visible light communication

Thread Starter

harishraj

Joined Aug 2, 2013
1
hey
i am making a project on visible light communication based on ofdm..any idea about the receiver circuit design..i need to filter ambient light..what photodiode shold i use..i am using white leds as transmitter ..and pin diode for receiver section
 
Last edited by a moderator:

crutschow

Joined Mar 14, 2008
24,977
hey yourself

Don't see how you can optically filter ambient light if you are using white LEDs. :confused: Why are you using white light? A single color would be much easier to filter.

One thing you might do is make a very high dynamic range receiver input stage and block the DC ambient light signal (AC couple) to the following stages so that only the AC signal is amplified.
 

cornishlad

Joined Jul 31, 2013
218
As these systems are invariably very directional you could put your photo transistor in a tube with the inside painted black to reduce the effect of ambient light. Longer the better !
As white light is made up of all colours so I don't see colour filtering doing much good.
 

Shagas

Joined May 13, 2013
804
Hmm even if doing something like communication in visible light (which I don't know why anyone would do that over using IR ) then wont the ambient light not matter since the information passed is changing levels all the time and ambient is more or less constant .
I'm thinking about this analogically to using an ac coupling capacitor to get rid of the DC voltage and only couple the ac . Can we do this the same way with light?
 

wayneh

Joined Sep 9, 2010
16,390
... the information passed is changing levels all the time and ambient is more or less constant .
Ambient is noisier than you might think. Especially indoors if there are fluorescent lights around.

Aren't IR LEDs generally much higher power than visible light, like 100ma instead of 10mA? That puts you at 10X disadvantage right there.
 

joeyd999

Joined Jun 6, 2011
4,394
Hmm even if doing something like communication in visible light (which I don't know why anyone would do that over using IR ) then wont the ambient light not matter since the information passed is changing levels all the time and ambient is more or less constant .
I'm thinking about this analogically to using an ac coupling capacitor to get rid of the DC voltage and only couple the ac . Can we do this the same way with light?
Don't try this under florescent or other gas discharge lamps (i.e. sodium). The lighting is definitely not DC!

And much of the new LED lighting uses PWM to control brightness.
 

JohnInTX

Joined Jun 26, 2012
4,083
Hmm even if doing something like communication in visible light (which I don't know why anyone would do that over using IR ) then wont the ambient light not matter since the information passed is changing levels all the time and ambient is more or less constant .
I'm thinking about this analogically to using an ac coupling capacitor to get rid of the DC voltage and only couple the ac . Can we do this the same way with light?
The problem is the sensor itself. Its output is proportional to the incident light (in your case a very broad spectrum). When the sensor is exposed to enough light, steady or not, it will saturate. End of comms. Your idea of AC coupling is valid to discriminate data from background light but only if the background light is not saturating the sensor.

IR sensors not only have circuitry to separate background light from data by frequency discrimination, they also usually have optical filtering to filter out all but light in a very narrow band of spectrum that is unlikely to be present in the anticipated environment.

EDIT: wayneh and joeyd999 make some excellent observations.
 

joeyd999

Joined Jun 6, 2011
4,394
Ambient is noisier than you might think. Especially indoors if there are fluorescent lights around.

Aren't IR LEDs generally much higher power than visible light, like 100ma instead of 10mA? That puts you at 10X disadvantage right there.
Believe it or not, you can reliably blast up to 1A or more through a low power IR emitter, as long as you keep the pulses short and the average power dissipation less than the max stated value. This can give you a 100:1 improvement in light output. Add some AM or PFM to the driving signal, and a good detector, and you've got a nice IR communications setup.
 

LDC3

Joined Apr 27, 2013
924
IR sensors not only have circuitry to separate background light from data by frequency discrimination, they also usually have optical filtering to filter out all but light in a very narrow band of spectrum that is unlikely to be present in the anticipated environment.
Just because you can't see it doesn't mean it is not there. There is a lot more IR light present than you think. Incandescent lights put out a lot of IR. On the other hand, UV light is not as common, except with the sun.
 

GopherT

Joined Nov 23, 2012
8,012
hey
i am making a project on visible light communication based on ofdm..any idea about the receiver circuit design..i need to filter ambient light..what photodiode shold i use..i am using white leds as transmitter ..and pin diode for receiver section
harishraj,
If you ever decide to return to this thread, could you please explain some of your reasons why visible? Will this be transmitted through free air or on fiber optic cable (what distance)? Will you have access to visible light optical filters for your OFDM - I would assume it is required. What type of transmission rate are you hoping to achieve? How many multiplexed signals will you have?
 

Shagas

Joined May 13, 2013
804
True that most ambient lighting is not Dc BUT if our data transmittion is at much higher frequencies can't we filter out the low ac lighting frequencies?

Yes led lighting is controlled with PWM but I haven't seen ambient LED lighting yet.... only in special places.
 

joeyd999

Joined Jun 6, 2011
4,394
True that most ambient lighting is not Dc BUT if our data transmittion is at much higher frequencies can't we filter out the low ac lighting frequencies?
Yes, though one problem is harmonics. Fluorescents have a fundamental frequency of 120 Hz (100 Hz across the pond), but it's not even close to a sine wave. Closer to full wave rectified AC -- but that is not even true because light output is non-linear with respect to voltage, IIRC. So, there are harmonics extending far off in multiples of 120Hz, plus other frequency components due to noise sources in the AC lines.

Also, at reasonable distances, ambient lighting is far brighter than the received light signal. So the detector must be designed so that it does not saturate under worst case lighting conditions. It follows that the resultant signal will be quite small (many 10s of dBs lower) compared to the noise.

In the absence of filtering the actual wavelengths in question, some modulation technique must be used to get reasonable range and S/N. And the detector's band-pass filter(s) need to be reasonably tight. A lock-in amplifier works well in these kind of applications.

I have no idea how OFDM (as proposed by the OP) performs vs. external noise sources, BTW.

EDIT: AFAIK, OFDM's prime advantage is that it prevents neighboring channels from interfering with each other, on an otherwise clean wire (or fiber). It may not perform well in noisy environments, though.
 
Last edited:

THE_RB

Joined Feb 11, 2008
5,438
Most types of high frequency AM or FM modulation will work pretty well, provided the frequency is higher than (or very different to) the frequencies of the light noise from mains Hz or LED lamp PWM Hz.
 

balckbox

Joined Jan 6, 2014
1
Hi
Harish I am doing my final year project on VLC. And i am using Ethernet for data transfer. I need your help regarding this.
<SNIP>
 
Last edited by a moderator:

alfacliff

Joined Dec 13, 2013
2,458
one way if your just experimenting is to use the same type led for a detector as you use for the transmitter. led's generate dc voltage when exposed to light and the most when exposed to the same color as they are designed to put out. for a quick demo, put an infra red led on your scope probe, and point it at a tv remote, you should see the pulses the remote is sending.
 
Top