Wi-Fi Light Bulbs Shipping Soon

An anonymous reader writes "Computerworld has an interview with an Australian startup called LIFX, producing WiFi-connected LED light bulbs. Each light bulb is a small computer running the Thingsquare distribution of the open source Contiki operating system that creates a low-power wireless mesh network between the light bulbs and connects them to the WiFi network. The wireless mesh network lets the light bulbs be controlled with a smartphone app. Through a Kickstarter project, the company has already raised a significant amount of money: over one million USD. "

Comments

Wi-Fi toothpick

[justthinkit]

I'm waiting for the Wi-Fi toothpick.

Re:Wi-Fi toothpick

[ganjadude]

yeah, and the wi fi doors.

look, i dont want my 45 cent light bulb costing me 50 bucks. I dont need a light bulb with a computer in it, can i think of fun things to do with it? sure but when i have over 100 light bulbs in my home, i dont want them all costing me a months pay to replace. what is wrong with a good old fashioned light bulb??

Re:Wi-Fi toothpick

[thereitis]

Seems to me that wifi-enabling the light switch would be more useful and cost-effective (for most people) than doing the same to the bulb.

Re:Wi-Fi toothpick

[tibit]

I think it won't be too long till the added cost of Wi-Fi in a bulb will be a couple of USD. Remember: a Wi-Fi chip has a microcontroller inside of it. That microcontroller should be enough to run Wi-Fi and a simple mesh network. It doesn't need a full-blown webserver, but even that could be done on a micro. The volume lets you optimize the heck out of everything. It would cost $0.0 in materials to have this chip control the light that already needs to have a power supply built into it anyway. In fact, the added on-top-of-Wi-Fi chip cost might even be negative if they let the Wi-Fi chip do power regulation in firmware as well. That way you lose the voltage regulation silicon, so that may cut some cost. Why the heck not, every modern industrial servo drive does all of the control in software anyway, and it's a couple control loops worth more complex than a power supply :)

Re:Wi-Fi toothpick

[Anonymous Coward]

switching 110/220 is a big deal involving mechanical relays that tend to stick open/closed or flutter when they fail.

what's more interesting is someone writing a virus/trojan that scans for these devices and then tries to trigger an epileptic fit by flashing all the lights on/off when it's night time.

Re:Wi-Fi toothpick

[davetv]

a semiconductor device - a triac - is more widely used for AC switching than relays

Re:Wi-Fi toothpick

[raymorris]

what is wrong with a good old fashioned light bulb?

It used energy, and was made of sand. Therefore your 50 cent light bulb needed to be replaced with a $50 biohazard made of mercury and other toxins sold by campaign contributors.

Re:Wi-Fi toothpick

[willy_me]

Yes, but they result in very dirty power that tend to kill cheap ballasts. If you're making a switch for lighting, a relay is the way to go if you want to support compact florescent bulbs. But triacs are good for incandescent bulbs.

Re:Wi-Fi toothpick

[hairyfeet]

If it costs just 10c its STILL stupid and wasteful, okay? Just because a dumb thing can be done cheaply doesn't keep it from being a dumb thing, especially when others have pointed out a VERY obvious way to get the benefit without the stupidity, and that is putting the Wifi in the socket NOT in the bulb. This would keep the electronics farther away from the heat, let you build a better antenna because you'll have more room, it just makes a hell of a lot more sense to just put it in the socket than it does in the bulb.

Personally I'd go one better and put it in the switch, as most wall mounts are in hallways anyway, putting it in the switch gives you an easy manual override (the switch itself) and since one switchplate can have 2 or 3 light controls you could use less chips by having one master control say 4 switches and cut the cost down further.

Re:Wi-Fi toothpick

[tftp]

Putting the brains in each bulb makes it more generally accessible and effortlessly scalable.

Unless cost is a factor.

Re:Wi-Fi toothpick

[thegarbz]

Erm sorry but that's false. TRIACs only result in dirty power if they are used to chop part of the sine wave such as in a dimmer circuit. If you connect it high via a transistor or optocoupler it will start conducting from the very start of each half cycle and will not result in any harmonic distortion.

