Sally Williams @ Fab Academy 2015

Assignment: 

• Measure something: add a sensor to a microcontroller board that you've designed, and read it


Background

Starting week 10 on input devices quickly revealed to me that I did not have sufficient knowledge and understanding of electronics to progress quickly.

Without a more basic foundation in the principles of connecting components and direct guidance on routing the board in Eagle, I ended up wasting a lot of time unnecessarily this week.

I had wanted to use Input Device week to make a useful contribution towards my final project.

In essence I was developing a wrist wearable digital mala for recording and training the mind in mindfulness. In some buddhist traditions, there are certain practices which need to be accumulated (up to a specific number - which reflects the amount of practice one has accomplished). Therefore, I want to integrate an additional feature on the wrist wearable  - a semi-automatic and intelligent means to count (‘hinder-free’), accumulations of yoga postures (such as Sun Salutations or Prostrations for example).

So, here, the use of the phototransistor becomes very useful; when a single yoga sequence is completed, the body lies flat to the ground with arms extended over one’s head, flat on the floor. With the phototransistor embedded within the inside of the wrist wearable, when the wrist makes that contact with the floor, the light is blocked. When this event occurs, the LED would confirm a single count, and the code would count 1. As this action is repeated, the yoga postures being accumulated are automatically and effortlessly counted and recorded.

Plan A

Therefore, I wanted my board to include the following:

1 - use a slide switch to enable and control a specific function relating to the phototransistor, for a switched-on period of time (to disable accidental counting whilst blocking the light);
2 - use a phototransistor as a means to measure when the light is blocked (the wrist wearable has made contact with ground and light collection is disabled; reflecting a single sequence of yoga movement is completed);
3 - an LED confirms that the phototransistor has recognised the event and a single count has been recorded.
4 - an additional button to enable counting moments in mindfulness training.

Therefore, the board that I tried to create in Eagle (schematic), included a phototransistor, a button, an LED and a sliding switch, using ATTiny 44.




However, when it came to routing the board feature, I could not get it to a workable configuration. I tried auto-routing and manual routing (for days!). I also had configuration errors which even Eagle couldn't tell me about.  So it was back to the drawing board to a simplified version. I also needed more guidance to progress.

Plan B

The most important feature that i required on my board was the phototransistor; the sliding switch could be simulated by the conection and disconnection of the usb; and I could worry about the button later on. This simplification meant I could downsize to the ATTiny 45.

Pin out for ATTiny45:




Components added in Eagle Schematic View showing simplified board using the phototransistor and LED:



Components ready to be configured in Eagle Board View:




Board routed successfully (finally) :




Components soldered onto board

The board was milled succesfully and then the components were soldered into place.



I inspected the board, and all connections looked good to me. Wherever I had a doubt (slighlty dull solder), I returned to re-solder. 

I conneted the FTDI connector to power up the board (from the laptop) and the Fab ISP to program the board. Having selected all the relevant parameters (Board - ATTiny; Clock - 8MHz internal; Processor - ATTiny45; Programmer - USBTiny ISP), I proceeded to burn the bootloader.

An error appeared : "Programme operation not supported. Error while burning bootloader".

I could discount the peripheral boards as these were in 100% working order (from previous assignements) and were recognised without error by the laptop:




The board was now checked thoroughly with use of a digital multimeter, and with the help of a tutor.



There was one connection that did not display a 100% solid connection.
I re-soldered and tried burning bootloader again. The same error appeared repeatedly.

I also downloaded Python and ran the command to "flash" the bootloader, but it ran the same errors.




This was unfortunate as it meant that I could yet proceed to test the measurement of light into the phototransistor. I had compiled my code ready in Arduino and had managed to successfully interpret and correct all the errors. 

Shown below is the code to measure and count the yoga posture  sequences by use of the phototransistor:




Milling and soldering a new board

Having checked for possible errors on the board, it was possible that my Attiny processor either faulty or overheated (in soldering - although there was no yellowing on the board to indicate this), therefore I made a new board. For speed, I decided not to try and redesign the board in Eagle; I would use a physical bridge over the Attiny.

New board:



Connecting the board:



Programming the board:

Dowloaded NG .c code and makefile (placing and renaming makefile in dedicated folder on desktop). In terminal window, changed directory ('cd' into the folder) and then ran the program with 'make program-usbtiny'. For the first time, the new board flashed successfully.

Additions and modifications to the code

With the help of my mentor, the following parts of the code were added to NG code:

• led pin definitions;
• set threshold on the phototransistor for the led light to come on;

For my own record, I have extracted and included below, aspects of the code that was added or modified:

  


The board worked successfully... when the phototransistor was close to being covered, this would be recognised and the led would confirm this, as shown below:

 
   

 
It was noted that the phototranistor was very sensitive to any changes to the ambient light in the room. This would make it harder to establish a reliable light-to-darkness change range with my final project device.

When I repeated the test at home, the phototransistor was even more sensitive, as there was a flood of light from a near window. I experimented with pushing blutack around the PT to block ambient light. Once the body of the PT was encased in this opaque material the results were much more stable (without any oscillations) and repeatable:

  


I also ran the Python program via the terminal window (using "python hello.light.45.py /dev/tty.usbserial-FTHHQE77") and this provided a visual display of the varying states of the phototransistor, shown below:



 


 

Download Eagle schematic here
Download code here
Download traces here
Download interior here
Download code here