Alexa, turn on Red Alert!

Red Alert

I love Star Trek (any good SciFi really) and who doesn’t? I also have an Amazon echo and I have been making my home smarter and adding automation where I can, very SciFi’ish yeah? The other day I thought to myself how cool would it be if I could activate a red alert”? Well I do have the Dot and I do have Philips Hue bulbs. I do have a home server and I do have Raspberry Pies. So I decided I wanted to have this feature and set out to get it done. As it turns out I am not the only one that wanted to be able to do this. After Googling I found a few other people that went through a similar process. Although I have not seen anyone do it the way I did. The ones I found mostly accomplished it with a Google Home Assistant and some used Node-Red. Hey, I have Node-Red. I guess the Home Assistant can play audio files. I have found a few pages on Alexa doing this recently but I have not gotten into making skills, yet. All the pages I saw also seemed to rely on an outside service of some kind (minus the voice assistant). They pulled the audio from the web or used IFTTT (which I hate) to do something. I don’t want that. I like to be as self contained as possible. Here’s what I did.

I have my home server setup with NR and it takes it all the MQTT in the house and does all the NR handling for the house. Then I have a Raspberry Pi (that also sits on top of the server), this has a temperature sensor on it and it handles the audio portion of the red alert. Since it is sitting there I also have it monitor the server, and the server monitor the Pi.


Yes I have a server and a Pi, yes they both also run NR. Why do I not use just the server instead of the Pi? The server doesn’t have a sound card and I don’t have an extra one. So yeah.


Flow 1
Flow 1

This flow is where I make the color changes. I use the Wemo Emulator node to create a device Alexa can discover, that also allows me to choose my own trigger word. That node gives a 1 or a 0 (on or off). I pass that to a function that contains the hue bulb color and activates the alert pulsing, this all flows to the hue bulb and out via MQTT. In the flow I only have one hue bulb connected. I have since connected the red alert to all the bulbs I have (currently 4 colored ones).

Flow 2
Flow 2

Here is where I activate the audio. I initially tried omxplayer. I found a shell script that looped the audio but it gave me issues when trying to kill the process. It only worked the first time. The processes didn’t die completely. I want to be able to stop the red alert also. So I continued searching. I came across a post in the NR Google groups and they had created a flow for playing a sound off of motion detection. I was able to take the kill (killall, duh) command he used to stop my flow. I also used the player (mpg123) he used because it has looping options builtin. Sweetness.

I am now able to tell Alexa to turn on a red alert and have her stop it as well. This doesn’t used IFTTT or rely on any other outside source (besides Alexa). It does not require an Alexa skill either. I have future plans to make a custom skill for this so I can change the phrase from “turn on red alert” to something more comfortable like “activate red alert”.

Onward to the flows!

Flow for the red alert lights:

Flow for the red alert audio:

(You may notice that the flows are different from the images, I cleaned up the flow before exporting and pasting the code.)

UPDATE 8/29/2017: I added updated code/flows so that you can disable the Red Alert lights. Previously a “stop node red” command would silence the sounds but not cancel the flashing lights. With the new update the sounds stop and the lights stop flashing and turn white. Still working on getting the lights to default to white after the Red Alert times out. 



Stop Red Alert


Some pages I found helpful.

Where I got the red alert sound

Play Ambient audio on motion detection!searchin/node-red/audio|sort:relevance/node-red/vwQq8Plk0Zg/6DV5ZYMRCAAJ

The code for the ambient audio

Google Home Assitant and an RPi with video

Static IP on a Raspberry Pi running Jessie

Raspberry Pi

I know this is out there already in the Googlesphere, this post is more for me in the future than it is for you. I have a few Raspberry Pies and I like to run static IPs. It just makes things easier and I like having my network “all fancy”.

