Feeding data to an LED

[chattr]

I'd like to create an one-way optical data link between two computers. I could use two fiber optic media converters for that, but they are rather expensive and in the end I want a single small device. I took a look at IRDA and VLC, but my main problem is how to get the data out of the computer and convert it into light. Imho, the easiest would be ethernet and 10Base T with its Manchester encoding. But I need something faster like 100 mbps. 100Base TX ethernet uses MLT 3 encoding and I guess I can't drive a LED with that because it uses three voltage levels (e.g. -1, 0, +1) and a LED can only be driven by two (am I right?).
I looked for ICs that can do interconversion between the different encoding techniques but up to now I couldn't find anything.
Is there any way I can make a LED understand MLT 3 encoding?

 

[Ian Rogers]

I did something similar for my project way back when.... I took three analogue signals and multiplexed them onto a 40Khz carrier and sent it to a PC running a program I wrote in borland C...

I did use infrared LED's and photo transistors on the PC side...... I could get quite a decent distance...
If you want to do it with SIRC take a look at Nigel's PIC tutorials and you will be able to do same....

 

[AnalogKid]

An LED is a linear device. That is, over a certain range of current values it makes a range of brightness. So amplitude modulating it with something other than on/off is entirely possible. But the receiver will have to have a really fast AGC loop to present stable voltage levels to the comparators or whatever you use to translate the 3 light levels back into voltages. Also, like all AM radios, the signal will be susceptible to amplitude noise.

ak

 

[chattr]

I did use infrared LED's and photo transistors on the PC side...... I could get quite a decent distance...
If you want to do it with SIRC take a look at Nigel's PIC tutorials and you will be able to do same....

A high distance does not matter in my case. I only need a range of a few centimeters. The PIC IR tutorial looks interesting. Do you know which IC was used in this case? I can't figure it out from the small image. I'm also unsure if I can directly use the MLT 3 encoding for transmission.

An LED is a linear device. That is, over a certain range of current values it makes a range of brightness. So amplitude modulating it with something other than on/off is entirely possible.

I was also thinking about this but wouldn't I need voltage levels of e.g. 0, +1 and +2 V? The MLT 3 encoding also uses negative voltages but afaik applying these to the LED would destroy it.
So either I need something which on the LED side adds +1 V and subtracts 1V on the receiver side or I need some kind of LED which also lights up under a reverse voltage. I don't know if either of those is even possible or how to accomplish that.

But the receiver will have to have a really fast AGC loop to present stable voltage levels to the comparators or whatever you use to translate the 3 light levels back into voltages.

Do you know any receiver that can do that? I was thinking of photodiodes which are also used in fiber optic links. They should be fast enough but I don't know if they can handle the three levels. These links use either Manchester or NRZI encoding which only have two levels.

 

[Ian Rogers]

Do you know which IC was used in this case?

pic16f628..

 

[Tony Stewart]

IR LED's do not have enough power and bandwidth to support 1mbps beyond 1 meter. SO you will have to invest in a laser transmitter.
Is that what you can afford?
MLT-3 encoder / decoder is easy.
I have used IRDA2 up to a few meters at 100kbps but beyond that, it's easy to use with binary , as the decoder has AGC, but marginal with path loss unless you get a high power IR LED <10 degree and pulse it at 1 Amp for <<1us at 100kbps.


IRDA2 was design for wireless portables from phones and cameras in Mbps but very short distances << 1m

SO what is your realistic expectation for speed and distance?

 

[Mikebits]

Agreed, a standard IR LED would never handle 100 mbps.

 

[Tony Stewart]

Agreed, a standard IR LED would never handle 100 mbps.

and would be lucky to get 1mbps in 1 m.

 

[chattr]

IR LED's do not have enough power and bandwidth to support 1mbps beyond 1 meter. SO you will have to invest in a laser transmitter.
Is that what you can afford?

Afaik common sfp transceivers or media converters use IR LED's. They can handle a speed of 100 mbps and are rather cheap. Of course this is not the same kind of IR LED you use in a TV remote control. The Ronja project uses IR LED's for 10 mbps data links over several hundred meters: http://ronja.twibright.com/

MLT-3 encoder / decoder is easy.

So how can this be done? All I could find up to now are encoders/decoders which are integrated in the ethernet chip.

I have used IRDA2 up to a few meters at 100kbps but beyond that, it's easy to use with binary , as the decoder has AGC, but marginal with path loss unless you get a high power IR LED <10 degree and pulse it at 1 Amp for <<1us at 100kbps.

I found some cheap IR transceivers from Vishay which support a couple of mbps, though, still well below 100 mbps. There is also UFIR and GigaIR which support bandwiths of 100 or 1000 mbps. But I could not find any commercial device supporting these specifications.

SO what is your realistic expectation for speed and distance?

100 mbps on a distance of 1 m (but more likely only some centimeters). I know that this is possible because fiber optic media converters/ sfp transceivers easily support this. There are also lots of publications on visible light communication or free space optics which achieve these data rates. But except media converters or sfp transceivers they don't have ethernet as an input. In the end I would need something like a fiber optic media converter with two ethernet connections.
The dataflow would be: pc1 --> ethernet1 --> led --> photodiode --> ethernet2 --> pc2

 

[Tony Stewart]

Afaik common sfp transceivers or media converters use IR LED's. They can handle a speed of 100 mbps and are rather cheap. Of course this is not the same kind of IR LED you use in a TV remote control. The Ronja project uses IR LED's for 10 mbps data links over several hundred meters: http://ronja.twibright.com/

So how can this be done? All I could find up to now are encoders/decoders which are integrated in the ethernet chip.

I found some cheap IR transceivers from Vishay which support a couple of mbps, though, still well below 100 mbps. There is also UFIR and GigaIR which support bandwiths of 100 or 1000 mbps. But I could not find any commercial device supporting these specifications.

100 mbps on a distance of 1 m (but more likely only some centimeters). I know that this is possible because fiber optic media converters/ sfp transceivers easily support this. There are also lots of publications on visible light communication or free space optics which achieve these data rates. But except media converters or sfp transceivers they don't have ethernet as an input. In the end I would need something like a fiber optic media converter with two ethernet connections.
The dataflow would be: pc1 --> ethernet1 --> led --> photodiode --> ethernet2 --> pc2

YES but these were not ordinary IR LEDs and ordinary Transmitters
●LED + 13cm lens, 17mW of light
●10MBaud, 100% depth
● 14' divergence , 4m spot @ 1km

After you build your optical link and determine the fastest transition frequency it can support, I can design the encoder /decoder. I designed a 10mbps fiber link before 30 yrs ago, so I should be able to do it for fiberless or "Visible Light Communication" (VLC) as it is called now.

The usual problems of dispersion, group delay distortion, eye pattern asymmetry must be addressed with rising capacitance with emitter power but that isn't hard to me.
**broken link removed**