The only time you'll see dirty power while a TRIAC is used as a simple switch is when it's not conducting. The leakage current is not constant. However if a few milliAmpers are likely to kill your ballast it was well and truly time to replace it anyway.

Re:Wi-Fi toothpick

[FuzzNugget]

Cost vs. what? Plain old "dumb" bulbs? Sure, but then you're obviously not interested in this type of thing anyway.

But cost of this vs. other systems? OK, maybe a bit more expensive, but you're *getting* so much more. LIFX is an open system, uses existing standardized networking protocols and is *programmable*. That last point, I think is what makes LIFX so much more than other systems (oh, you can turn your lights on with your phone and select 8 different preset colors? How cute!)

I'm not astroturfing, I swear, I've been excited about these totally hackable lights since I heard about them months ago.

Re:Wi-Fi toothpick

[tftp]

I use Insteon and Z-Wave at this house. They run on cheap 8-bit processors and do not require fancy clocks, or complex modulation, or multiple channels. WiFi is overdesigned for the task, is chatty, and needs configuration of some sort.

If you are interested in home automation, the last thing you want to do is to jump on a technology that is advertised to you through Slashdot. There is not much you can do with a single light bulb; however if you get a proper set of sensors, switches and stuff (Z-Wave on more expensive end, Insteon in the middle, and X10 at the dirt cheap prices) then you can build a usable system out of that, one that detects your activities and does the right thing. A single light bulb might be fun to play with ... for about five minutes. To do more you need more. Insteon market has many devices that do what you need. Z-Wave manufacturers make new devices every day.

Re: Wi-Fi toothpick

[Mashiki]

The bulb will last more than 15 years. Your arguments are ludicrous.

I seem to remember them saying 5-10 years on CFL's, odd that up until I dumped them all and went back to incandescent's, I'd replaced a dozen of them at least twice--though under warranty until I'd simply had enough.

Hell, even the new replacement 36" mini bulbs that they're pushing to replace the 48" florescent tubes, rarely last 2 years. The bulbs might last a year, maybe. And I've replaced 8 arrays in the last 4 months(all with a standard 2 year warranty), made by sylvania, and phillips. The 48" jobs that I still have, have ballasts made in the 80's and are still working. Hell I've got one tri-bulb 36" assembly that was used in street lights in the 70's where I live, and the ballast is still good.

Re:Wi-Fi toothpick

[Nutria]

If you're still using incandescent lights an LED will save a few times its cost in electricity.

Since LEDs are a metric ass-load more expensive than incandescents, they'll need to last an appropriately metric ass-load longer than incandescents.

But since CFLs were also supposed to last 10-15 years but don't, color me skeptical in believing LED light manufacturers claims of light bulb nirvana.

Re:Wi-Fi toothpick

[PNutts]

And for full disclosure, I discovered as I replaced more incandescents with LEDs my natural gas consumption increased ever so slightly during the winter because I wasn't heating rooms with 60W / 100W bulbs any more. I don't have air conditioning, but if I did that increase would be offset by reduced cooling costs in the summer.

I don't get nearly as cranky when my daughter leaves her closet light on all night. It now uses 6.5W instead of the 60W bulb that came with the house and then the 40W I replaced it with.

Re:Wi-Fi toothpick

[Anrego]

x10 .. x10

Somewhere in the recesses of my brain that sounds fami.. oh god.. OH GOD!

Kidding aside, I played with x10 for a while and if anyone is thinking about it, my suggestion is: don't

It's a terrible and outdated protocol. A quick "of the top of my head" list of the major problems:
- It's one way (for the most part). There was a kind of handshake thing out there but it was never used.
- The signals are easily lost in what were called "signal suckers" in many x10 circles. Basically any device using cheap filtering could kill a signal. This was a bad combination with the first one. It was common recommended practice to send a command 3 times at a 2s interval..
- False positives! The protocol is insanely simple and came from olden times when there were generally few noisy devices plugged in. The result is the right burst of noise can actually be a valid message and result in anything (but normally it was your bedroom lights turning on in the middle of the night).
- Slow. I don't know what the actual command throughput was.. but it wasn't good.