I went to setup a static IP the other day and it wasn’t working. I rebooted a few times until I figured out something was up. In my Googling I found out that the standard linux way of assigning a static IP address won’t work with the new version of Jessie for a Raspberry Pi. They made some updates and the old configuration now gets ignored. So editing the interfaces at  /etc/network/interfaces is a no go. I mean you can, but it’s pointless. Not to worry the new way is just as easy as before it is just located in a different file now. Now you need to edit  /etc/dhcpcd.conf, and the way you set the address is slightly different.

The old way of doing things:

And the new way of doing things:

It has become a little easier if you ask me, you just need to know where the new file to edit is.

All thanks to this page:

Automated Keyboard Light with Alexa

Since I have been fiddling with Alexa I was able to get a light working with Wemo emulation. Both on the Raspberry Pi and on the ESP itself. I am mostly using the ESP with Fauxmo to act as physical devices. The Wemo emulation being done on the Pi is for running a bunch of scripts with MQTT or (hopefully in the future) gettin’ data from sensors and such. Still trying to find a way to get Alexa to read whatever I give her from MQTT, that would be righteous. But for now I have an automated keyboard light with Alexa.

(TLDR; Made a keyboard light on an ESP with a relay that emulates a Wemo plug and is voice activated by Alexa. Skip to the bottom for the code I used.

Any who, I replaced my old keyboard light switch made out of an old telephone biscuit jack with a toggle switch. I upgraded. I can now voice activate my keyboard light with Alexa. Man I’m lazy, and man that is cool. Not the lazy part the keyboard light. I do have to admit this was not my first attempt at this build. I tried two times before I finally got it right. The first two times I was trying to use 2N2222 and 2N3904 transistors and neither would work right for me. I was able to get it all working on the breadboard just fine but as soon as I transferred it to a PCB it failed. I think the problem is with the transistor. From my measurements it keeps leaking 12v back through the base and I don’t know enough about electronics to be able to figure it out yet, obviously, I tried twice.

So the third time I used the pre-made modules I have; 5v relay module. I put together a small PCB for the ESP and a DC-DC converter and added some pins to use jumper wires to attach to the relay. Soldered the power to a barrel jack and hooked up a toggle switch and connected it to the relay. So if I flip the switch it bypasses the relay and I get light manually. Always good to have a backup. The switch will work with or without the ESP plugged in. I plugged it in and bam! It worked. I gave Alexa a few commands and on and off the relay clicked. Beautiful.

Then……it failed, it started flickering the relay. It took me a minute to figure it out. I forgot the current limiting resistor on GPIO2 for the relay. Oops. That’s an easy fix luckily. The green jumper wire in the pictures goes to the pin header from GPIO2, so all I had to do was remove the jumper wire and replace it with a 1K ohm resistor. Easy. It was getting late so I turned it off and removed the ESP. The next day I go over to my computer and I can smell the lovely aroma of burnt electronics. Fuck. I look down and I can see the DC-DC converter sparking on the underside of the PCB. Turned out to be a bad solder job on my part. Since my liver transplant I have to take a shit load of pills, and some of these pills cause my hands to shake. Sometimes it’s not so bad and other times it’s ridiculous. I guess they were shaking more than I thought that night.

So I had to rebuild the whole thing. Again. Live and learn. This time I was sure to leave extra space in my solder routing just in case. The Mark IV has been up and running with zero problems for two days now. I think I worked out the kinks. And it is awesome to be able to sit down and tell Alexa to turn on my computer room and keyboard lights. Hell with Node-Red I could even WOL my computer!

Now behold, pictures…

View Post

Alexa Skills

Amazon Echo Dot

I purchased an Amazon Echo Dot a couple weeks back. In the past I had stated that I would never put one in my house, and generally had a distain for voice activated things. I had done a little bit of reading on the Dot and I have been getting further into home automated things. I finally had a light bulb moment where I now understand all the coding. I am now dangerous. Haha.

