by Bryan Bergeron, Editor by Bryan Bergeron, Editor
The US energy efficiency legislation for light bulbs has been a boon for alternatives to the traditional
incandescent bulb, including the LED. Energy issues aside,
I use LED lamps for my desk and workbench because
they’re small, unobtrusive, and give off little heat.
Still, there’s a lot of energy being wasted by LED-based lighting — especially if you think of the light they
generate as unused bandwidth. Making use of this
bandwidth is the basis of LI-FI (Light Fidelity) which has
many parallels with WI-FI.
Proponents of LI-FI contend that room light,
automobile lights, interior aircraft lights, and street lights all
represent under-utilized bandwidth that can carry
everything from voice communications and HDTV to
intelligent navigation signaling between cars. If you want a
great vision of what’s possible with the technology, check
out the free TEDtalks podcast, “Wireless Data From Every
Lightbulb,” available on i Tunes.
Communications with light is nothing new. Roman
soldiers used polished shields to reflect sunlight when
signaling during battle. Commercial, secure, line-of-sight
laser communications systems have been around for
decades. I’ve used a DIY He-Ne gas laser and telescope
system to communicate over five miles, but at modest
bandwidth. And, of course, much of the phone system is
based on optical transmission and switching.
IR light signaling has been around for some time. You
probably own at least one IR remote. However, as with
UV light, you wouldn’t want to bathe a room with IR light
because it can damage your vision. The real R&D activity
is around visible light. There’s plenty of bandwidth
available, and it’s not likely that you’ll inadvertently stare
into a bright light and burn your retinas.
As with RF communications, the transmission process
is straightforward and doesn’t take much in terms of
hardware. Reception is the sticking point. There are issues
of inadequate signal strength, how to handle multipath
signals formed by light reflecting off of different objects
and arriving at the receiver out of time, and interference
from other light sources, just to name a few.
So, assuming I’ve piqued your interest, where do you
begin? Start simple. Get an ordinary red LED and
phototransistor — each coupled to your favorite
microcontroller — to communicate with each other. Once
you’ve managed a simple simplex serial connection, go for
full duplex (two-way) communications.
To give you an example of what’s involved, I’ve been
8 December 2011
experimenting with parallel data streams using red, green,
and blue LEDs. Driving low power red, green, and blue
LEDs with an Arduino is trivial. I use external switching
transistors with each LED so that I can handle high power
On the receiver end, I’m working with two different
approaches. The first uses separate phototransistors for
red, green, and blue light. I use standard photographic
acetate as a filter for each phototransistor. Red, green, and
blue transparent film from acetate report covers works just
about as well.
I’m also working with color-sensitive light sensors,
including the TCS3200-DB color sensor from Parallax. At
almost $60, it’s expensive, but it has a built-in array of
photodetectors with red, green, and blue filters. It’s worth
looking at the spec sheet (downloadable from
Parallax.com) to get an idea of how they’re using the
chip. Parallax also sells the TSL230R light-to-frequency
converter. As with the TCS3200-DB, the $6 chip isn’t
designed for color light communications, but it has
potential that’s worth exploring.
SparkFun electronics sells an inexpensive TEMT600
($1.50) light sensor that can be put behind a color acetate
filter. As with standard phototransistors, these devices are
more sensitive to certain portions of the visible spectrum
than others. SparkFun also sells the Avago ADJD-S311-
CR999 that has built-in red-green-blue filters ($5).
Unfortunately, you’ll have to be skilled at SMT mounting
to get at the input and output leads.
Of course, the real work for harnessing broadband
light communications is in the software. Checksums and
other error-detection mechanisms provision for
interference from other visible light transmitters. There’s
the approach used by most IR remote controllers — that of
modulating the bean at about 38 kHz. Take a look at the
spec sheet on the IR receiver TSOP85 — available from
SparkFun — to get an idea of what’s typically involved in
Of course, if you’re on an unlimited budget, then you
could start considering diffraction gratings, prisms, and
other commercial-grade tools developed for the fiber
optics community. Edmund Scientifics is a good place to
look for information and products.
There are obvious technical and behavioral issues that
must be addressed before LI-FI becomes ubiquitous. For
example, there will be no more tucking your phone in
your pocket or purse. You’ll have to expose at least part of
the phone to light — perhaps as Bluetooth chest-pin
communicators akin to those on Star Trek. Then, there are