The whole thing was a terrible experience, and ultimately the novelty of it dies pretty quick. The very few useful implications are easily dealt with using much simpler technologies. One of the nicer things was always turning off the bedroom lights while laying in bed. Now I've got a self contained wall switch/remote dealie that works _perfectly_ and didn't even require a neutral ground wire or anything.. literally just swap and go.

I still have most of my old x10 gear. I will usually pull some of it out during christmas time.. few appliance modules controlling christmas lights and such.. but I'd never even think of trying to automate a home with it.. stuff is garbage.

Re:Wi-Fi toothpick

[dbIII]

Since LEDs are a metric ass-load more expensive than incandescents, they'll need to last an appropriately metric ass-load longer than incandescents.

That's exactly what they are doing with traffic lights.

Re:Wi-Fi toothpick

[Anrego]

Never really looked into z-wave, but I've heard insteon described as "x10 with the bugs worked out", which sounds nice. The idea behind x10 made a lot of sense, the implementation was just terrible.

Unfortunately my home automation craze ended. Kind of depressing really, it was real "house of the future" stuff and I was totally into it when younger. Now that I actually have the money and knowledge to do it properly, I no longer have the motivation.

I guess the gee-wiz factor of it died for me. With my remote controlled bedroom light, automatic coffee maker (or more commonly the tim hortons on the way to work..), and "smart" thermostats.. there's really not much else that makes sense to automate for more than the novelty of it. I know we are supposed to look at the "but why" question with distain.. but I can't help it :(

Re:Wi-Fi toothpick

[fuzzyfuzzyfungus]

While I agree, architecturally, legacy infrastructure has serious inertia, which results in the world being largely held together by a mixture of dumb choices and dirty hacks laid on top of antiques that nobody wants to replace.

Even in situations where there is a logically-separated arrangement widely available(as with fluorescent tubes, where mechanical and electrical standards for fixtures with discrete ballasts have been established for decades), the market is still flooded with ghastly all-the-driver-electronics-crammed-into-an-E27-base-package models that usually fall over and die because their driver circuits are complete junk. People still buy them, because the alternative involves mucking around behind the wall with mains voltages.

With something like an LED fixture, especially if you want fancy color controls or dimming, or both, there really aren't any existing standards for sockets. The closest thing is probably gear designed for 12v halogen bulbs, which makes driving an LED array pretty painless; but that has no data/control channel. Power only.

If you had the luxury of doing a legacy-free design, top to bottom, things would definitely turn out much better; but unless you could do that and be able to get replacements from more than just a single vendor who may or may not go out of business and/or gouge you, you aren't likely to displace existing lousy but compatible solutions.

(Incidentally, this is probably why Wifi keeps popping up in home automation at all: it's brutally overpowered for the purpose, as well as relatively expensive, power hungry, and complex; but its sheer ubiquity and near-absence of vendor market power keep inspiring people to cram it into dubiously suitable places just because the alternatives are overpriced and proprietary, or only compatible with themselves, or both.)

Re:Wi-Fi toothpick

[tftp]

There also the significant downside in that all your light switches become useless because you have to leave them all on, all the time.

There is a more significant downside, but to learn about it you have to have a few of those smart switches in your house.

That downside is that they consume a lot of power, regardless of whether they are on or off. They don't have transformers; but that wouldn't be a good solution either because 1 to 2 watt transformers are horribly inefficient (I measured them personally.) A switcher would be efficient, but they cost too much, and they require medium frequency, ferrite-based transformers, and an optocoupler... it gets expensive fast.

So what are these switches using instead? They have a small capacitor, a small resistor, a rectifier bridge, and a Zener diode. That's basically all that they have. This is a horribly inefficient system; a typical dimmer switch with remote control (Z-Wave or Insteon, doesn't matter which one exactly) easily dissipates a watt of power. Multiply by 24 and 365.25, and you end up with about 9 kWh wasted just on that one switch. If you have ten of them, that'd be 90 kWh per year. Not much, but it's energy that you did not have to burn. To put it differently, each switch consumes enough energy to keep an incandescent 100W light bulb on for more than three days and three nights. If you use an LED or CFL light bulb, it takes 10% of that power, and then the switch itself burns enough power to keep that light on for a whole month.