Logically the next step was integrating Alexa with my custom build devices. All of my devices work off an ESP or Raspberry Pi (or will). I found Fauxmo Wemo emulation which is fantastic. That covers about 60% of my use, a simple off and on. But for things like my temperature sensors, hyrgrometers and water lever sensors I need to be able to pull data and have Alexa read it back. Like “Alexa, ask server for the temperature.” Or “Alexa, ask the garden for an update.” All of my devices communicate via MQTT. I started with Node-Red without Alexa so this is the direction I went. I still think it was the best route to go. All I need is an Alexa skill to read back different inbound MQTT topics (for different devices). I can’t believe there isn’t a skill already out there to read back MQTT data or homebrew temperature sensors. Or is there and I haven’t found it yet? The majority of the available Alexa skills fucking suck. There are very few actually useful skills to use. The majority of the skills I have seen are garbage. One off skills you would use once or twice then forget about. More Alexa joke skills than there needs to be.

So I started looking into making my own skill. This is what I have found to get started with creating a skill.

  • You need an Amazon AWS Developer account
  • You need access Lambda
  • You must subscribe to basic free tier of service
  • You must subscribe to Amazon EC2, which is only free for one year.

So is this a catch with using custom skills and Alexa? Will my skills then start costing me money after 12 months. If I cancel after 12 months do I lose all my skills? I’m thinking yes, I will bet you Amazon is planning on cashing in on this later? Or is it already that late in the Echo game that I am just being unfoundedly paranoid?

I have not yet created a skill. Javascript is not my forte and I haven’t sat down to dig in to this yet.


Some info that I found hard to find:

Great walk-through on creating a new skill:

ESP-01, MQTT and an OLED (SSD1306)

OLED ScreenIt took me about a week to get this going. There are still a few kinks that need to be ironed out, or worked around. But I wanted to get this out there before I forgot about it. I could not find a sketch for using an Arduino or ESP with an OLED and MQTT. Not one that didn’t just display data from a connected sensor., that is not what I wanted. I wanted to be able to send messages to the OLED screen via MQTT. I spent quite a while looking for it and I couldn’t find anything. About a week or two ago I finally broke the wall that was keeping me from sending and receiving MQTT message on an ESP. So I took that code and the working code from the Adafruit OLED sketch and made a baby.

I was able to produce a sketch that will simply display any text received via MQTT. Perfect. I plan on using this cobbled together with Node-Red. That way I can have one screen and send multiple sensor readings to it with minimal coding and parts. There are a few things I still need to figure out. For example I can’t get the screen to clear. Perhaps thats what the OLED reset pin was for? But my OLED only has four pins; SDA, SCL, Vcc and Ground. No reset. What to do? I think that is actually the only thing I need to work out. The Adafruit code will wrap your text to new lines, so make a note of that. I have gotten around clearing the screen by injecting spaces via MQTT. Its a little more effort in Node-Red for now until I get that down. But it works. It doesn’t matter how you get there as long as you get there right?

I Program the ESP via the Arduino IDE, if you don’t know how to set that up take a stroll down Google lane. Tons of help there on that topic. As it stands in the code below, once powered up the OLED should flash the Adafruit logo until a message is received then it will display the message until a new one is received. Simple. The default topic is “inTopic”, and the default OLED font size is 1. I have tried a font size of 2 but no higher. I also stumbled upon this library here. Much smaller and designed just for text. May try that in the future, but if it ain’t broke don’t fix it.

Items Required:
(1) ESP-01 (mine is a standard ESP-01 from eBay)
(1) OLED I2C Screen (I used a 0.96″ 4pin OLED also off eBay)

IDE Library Requirements:
Adafruit GFX
Adafruit SSD1306

Some places that helped

This was the tutorial that helped me get what I have going. I couldn’t get the OLED to work quite right until I came across this post. I used it as the base for what I have.

This page helped but not until I specified the SDA, SCL pins for the ESP. I couldn’t get the display to work properly and do what I wanted.

And the code (below): [See the comments for updated code]
View Post