Slashdot Mirror
Ask Slashdot: Options For Cheap Home Automation?
New submitter goose-incarnated writes I'm looking at cheap and simple home automation. Unfortunately I'm not too clued up on what my options are. There are such a wide array of choices, none of which seem (to me) to be either cheap or simple. I'd like to:
Turn switches on/off (lights, wall sockets, general relays, etc); Read the status of on/off switches; Read analog samples (for example, temperature sensors); 'Program' switches based on analog samples/existing switches (for example, program a relay to come on at 30C and go off at 25C, thereby controlling the temperature); Similarly, program switches to go on/off at certain times; Record the samples of analog or digital inputs for a given time . I'd like to do the above using smartphone+bluetooth (for when I'm in the vicinity of the room), or smartdevice+WiFi (for when I'm in the house, somewhere), or even in a pinch, using HTTP to access a server at home from 600km away (which is what I'm willing to do). I'm definitely not willing to stream all my requests/data/responses through a third-party so third party cloud subscription solutions, even if free, are out of the question. Finally (because I know the Slashdot crowd likes a challenge :-)), I'd like something that is easily reprogrammable without having to compile code, then reflash a device, etc. What languages for embedded devices exist for home automation programming, if any. A quick google search reveals nothing specially made for end-users to reprogram their devices, but, like I said above, I'm clueless about options.
I'd recommend looking at Insteon. It is the closest to what you are looking for off the top of my head.
If you want to avoid going through some service they own/control, you might need to roll your own management system, but as far as cheap devices you can control programatically I think they're probably what you want.
I'd like something that is easily reprogrammable without having to compile code, then reflash a device, etc.
Architecturally you probably want the devices to be dumb. They can report information or accept commands. Leave any logic to some kind of centralized controller.
Bonus note: avoid x10 (if it's still even around). It's dirt cheap but pretty much the shittiest system out there. I lost much sanity to it back in the day.
Personally I did the x10 thing back in the day using at first an ocelot controller and then eventually my own hacked together system using the ocelot as a modem. The shittiness of x10 aside, I grew bored with it fairly quickly. It's all fun and has a neat "house of the future" feel, but I didn't find a great deal of utility in it, and what utility there was is already covered by purpose specific devices (smart thermostats, etc).
Bah, do what people have been doing for centuries ... have kids and make them get up and do it.
Lost at C:>. Found at C.
best part of DIY is spending three hours troubleshooting why the lights don't turn off when you could have just walked to turn them off
in NYC my electric bill is around $90 every month. breaks down to $25 for the electricity at $.095 per kilowatt hour, $55 for the wiring charge and the rest taxes and fees.
why spend $1000 to save $5 a month in electricity costs? because if i reduce my usage by 20%, that's $5 whole dollars a month in savings
You could look at maybe getting a Raspberry Pi and/or an Arduino board. There a quite a few 'home automation' tutorials/youtube videos out there on the subject with specific examples for door locks, motion sensors, temperature/humidity monitoring, automatic lighting, etc... This would most likely give you the most custom and cheapest option but requires building everything yourself, including the code if you can't find a working example.
All this is dirt cheap to do yourself if you're able and willing to spend some time to do it.
You can source Arduino Uno clones from China/eBay for as low as ~$4 apiece. A reed switch for a window/door sensor is ~$1-2, a Wifi module (ESP8266) ~$3.
You don't have to start from scratch putting everything together, there's plenty of schematics and code online to start you off (and complete projects).
You can get yourself a VPS to coordinate everything online, starting from $3/YEAR for a IPv4 NAT IP. Check out http://lowendspirit.com/
The home automation landscape has options differentiated by bandwidth, price point, device compatibility, and software capability. Low bandwidth devices are cheap but lack reliability, e.g X10. Expensive devices are often limited in device compatibility, Leviton. Basically, to make it all work you have to get the devices you need, then decide on a way to connect them all to a central server that can access them --- then start programming. You can see my home automation presentations at http://momjian.us/main/present....