I have these switches, and they are heating the walls of my house. This is particularly easy to notice in summer. I can afford that because all my power comes from the solar (PV) array. But it's a factor if you just want to "become green." The best way to be green is to use LED light bulbs and regular, dumb switches - and to use those switches whenever you enter and leave the area. Automation saves energy only in very specific circumstances, when you cannot expect the lights to be turned off promptly. Corridors of office buildings, at night, is one such example. At home most people neither want nor need home automation. They will not be better off with it. It's just a toy, and as most toys go it costs you more than you save.

Re:Wi-Fi toothpick

[AmiMoJo]

The problem is most western lighting is really, really shit. We usually have one or two hanging bulbs to light up a room, and then put shades on them which prevent light reflecting off the ceiling. Rooms are unevenly lit and we need expensive compact bulbs with the electronics crammed into a tiny space.

In the far east, particularly Japan, they have big (~0.6m) diffuse lights that put out 5500lm (about 5x what a 100W incandescent produces). They are LED based and only consume 50W of power. They last a long time as the electronics are properly spread out and cooled. You get a remote control that adjusts brightness and colour temperature. They don't even cost any more than typical gaudy western light fixtures.

Over there they also have more small low wattage lights that are used for specific areas where people work.

LEDs are already excellent, we just don't use them very well.

Re:Wi-Fi toothpick

[AmiMoJo]

The problem with putting it in the switch or the socket is that you then need an electrician to install it, or at least some knowledge of such things to do it yourself. With a bulb you just screw it in as normal, any consumer can do it. To the average person that is a big selling point.

Re:Wi-Fi toothpick

[thegarbz]

So what you're saying is when a TRIAC is always on and doesn't chop the sine wave then there's no distortion?

Are you a natural genius or did you just read my post?

Re: Wi-Fi toothpick

[thegarbz]

Reactive loads are typically filtered with with small LCR circuits as used in any universal dimmer capable of dimming CFLs and LEDs without flicker. They just aren't very common yet as they have only been on the market for a few years. Also the mA leaking through the TRIAC will not provide sufficient fwd voltage to make it to the caps. TRIAC control is used widely in smart lighting configurations in commercial buildings, yet I've never seen a light randomly flick on.

Re:Wi-Fi toothpick

[CastrTroy]

This is what I don't like about current LED designs. They really should come in a multipart design. There's no reason I should have to replace the AD/DC converter every time a few LEDs decide to break. Putting everything in a single form factor is convenient, but makes things much more expensive to fix when a single component fails. Since LEDs don't need to be in a vaccuum like incandescent filaments, and don't need harmful chemicals like CFL bulbs, they should be easily user servicable, allowing individual LEDs to be replaced, or allowing the working LEDs to be connected to a new DC converter (or whatever it's called) should that part die. If you want a computer controlled light, that part should be a separate pieced that connects to the rest of the bulb.

Re:Wi-Fi toothpick

[drinkypoo]

X10 is just worthless. It might have been a good idea once, but now there is so much noise that it never works.

On the other hand, Wifi is ridiculous overkill for this application. It quite literally makes no sense. Well, let me qualify that; one day it will make sense, when it costs much less to do, but today it literally makes no sense.

All of these problems could be solved trivially if you could just buy a power cable with some data wires bundled into it without spending a lot of money, then you could control all this stuff with wires (which could be brought together into a junction box next to the power junction) and use RS422 or what have you, like a pro. Of course, that'd only work for new installs, but it'd be better than all this fucking around with wireless.

Re:Wi-Fi toothpick

[drinkypoo]

LEDs are already excellent, we just don't use them very well.

Speak for yourself. I get plenty of light from my Cree lights. Both are in fixtures with good reflection and they cast a lot of light. In fact, I'm going to have to install a dimmer in the bathroom because it's too bright in there in the morning.

Over there they also have more small low wattage lights that are used for specific areas where people work.

Nothing prevents people from buying desk lights.

Re:Wi-Fi toothpick

[drinkypoo]

If you're smart enough to fix LED lights, you're smart enough to retrofit an LED light into your own enclosure using replaceable parts you buy from DX.com or similar. Instead of complaining, do something about it. Adding all that cost and complexity to something that doesn't fail if it's built worth a crap anyway (e.g. Cree-brand lights, which actually come with a respectable warranty) makes no sense while the prices are on a race to the bottom.

