same, but I think the microcontroller would be better an Arduino pro-mini. It'd be able to handle the switching and LED easier, be easier to program, using a different plug, and end up more configurable. The best feature of using the pro-mini is it's built in hardware USART.

USART Radio Test: The first test was 2400 baud and worked flawlessley, then up to 9600, then 19200 baud (plot shown right). You can see the slight delay (as a result of the slew mentioned above)  I was really surprised this even worked at all as with radio PSK, FSK are king. I'm still not entirely secure with this method as 15 feet away errors started, and at higher baud rates, the slew delay (I wish I could fix that!) starts causing errors.  You can see how that would happen in the lower sample. Hi-Lo is only 22.5uS, but the return to Hi is almost twice that.  It seems stable at 19200 otherwise, and generally I won't be much further than a few feet from the SIAB. This test also used pretty crappy non-tuned antennas, just bits of wire off the wish board. Because the antenna is 90% of the system usually (well 100% if it is disconnected!) I'm sure the range will be much better.

The '255's are to establish a stable sync and carrier. To conserve battery power, a shutdown of the stream would be desirable, but the "header" required can be slow. A time delayed stop transmit may work. If there is no motion for a minute, then the transmit stops. But that may become a problem as well!  The best solution was to find a maximum time that the transmitter would stay active, and keep the RX USART happy. The TX  module uses decoupling caps, so once the cap is charged, the TX stops. Therefore the maximum time for inactivity is 30mS (as tested) so it'll be 20mS to stay inside the safety zone with '255's, which are only a 1/10 byte pulse.  The logic analyzer output to the right shows it's all pretty stable. (TX is red) Me likey! Future Modifications:  I know I won't be able to leave it at this, as one never can right? Having holes drilled in the side of a rubber sole of a shoe is a recipe for disaster, especially in the wet climate I live in. One puddle and it's done! It needs to be sealed. The charger/Programmer is the big problem. It's very preferable to have a socket to plug into for that, which would prevent repeatedly pulling it apart to plug in the programmer. The charging could be done with a simple inductor pairing (you can actually buy those now!) by way of using two coils at resonance, one in the shoe, the other on a charge station. By oscillating the charge coil, the power will transfer to the shoe coil. This voltage must be rectified of course, then regulated to charge the battery.    The programming (FTDI) could be connected via a small header on the top of the shoe (somewhere..?) Once the programming is done, it could be tucked under the sole into a spare opening for safe keeping. The switch (push-in, push-out, or just pushbutton as Arduino could standby in low-draw "sleep" mode) could be mounted on a small board on the back of an opening facing outward, then it'd only take a push in the rubber side to "click" it on/off. I made a little CG drawing of the heel area and how this would work (in theory!) to the right. It's pretty flexible on the sides (the heel) between the gussets.  The RGB LED could be a fancy affair. Adafruit and their "wearables" comes to mind. What's wrong with a few twinkling LEDs at foot (sorry haha!) They *are* pretty so maybe I'll run them over the top or along the edge. We'll see!       OK that'll be it on this part of the project for a while, as the whole SIAB box has priority over this.

More to come!

Cheers! Sandy*
June 17th 2015