Hmm, maybe I shouldn’t call it the MatDuino. But I need a name for this project.
It stems from my recent musings about how wireless devices in the home, specifically #IoT devices, are subject to so many more potential problems than wired devices.
Things that got me thinking recently:
- Numerous comments on my YouTube videos about how I should consider using the ESP8266 instead of Arduinos wired into my network, each one I have rebutted;
- Reading this article on the Automated Home site, which is on the subject of inconvenience and security of 2018 Smarthome devices, not really wireless – but my thoughts went off at a tangent;
- This great video by Jon Oxer of SuperHouse.TV, who argues “wired over wireless every time”
- This chat on openHAB community which started out as “can someone recommend an RGBW device”, ending up with the inevitable “hmm, you should probably build your own, given what you need”
- Finally, culminating in my Yale Home Alarm being jammed from 8am this morning. This refers to the connection between wireless sensors and the bridge, which I understand works on 433MHz RF. Thanks to @PenTestPartners for writing this superb article on the subject: Alarm Systems Alarmingly Insecure, Oh the Irony
I have always loved the combination of Arduino Nano and Funduino (the W5100 network breakout board) for connected devices in the home. I’m talking about everything in the DIY #homeautomation home, for example RGBW strip drivers connected to the network, PIR sensors, temp / humidity sensors, etc.
I find housing both boards in a single enclosure is cumbersome and really takes away from the “nano” feel. My rule of thumb is: can you fit it in an electrical backbox? Yes, then we can start embedding low voltage things around the home (assuming you’re designing your house from scratch – I am).
Similarly I love the Mega for its vast IO count, and likewise one of these things with a W5100 breakout board in a wall backbox is a bit too tight for space.
Enter: the MatDuino. (Better names on a postcard please.)
The MatDuino, and the MegaMatDuino are two boards that exist in my mind – but are next on my project list – with the following design decisions:
- Based on ATmega328 and ATmega2560 respectively. I’d be using the TQFP package, not the DIP – small is the name of the game!
- So yeh, need to learn to solder uber-small things
- PoE networking.
- Will trade things like the onboard serial for PoE network using WIZnet W5500 or W5100. (Which to use? Does Arduino support W5500 in the default Ethernet library? I need compatibility with libraries, more than newness, on this one.)
- No serial / TTL onboard, but instead, the smallest 4-pin interface practicable, which we can plug a cheap USB TTL programmer into, for bulk programming of boards (i.e. sensors across the home)
- This saves space usually taken up by the Nano’s CH340 chip and the micro USB port, and the Mega’s chip and humongous USB port! (WHY ARDUINO, WHY?!)
- The board will expose as much IO as humanly possible. This means we will use serial TX/RX (i.e. we can use Pin 1 in our projects) as long as we add a tiny SMD toggle switch to isolate this pin during upload of sketches
- No interfaces for wireless – just pure IO for sensors, displays, etc.
- The thing should be really quite small. First version won’t be too small, as it needs to work. After that, we miniaturise.
- Oh, did I mention PoE? You know I’m crazy about including a Silvertel PoE chip so we can have 802.3af powering our projects. Take two bottles into the shower? Not me! We need one cable powering these things in the home, and it’s going to be the same one as the one we already need for a wired network device, dammit.
- Some kind of header layout that allows for adding top-boards. I’m thinking, my wall MQTT wall control panel (ATmega2560), my yet-to-be-built multisensor (ATmega328), and whatever else. Flexible form factor for these devices should be supported.
Anyone else interested in this build?