Re:Wi-Fi toothpick

[CastrTroy]

It really depends on a lot of factors. My lights are usually only on at night. In my neck of the woods, they charge less for power in the evening and nights. 6.7 cents per KWh. For 50,000 hours, a 60 watt bulb would cost $200. A 9 watt LED $30. So over the life time of the bulb, I would save $170. But then we subtract out the cost of the bulb. Let's assume it's only $30. So that's $140 saved over 50,000 hours of operation. Assuming I use it for 4 hours a night, that would be 34 years. I really don't trust that the light will still be working in 34 years. Something will break. I would be surprised if it lasted 5 years. So over 34 years, I'd have to replaced it 7 times. At a cost of $30 a bulb, that would cost me $210. Which is more than the amount the electricity costs me. I've left out the cost of the incandescent bulbs, but they cost almost nothing and I've had them last a couple years, sometimes longer than the CFL bulbs. I've still got a few in my house that I've been meaning to replace with CFLs since I moved into my house 3 years ago, but they haven't gone out yet. And those bulbs get turned on and off a couple times a day.

Re:Wi-Fi toothpick

[Monoman]

I swore off X10 simply because of their pop up ad assault many years ago.

Re:Wi-Fi toothpick

[drinkypoo]

No, really, the best LED lamps on the market (AFAICT) are the same price as the worst ones. Buy CREE LED lamps from the home despot, they are dimmable, have a ten year warranty and cost ten or thirteen bucks. They will most likely be on an end cap display someplace far away from the rest of the LED lamps. So far they are awesome. You can google around for a teardown to see that they have a much more credible-looking power supply than most LED lamps. No perceptible flicker, excellent color temperature. Very well heat-sinked. Barely longer than a normal lamp, but that is still a consideration.

hackers and have fun with this and maybe even driv

[Joe_Dragon]

hackers and have fun with this and maybe even driver people nuts / make them pay for all of this to go away.

This makes no sense.

[Chris Mattern]

Why would you put control circuitry that doesn't wear out into the replaceable part that *does* wear out instead of into the fixture that holds it?

Re:This makes no sense.

[Trepidity]

In principle it's not the right place for it, but it does make transitioning much easier. If you were going to put it in the fixture, you'd need to put out a whole range of devices with the new fixture, ranging from recessed ceiling lighting to lamps to whatever else people have in their houses. And once you did that, people would have to replace their existing fixtures with the new ones.

While if you put it in the bulb, you can just screw it in to any of your existing fixtures quite easily.

Re:This makes no sense.

[Belial6]

The same thing could be done with a screw in adapter. Adapter screws into the fixture, bulb screws into the adapter. Same feature set. More flexibility.

Re:This makes no sense.

[drinkypoo]

$10? That will outright pay for a 40W Cree equivalent with a ten year warranty. You're not getting them for the same cost, you're getting them for free (Well, three bucks more if you get the 60W)

OTOH I don't get any LED rebates, or if I do, they've not told me about them. Which is too bad...

Seems like a great way to...

[SemmiZamunda]

...absolutely POLLUTE the airwaves with junk wi-fi signals. Seems like this would add a ton of unnecessary interference on currently existing wireless networks.

Smarthome networked LED lightbulb

[Yahma]

While not WiFi, Smarthome has had a network connected LED bulb for over a year now. In my opinion, it is better suited for home automation than the WiFi bulb in the OP because it utilizes the Insteon Protocol, which is the replacement for X10.

Re:Smarthome networked LED lightbulb

[hjf]

Same with Philips' "Hue" lights.

Re:Smarthome networked LED lightbulb

[adolf]

More WiFi clients == less RF flying around?

No, not really: Insteon (and X10) are dead silent unless commands are being sent. Meanwhile, WiFi devices are inherently somewhat chatty; they all spend a significant portion of their time broadcasting "Hey, here I am! I'm still here! I'm still here! I'm still here! Hey, everyone! I'm still here! Are you there? Good! Because I'm still here!"

