We finished last time with a complete electronics prototype of our Netflix switch. Next we must implement the code to make this a working system.
By now we know what hardware we have available to us, but what are we actually trying to achieve? The goal for the original switch from Netflix was to turn on the TV and perhaps launch the Netflix app (if you have a compatible TV); perform a little light show using the LEDs we hooked up previously; and send messages to another internet enabled device. This other device could be used to extend the functionality of the switch; making API calls or performing some home automation.
Code
I have hosted the code as a 3 part gist. The main file is Netflix.ino and orchestrates the hardware we have already hooked up. The other 2 files implement a simple class that will generate a carrier wave from the Infrared LED enabling it to be used as a remote control. The specific carrier wave will be unique to the TV you have and needs to be known before compiling the application.
IR Carrier Wave
For the switch to work with your brand of TV you must first know the encoded signal to turn it on and off. This signal is just a series of pules emitted from the IR LED that we can represent as an array of timing information (how long to turn the IR LED on, how long to turn it off, how long to turn it on, etc…).
There is a great library that can capture this information for us written by Ken Shirriff. Specifically you want to be looking at this file but you will need to download and extract the entire library for this to work.
You will need to run this on a separate Arduino hooked up to an IR receiver.
This is fairly simple to do requiring very little setup. More information on
this can be found on
sparkfun. Once you
have your raw signal, the output should be assigned to the data array within
Netflix.ino.
Messaging Queue
As an optional part of this project I have setup a remote messaging queue on Heroku that I use to dispatch events when the switch is activated. This allows any arbitrary code to listen and react to these events and perform further actions.
The Particle SDK does already provide a cloud based service that can be used to publish events to any listening subscribers. However this requires the subscribers to be using the Particle library and tightly couples any application to the Particle ecosystem.
I prefer to avoid unnecessary coupling where I can and instead recommend setting up a RabbitMQ instance on Heroku. If you too are planning on implementing this feature, you’ll need to update the username and password sections of Netflix.ino.
Any subscribers now need to listen on the home/netflix topic of the RabbitMQ
server and can perform any action you want. I intend on writing a touchscreen
GUI application for the Raspberry Pi that will order me Dominos Pizzas when I’m
watching a movie. But you can create whatever you want.
Particle Build
Particle supply an online IDE from where you can compile and upload your sketches to your Particle Photon board. It is pretty basic, but that is all it needs to be.
Here you can copy across the code into 3 separate files, named exactly as they are within the gist.
If you followed along with the messaging queue section of this article you will want to include the MQTT library in your sketch using the libraries tool. If you do not intend on implementing this particular feature you will want to comment out anything relating to it otherwise the code will not compile.
When you have verified and uploaded the code to your Photon it should work just as mine in this video:
Future Work
From my time spent working with the Particle Photon I have noticed a lack of simplicity when working with the hardware timers available on the board. These timers would enable me to afford the multithreading support I intended to generate the Infrared carrier wave. Consequently I have been looking for an alternative WiFi enabled microprocessor.
RedBearLab WiFi Mini
The RedBearLab WiFi Mini is the most affordable WiFi capable microprocessor available at the time of writing that is also compatible with most Arduino libraries. It is built with the Texas Instruments CC3200 dual core MCU: an ARM Cortex-M4 core running at 80 MHz with a dedicated ARM core for WiFi network processing. It also features Over The Air (OTA) download of application firmware from other WiFi devices or the Internet, enabling similar firmware upgrades as the Particle Photon.
However the RedBearLab WiFi Mini cannot simply replace the Particle Photon. It requires its own IDE and some of the lower level features of the board (hardware timers) require more advanced C programming, which I will need to work through before I commit to working with this board.
Timely Library
The Timely library should go some way
to make this project possible using the RedBearLab WiFi Mini. Timely is an
Arduino compatible library capable of controlling the TIMER registers of both
AVR and Texas Instrument microprocessors. This is currently still a work in
progress and I’ll update this article when it is stable.
Conclusion
You should now have a fully functioning system that can turn on your TV, possibly launching the Netflix app if you have a Netflix recommended TV; put on a bit of a light show, and optionally dispatch messages to a remote messaging queue. In the next article we will finish off our switch by housing it within an enclosure and giving it the wow factor.