The reason this is hard is that there is so much variety in what people want to automate. With a car, the manufacturer totally controls the environment, and when you buy a new car, everything is new. With home automation, the home automation provider rarely controls the entire environment, so implementing home automation always requires customization. I suppose the only easy option is to buy a new home and have the builder install everything from scratch with home automation designed from the start. However, that rarely happens because people rarely want home automation from the start, and even if you do that, the home automation landscape changes so frequently that it would soon be obsolete.
Cheap, easy, reliable.
Pick 2.
I've been happy, for several years now, with my Insteon setup. I think it will meet many of your needs:
:-( I renovated/re-wired my home such that all the wiring ran first to the switches, then up to the lights - you need this arrangement.
:-) In fact, Insteon contantly refers to "scenes" where you would do exactly that: press one single swtich to bring down your home-theater screen, close curtains, dim lights, etc. Personally, I don't use this feature
-For a central controller, I use the Universal Devices ISY-994i. This provides a web-interface for status, and quick toggle-controls, but it's also where I do my "programming"
-if you buy and install an Insteon switch, it will "just work" as a regular switch - others in your home will have zero learning-curve.
-they do have some universal I/O modules with dry contacts, etc. You may be able to work with these to perform your temperature monitoring and control, along with water-level detection, etc. See if it would work for you.
And, here is a bit more "geek" for you to ponder:
The Insteon signalling uses both a data-burst on the AC power-line, and also a radio-frequency data burst. Each device will repeat a burst (up to 3 hops only, or something like that) - in practise, I've found it really, really reliable. I started with just a controller and a couple of (AC-powerline-only) switches, but this required me to add a hardware phase-bridge, so data-bursts on one phase of my house-wiring could reach devices on the other phase. As I added more devices, especially more wireless devices, the mesh got better and better.
The Insteon switches will require 3 wires in your switch-box:
1) Hot wire. The Insteon switch will use a bit of power, as well this is what powers your light or other device(s). Typically it's a black wire.
2) Load wire. This is what goes to your devices.
3) Neutral wire. Typically this is white, and can sometimes be a problem. My older (1950's) home had all the hot-and-neutral wires run to the lights first, then a pair of wires ran down to the switches... this won't work
Each Insteon switch is internally pretty cool: it's comprised of two parts:
1) a switch part, that detects when you press the switch, and reports this data-event
2) a responder part, which receives a data-event and closes the circuit (with relay, or whatever).
So, you can actually have a single switch act to control several devices, if you want
The "programming" is not really very geeky: it's more just a set of conditions and actions, selected from drop-down dialogue boxes. That said, you can do some fun stuff, like:
- change actions based on sunrise-sunset times (great for lighting)
- use a motion-detector, which also contains an ambient-light sensor! I use one of these outside, to tailor my lighting to the Pacific Northwest's gray and dreary winter days.
HTH,
Use a Belkin Wemo switch: http://www.belkin.com/us/p/P-F...
I've written a simple utility to switch them on / off based on serial number:
See: https://github.com/d-w/toggle_...
Works for me.
Cheers,
Dave
PS: For extra points, wear & monitor a Neurosky EEG monitor: http://neurosky.com/products-m... ... yet)
Write a listener to recognise when you want the switch on (easy to recognise certain meditation patterns) and then use this to toggle the wemo switch closest.
(I've done this - just not on Github
I wouldn't trust either to control any device with actual destructive capability.
X10 doesn't need a path to the internet. With such a primitive protocol all it takes is a dying fridge or UPS to make devices randomly turn on and off (this is actual experience talking).
The Experimental Physics and Industrial Control System. It's commonly used on particle accelerators and will interface to anything. And it's open source. Runs on Linux, Window, OSX operating systems and all hardware including Raspberry Pi, Beagleboard, etc. http://www.aps.anl.gov/epics/i...