Re:Smarthome networked LED lightbulb

[Belial6]

There is no good reason that a light bulb that is designed to screw into a standard socket should ever use any kind of wireless technology for it's control. The thing, by it's very nature, is already connected to a wired network in the home. Using wireless pollutes the WiFi spectrum while simultaneously exposing the device to hackers.

Re:Smarthome networked LED lightbulb

[dfghjk]

Spoken like someone with no knowledge or experience with Insteon.

Curious how you think WiFi light bulbs would *reduce* the amount of RF "flying" around compared to Insteon (which would be powerline). Even in RF form Insteon is low power and low range in comparison...oh, and Insteon is "already here" too as are the networked Insteon bulbs. All you need to use them is a switch and a powerline bridge. No big deal.

Re:Troll level: 99

[girlintraining]

They actually have encrypted connection, so as long as the people can't actually get on your internal network (aka, your wifi is secure), you will not be able to do this. Maybe try to read something about the whole thing before judging. I know its hard but it makes everything a lot easier.

Anonymous Coward: Not having a sense of humor since 1997.

What I expect with WiFi bulbs

[manu0601]

There have been experiments of light bulbs as down link: the bulb adds HF data signal in light emissions, and mobile devices can use it, leaving traditional WiFi spectrum used for just up link. I thought this was what this story was about and I must confess I am a bit disappointed.

What a dumbass idea

[Tridus]

Yay, let's significantly increase the cost of making light bulbs (instead of simply making an attachment that screws into the socket and then takes a normal bulb), so we can increase the power requirements to run the light bulbs, so we can add yet more signals and interference to an already overcrowded wifi spectrum, so that we can make our light bulbs hackable... all in an effort to do what? Avoid having to flick a switch?

About the only thing they're not doing is wrong is suckering people out of money on kickstarter.

Lightbulbs aren't pricey enough as it is...

[HockeyPuck]

For a 60W (or equiv brightness) bulb at Homedepot...

Incandescent bulbs are dirt cheap at $.40 a bulb.
CFLs... at $2.25
LED is $13.

You now want to put wifi in this thing? It takes a long time to recoup the cost of a $13bulb... I can't imagine what it would take to recoup some $25 wifi enabled bulb with encryption.

Wouldn't it also be the ultimate power vampire? You'd now be putting your lightbulbs into standby if you wanted to turn them on and off via some smartphone app. Last I checked when I turned them off via the wall switch, they actually went off.

Re:just what the world needs

[Nutria]

Why bother to have switches at all when you can have lights controlled by your smart phone?

Not everyone has a smart phone.

Not everyone carries it around every moment of the night and day.

You can flip a light switch with your elbow when your arms are full.

The many 10s of millions of people with presbyopia and myopia don't need glasses to flip light switches.

Requiring everyone in a family -- from the very young to the very old to carry a smart phone, and to pay for all those contracts, is plain, fucking stupid.

Dumb

[Animats]

It lights up. It can be turned on and off and dimmed remotely That's where we were with X10 in the 1980s. It doesn't relay data around for other WiFi devices.

It has over-the-air firmware updates. Your smartphone doesn't really talk to the lamps. It talks to their "cloud server", to which the lamps phone home. What could possibly go wrong?

Defeats energy savings, reduced lifespan

[Khyber]

Great, so now you have a wifi transmitter and computer inside of an already-heated LED light. Which means as the whole enclosure gets warmer from the added electronics, LED efficiency drops, as well as the overall lifespan. To boot, these LED bulbs aren't even as efficient as cheap Chinese LEDs that can run from a single network-connected smart dimmer TO BEGIN WITH, so their efficiency becomes roughly that of CFL, totally negating any power savings benefit unless you run the LED bulb at less-than-optimal power.

This is about as pointless as the 'fluid source' LED on kickstarter, which has a pathetic 50 or so lumens per watt - not even as good as a CFL.

To all you LED people on Kickstarter - hire some real engineers that know what's going on; your guys don't have a clue and you're making our industry look bad.

God I feel for you clowns way back in 2013.

[Impy the Impiuos Imp]

I predict in a few years, people will think back to the bad old days, you know, when you had to get up to flip the light switch on and off.

Much the same way we used to have to take a pillow off the sofa and lie in front of the TV in order to change the channel and surf during commercials. Seriously.