Felix Salmon's screed is poorly informed about LED bulbs and the current state of the art in wifi.
First let's go over the current marketplace:
• There are very nice 60 watt equivalent bulbs available for <$25 in quantity one.[1] Without looking directly at the bulb you will not tell them from incandescent. Single color, using phosphors to shift blue LEDs down to nicer frequencies.
• At quantities >500 LED bulbs which look like the illustration on the kickstarter page are under $10. [2] This gives a rough estimate of cost of materials assembled.
• A thumbnail sized microcontroller module with wifi is <$30 in quantity one.[3]
It certainly looks like there is room for a $50 that includes a light and a wifi enabled microcontroller. The cost model gets easier though. Each customer only needs one master bulb with the wifi, the rest only have the cheaper 802.15.4 mesh radios.
Before we look at the article anymore, lets consider what is different about the Lifx than the $10 bulb above.
• More efficient. It needs to be a bit brighter to meet the Lifx specs. This adds cost to the LEDs to run more current within the thermal limits.
• Tri color. It is going to need at least three LED colors. Ideally the sum of them is going to be a pleasing white so you get a usable color at full brightness. The other colors are achieved by running some of the LEDs at lower duty cycle. Notice they chose a bulb with a diffuser. That keeps you from getting blotchy spot light effects.
• More power switching FETs. Their total power handling will be the same, but it will be a higher component count.
• A microcontroller and radio added to the circuit board.
• Software, a couple weeks for the bulb innards.
For that they get to roughly quintuple the sale price. Sounds doable to me.
Most of the article is just an "On no! The sky is falling!" piece on Kickstarter feeling like a store. He uses some other company that took a run at the "just like incandescent" market and apparently got beaten by Philips. For some reason, he has a side-by-side set of photos which the Lifx illustration bulb and possibly the ugliest LED bulb I've ever seen. (Said ugly bulb is proclaimed indistinguishable from incandescent when inside a lamp shade, so it has a use case. Just don't let anyone see it naked.)
As the article progresses it just gets embarrassing. He declares it impossible because other smart people haven't done it and confidently predicts every negative lightbulb attribute he can think of will afflict the bulb. He throws out some ridiculous straw men for the product and imagines a huge barrier for configuration of a UI free wifi device (hint: solved long ago by many products, e.g. Apple's Airport Express) and that somehow the 0.05 watts from the wifi will overheat the bulb which is dissipating ~10 watts already.
At least it was run as "Opinion" instead of "Fact".
Here's my opinion on Lifx: The masterbulb concept is a mistake. Too confusing. Have a controller that plugs in the wall. Keeps the confusion down but also keeps people from switching off the master bulb or having to have a battery in it. They may need to give a bit on the total brightness to keep the heat under control. If they try to grow to something large, they will probably be killed by a patent lawsuit. If they make it far enough to make a nice PAR20 version (diffusion is much harder), I will happily buy 30 of them.
I think this is an excellent post. But perhaps not for the reasons jws thinks it is :-).
Look at that Phillips bulb, isn't it neat? When they won the L-prize for a bulb that could do 60W equivalent for 10W input, some of the coverage said that they had spent just over $28M developing that one bulb. They didn't open source anything they discovered in that process about making LED light bulbs. And granted LEDs have improved in the last 3 years since the Department of Energy ran its $10M 'kickstarter' campaign, the bulb linked is a 12W for 60W equivalent, and its only one color.
So one has to wonder, if it seems easy but the only one company was even able to complete and entry for the L-prize, what did they know that I don't?
I wonder if it is a form of suspended disbelief, if its far enough away from your direct expertise but seems practical and one really loves the concept, one can invest a level of belief into someone saying they can do what you want.
The one that will be interesting will be a Kickstarter to bring an herbal supplement to market that stops aging. I'm sure its coming, or a perpetual motion machine, maybe a project to get the Rossi cold-fusion machine into production. They keep saying if they have a million dollars they can build a functioning power plant.
First let's go over the current marketplace:
• There are very nice 60 watt equivalent bulbs available for <$25 in quantity one.[1] Without looking directly at the bulb you will not tell them from incandescent. Single color, using phosphors to shift blue LEDs down to nicer frequencies.
• At quantities >500 LED bulbs which look like the illustration on the kickstarter page are under $10. [2] This gives a rough estimate of cost of materials assembled.
• A thumbnail sized microcontroller module with wifi is <$30 in quantity one.[3]
It certainly looks like there is room for a $50 that includes a light and a wifi enabled microcontroller. The cost model gets easier though. Each customer only needs one master bulb with the wifi, the rest only have the cheaper 802.15.4 mesh radios.
Before we look at the article anymore, lets consider what is different about the Lifx than the $10 bulb above.
• More efficient. It needs to be a bit brighter to meet the Lifx specs. This adds cost to the LEDs to run more current within the thermal limits.
• Tri color. It is going to need at least three LED colors. Ideally the sum of them is going to be a pleasing white so you get a usable color at full brightness. The other colors are achieved by running some of the LEDs at lower duty cycle. Notice they chose a bulb with a diffuser. That keeps you from getting blotchy spot light effects.
• More power switching FETs. Their total power handling will be the same, but it will be a higher component count.
• A microcontroller and radio added to the circuit board.
• Software, a couple weeks for the bulb innards.
For that they get to roughly quintuple the sale price. Sounds doable to me.
Most of the article is just an "On no! The sky is falling!" piece on Kickstarter feeling like a store. He uses some other company that took a run at the "just like incandescent" market and apparently got beaten by Philips. For some reason, he has a side-by-side set of photos which the Lifx illustration bulb and possibly the ugliest LED bulb I've ever seen. (Said ugly bulb is proclaimed indistinguishable from incandescent when inside a lamp shade, so it has a use case. Just don't let anyone see it naked.)
As the article progresses it just gets embarrassing. He declares it impossible because other smart people haven't done it and confidently predicts every negative lightbulb attribute he can think of will afflict the bulb. He throws out some ridiculous straw men for the product and imagines a huge barrier for configuration of a UI free wifi device (hint: solved long ago by many products, e.g. Apple's Airport Express) and that somehow the 0.05 watts from the wifi will overheat the bulb which is dissipating ~10 watts already.
At least it was run as "Opinion" instead of "Fact".
Here's my opinion on Lifx: The masterbulb concept is a mistake. Too confusing. Have a controller that plugs in the wall. Keeps the confusion down but also keeps people from switching off the master bulb or having to have a battery in it. They may need to give a bit on the total brightness to keep the heat under control. If they try to grow to something large, they will probably be killed by a patent lawsuit. If they make it far enough to make a nice PAR20 version (diffusion is much harder), I will happily buy 30 of them.
EOM
[1] http://www.homedepot.com/Electrical-Light-Bulbs-LED-Light-Bu...
[2] http://www.alibaba.com/product-gs/601830276/led_light_bulb_e... For ballpark reference only. Never trust the specs.
[3] https://www.sparkfun.com/products/11395