Central control means nothing. X10 is completely insecure. Your power line is a shared bus. I had to coordinate with my neighbor to use different X10 channels, because we kept turning each other's stuff on and off. Anyone can just plug in a controller, and every X10 house downstream of the step down transformer will see the signals and respond. At one point I had a sniffer/sweeper running to debug the damn problem and that's when I got my neighbor's attention cause his house went bonkers when I swept the A channel. He moved to B and all was well after that. But still very insecure.
X10 is lacking in many ways, and I still use some of it, but the biggest problem I've had was with durability. Many of the components were horribly made and just broke. Stanley, GE, Radio Shack, didn't matter. Old Sears parts lasted longer but still just died. I can't comment on how vulnerable the others are, but X10 is very vulnerableâ"from outside sockets and other units on the same transformer. North Koreans aren't going to hack your house from Pyongyang, but sneaky neighborhood kids can. I think you can filter the signals from the incoming power lines, but that's not commonly done. Also, you need to bridge the two incoming 110V legs so that X10 signals get across, otherwise parts of your house might not talk to other parts. Still, to play around a bit, there's no cheaper way than X10.
Many models don't function properly when installed in Mom's basement.
Some interesting things have been happening in home automation communications infrastructure standards. I'll preface this by saying that I'm not an expert, but have been trying to follow developments here for a few years for my own system. I can see four or five different media that you may want to consider. If you don't want to be stuck on an island, the problem to be solved is one of compatibility, especially looking towards the future.
Power Line Communications - these are attractive because your house is already wired for them and the source of power for a device interface is rather obvious. Unfortunately, many solutions are low bit rates, unreliable, or proprietary. Universal Powerline Bus (UPB) seems to be the current raining champion here. Much more reliable than X10, but slow slow slow (240 bits per second). Works for on/off, but won't be useful for higher-end sensor applications. There are other low-speed PLC approaches, some of which are faster (4800 baud even!). What may be the way of the future in PLC is HomePlug AV and HomePlug Green PHY. The former is a high-speed PLC approach (up to 200+ Mbps) that costs a lot per device, the latter is a dumbed-down version of the former that has a much lower bit rate but supposedly costs 75% less in both $$ and power. Its bit rate is still supposedly much higher than UPB. Here the target is something like 250kbps which is touted as necessary for "Smart Grid" applications. IEEE 1901 is a standard that incorporates the HomePlug AV technology; I don't think it includes the HomePlug Green PHY yet, but HomePlug Green PHY is supposed to be compatible (on the same power line media) as HomePlug AV.
Low-voltage bus - here you have to run wires. Ethernet is the reigning champion, running over UTP, coax, fiber (ok - not exactly low voltage but it fits the mood). Great bandwidth, great reliability, but sucks down power and costs $$ for interfaces. PoE makes this useful for remote sensors/controls that don't have a local source of power.
High-speed wireless - you don't have to run communications wires, but the solutions here require more power than you want to supply if running off batteries in a remote location. 802.11 is the champion here.
Low-power wireless - lower bit rates (250kbps) but much less power consumed so that even battery-powered remote devices can be used. ZigBee is one example, which I think is interesting, because...
Ideally you want to be able to tie all of these together so that you don't need to plumb multiple media interfaces into your controller, and ideally you want to be able to use TCP/IP as a high-level protocol. Here is where IEEE 1905 comes into play. The IEEE 1905.1 standard provides an abstraction layer to established powerline, wireless, coaxial cable and Ethernet home networking technologies - IEEE 1901 / HomePlug AV, Wi-Fi, MoCA®, and Ethernet. I think there is some effort underway to get ZigBee into the fold as well.
If vendors see an advantage in following the standard, and interoperability becomes a selling point, perhaps we'll get away from the multiple proprietary islands.
A whole lot of string and pull switches. My dad did the whole basement lighting system that way, back in '79. As for scheduling, rig something up with egg timers. Eye hooks, duct tape, and popsickle sticks are also recommended. Maybe a toaster and a pulley. YMMV.
Everything's been downhill since the TRS-80