Smart Home Shade Controller

This project includes the software and 3D printable parts for an Alexa enabled controller for roller shades. It includes: * A web application that manages shade controllers on the local network, relays commands from Alexa, and provides a simple web UI

* A Node-JS script for an Amazon Lambda service to relay requests
  from a custom Alexa skill to the web application

* An arduino sketch for a NodeMCU microcontrollers

* OpenSCAD files for the housing and driver wheel for the
  controller

Building The Server

First, edit shades_server.ml to include the email address associated with your Amazon account as the authorized_email. This is used to authorize requests from the internet (notably from the Lambda function) using the Login With Amazon auth credential. At line 6, replace the string "[email protected]" with your email address.

$ sed -i -e 's/user@host\.com/[email protected]/' server/shades_server.ml

Next, install the dependencies, assuming you have opam installed:

$ sudo apt-get install mosquitto
$ opam depext mosquitto
$ opam install jbuilder core opium lwt yojson ppx_deriving_yojson mosquitto

Then compile

$ cd server && jbuilder build shades_server.exe

The server is intended to be run a Raspberry Pi Zero W with a touch display for the web client but should work fine on any system capable of running ocaml. Building the dependencies (notably core) on the Pi Zero W may require provisioning a swap disk:

$ dd if=/dev/zero of=./swap.img bs=1024 count=$((1024 * 1024 * 4))
$ mkswap ./swap.img
$ sudo swapon ./swap.img

Hardware

The SCAD files are designed around the following components for the controllers:

* NodeMCU ESP8266 microcontroller + Motorshield:
  https://www.amazon.com/ESP8266-Development-NodeMCU-CP2102-Shield/dp/B075VMNLZR

* Pololu 99:1 Metal Gearmotor 25Dx54L:
  https://www.pololu.com/product/1587

* Batteryspace LiFePO4 18650 Battery (6.4V x 1500 mah):
  http://www.batteryspace.com/lifepo418650battery64v1350mahflat864wh4aratewithpcbandpolyswitch.aspx

* Amrka 6V 12V 10A Auto Solar Panel Charge Controller Battery Charge Regulator PWM
  https://www.amazon.com/gp/product/B072WSQPJN

* US100 Ultrasonic Sensor:
  https://www.bananarobotics.com/shop/US-100-Ultrasonic-Distance-Sensor-Module

* NUZAMAS 3.5W 6V 600ma Mini Solar Panel Module:
  https://www.amazon.com/gp/product/B071R3NQBP

The solar charging isn't fully worked out yet, I'm currently using one of the above listed panels and a smaller 330ma panel, I'm planning to switch to 3x600ma to (hopefully) get better results.

The SCAD files should be exported to STL and should be printable on any 3D printer (I used a M3D Micro which is about the smallest printer on the market). The parts are:

• driver-wheel (x1): the actual wheel to put on the motor shaft and loop the shade string around. Assumes a simple rope for the shades. I recommend putting a few rubber bands around the wheel to increase friction. I print this with model-on-model support enabled.

• housing (x1): the main housing for the controller, battery, motor, and solar charge regulator. Print with model-on-model support enabled.

• housing-lid (x1): a cover for the main housing to make everything neat and tidy.

• motor-top (x1): a rounded piece to fit between the motor and the top of the housing, this gives the thumbscrews something to push down on to tension the motor/wheel into place. No support needed.

• screwplate-6mm (x2 or 3): these should be inserted in the motor housing above and below the motor. One is inserted at the top of the housing to give the thumbscrews something to push against into all the layers (so the tension isn't shearing across the layers of the housing). One or more may be inserted below the motor to give it something to rest against (you may need to vertically scale this depending on your mounting).

• ultrasonic-sensor-us100 (x1): split into two parts, base and shell, to house the ultrasonic sensor. Export once with base = true and shell = false, and once with base = false and shell = true (lines 23 and 24).

• solar-panel-hook (2x # of panels): this is designed to hook to the underside of a siding J-strip to mount the solar panels below a window.

• settings (x0): don't print this, it's just a master settings file included by the other scad files to ensure consistency.

See the thingiverse page: https://www.thingiverse.com/thing:3014140 for pre-exported STL files.

Configuring Controllers

Individual controllers are configured by modifying the config.h file in the shade-controller subdirectory. Just edit this file to include your WiFi network SSID and password (assumes WPA2), and for each shade controller substitute in a unique id, a human friendly name, and a description.

Once you've edited the config.h file, use the Arduino IDE to compile and upload the firmware to each control unit.

Wiring

• The motor should be wired to the Motor A port of the motor shield
• The ultrasonic sensor should have TX connected to D8 and RX to D7 and voltage and ground connected appropriately
• The battery should be wired to the Vin/Gnd terminals on the motor shield, with the Vin/Vm jumper installed

Calibration

Prior to first use, you need to calibrate the maximum height of your shades. With the controoler on, log into your MQTT broker and use the mosquitto_pub command to issue the CALIBRATE instruction.

$ mosquitto_pub -t shade-controller-00 -m "CALIBRATE"

This should cause the controller to run its motor to raise the shade until it stops moving, then set the current distance as the max distance. In practice, it's easiest to just manually open the shade all the way prior to sending the command. You can also use SETMAX mm to set the maximum distance to mm in millimeters.

$ mosquitto_pub -t shade-controller-00 -m "SETMAX 1200"

Once the controller has been calibrated it will store the maximum distance in EEPROM, so it shouldn't require recalibration. Adding a UI for calibration is on the todo list.

Alexa Skill

Create a new Amazon Alexa Smart Home Skill. See https://developer.amazon.com/docs/smarthome/steps-to-build-a-smart-home-skill.html for details.

As the backend, create an Amazon Lambda function, and upload the file alexa-backend/index.js

In the lambda function configuration, set the process environment variables BACKEND_HOST, BACKEND_PORT to the IP address and HTTPS port of your server respectively. And set USER_EMAIL to your Amazon email address (used for authorization).

Enable the skill in your Alexa app, then login with your amazon account information and it should pick up all of your shades.