Laser communication

[SScattered]

Hi,
Recently I've built a system where I can send data using a laser. This system uses 2 ESPs as the sender and receiver. It's short range and I only tested around 1m gap between laser and receiver. I thought of giving another shot for this project using FPGAs. I bought Tang Nano 9K FPGA but I'm new to FPGAs. I learned some lessons while I was in the uni, but there was no use of FPGAs for me until now.

My idea is to implement manchester encoding and hamming(7,4). So, I thought FPGA would give me more data rate. Current data rate is around 74kbit/s, it takes a considerable time to transfer even a 2MB file. Can anybody tell me where should I start and will there be improvement of data rate using FPGAs?

Also note that Tang Nano 9K has a SD Card slot also, maybe I can use that instead of using a microcontroller with FPGA to store data there?

PS: I don't use fiber optic. This is going to be short range free-space optical communication
PS: I'm an electrical & electronic engineer

[tggzzz]

Whatever you do, make very sure that it cannot blind people passing by, even when there is a gross failure in the system. For the avoidance of doubt, I don't care if you blind yourself.

What happens if the laser is turned on permanently? Or the current through the laser is too high? Or the optics are knocked/damaged?

Make sure you understand the eyesafety limits; many lasers on the market exceed those by a factor of 100 or 1000

[jonpaul]

simpler with fiber optic. Toshible TOS-LINK etc,

j

[moffy]

As jonpaul mentions, fiber optics is easier because one can buy integrated packages with all of the optics and drivers built in. If you want to go free air then you will probably need to look at collimating lenses and your own driver and receiver. Optical communication starts with a fast LED/Laser and driver and a fast low noise receiver, possibly using a PIN diode. The optics and filtering are as important as the transmitter and receiver. Once you have the physical link established then you can look at the data transmission and reception protocols. So is 1mm the physical link distance you are looking for or do you plan to use it in some other application? And as tggzzz mentions eye safety is a critical factor and if you can stay for lasers at or below 1mw you are better off.

[coppercone2]

I have a feeling you don't need power unless you need to get through rain.

Good optics reduce the power requirement drastically.


It made me wonder if an array of low power lasers might work in rain, since maybe one of the beams will have better integrity for a short amount of time with a random distribution of rain droplets. Fog might be the bigger problem.


What you need for safe experiments on lasers in open environment
1) heavy support structure. Not some dinky tripod with ball joints that barely fit, I mean that you need to tighten down a big bolt
2) mechanically seated components (i.e. glue joint failing won't let a optic get out of the safe zone)
3) operating area is smaller then safe zone so you can account for things like optics becoming dislodged etc'


what is a bad idea is doing this in a metropolitan area


Alot of optics are just glued ,which is obviously a problem because super glue or epoxy can fail.


For making high power lasers safe, IDK. I think you just need everyone that can be effected to wear goggles. Even if you have a rugged hardware setup, its still possible for glass to fracture and allow a beam to go into basically a random direction. Long shrouds over the output of the laser might help restrict it, but really that IMO needs to stay in a physics lab.

Making safe links for consumers at high power levels is probobly never gonna be a reality because joe dumbarse is going to want to know if the thing is down and look down the optic output to see if the 'bulb is weak or somtin'

[tggzzz]

For making high power lasers safe, IDK. I think you just need everyone that can be effected to wear goggles.

Correct. Then define everyone that could be affected, and how you will ensure they wear goggles.

Good luck with that!

Even if you have a rugged hardware setup, its still possible for glass to fracture and allow a beam to go into basically a random direction.

Never forget specular reflections; they make beams go round corners, through open doors etc.

Long shrouds over the output of the laser might help restrict it, but really that IMO needs to stay in a physics lab.

Making safe links for consumers at high power levels is probobly never gonna be a reality because joe dumbarse is going to want to know if the thing is down and look down the optic output to see if the 'bulb is weak or somtin'

Yup.

[coppercone2]

could a large grid of low power lasers help deal with rain related losses if some algorithm was applied to maintain safety?

[NiHaoMike]

I recall there was a project a long time ago to run 10Mbps Ethernet over laser pointers.

[daisizhou]

Can you share some details?Although the fiber optic connection is good, it is a wired connection, which is very inconvenient.
I'm very interested in your project

[tggzzz]

I recall there was a project a long time ago to run 10Mbps Ethernet over laser pointers.

TAoE section 9x.23 is "Sending power on a beam of light".

They note the product https://www.lumentum.com/en/products/diode-laser-830-nm-fiber-coupled-2w-2486-l4-series

The datasheet contains this. Note all the words

[SScattered]

Whatever you do, make very sure that it cannot blind people passing by, even when there is a gross failure in the system. For the avoidance of doubt, I don't care if you blind yourself.

What happens if the laser is turned on permanently? Or the current through the laser is too high? Or the optics are knocked/damaged?

Make sure you understand the eyesafety limits; many lasers on the market exceed those by a factor of 100 or 1000

Hi, thanks for the reply! Yeah, for sure. Currently I'm using a 5mw red-laser diode. My current testing distance will be around 1m so when I increase the distance I sure will need the lenses.
So I think with this laser and the distance it's not much of an issue to do testings?

[SScattered]

As jonpaul mentions, fiber optics is easier because one can buy integrated packages with all of the optics and drivers built in. If you want to go free air then you will probably need to look at collimating lenses and your own driver and receiver. Optical communication starts with a fast LED/Laser and driver and a fast low noise receiver, possibly using a PIN diode. The optics and filtering are as important as the transmitter and receiver. Once you have the physical link established then you can look at the data transmission and reception protocols. So is 1mm the physical link distance you are looking for or do you plan to use it in some other application? And as tggzzz mentions eye safety is a critical factor and if you can stay for lasers at or below 1mw you are better off.

Yes, it will be short range free-space optical communication. The laser diode I have is 5mW (mentioned in the website tho)

[SScattered]

I have a feeling you don't need power unless you need to get through rain.

Good optics reduce the power requirement drastically.


It made me wonder if an array of low power lasers might work in rain, since maybe one of the beams will have better integrity for a short amount of time with a random distribution of rain droplets. Fog might be the bigger problem.


What you need for safe experiments on lasers in open environment
1) heavy support structure. Not some dinky tripod with ball joints that barely fit, I mean that you need to tighten down a big bolt
2) mechanically seated components (i.e. glue joint failing won't let a optic get out of the safe zone)
3) operating area is smaller then safe zone so you can account for things like optics becoming dislodged etc'


what is a bad idea is doing this in a metropolitan area


Alot of optics are just glued ,which is obviously a problem because super glue or epoxy can fail.


For making high power lasers safe, IDK. I think you just need everyone that can be effected to wear goggles. Even if you have a rugged hardware setup, its still possible for glass to fracture and allow a beam to go into basically a random direction. Long shrouds over the output of the laser might help restrict it, but really that IMO needs to stay in a physics lab.

Making safe links for consumers at high power levels is probobly never gonna be a reality because joe dumbarse is going to want to know if the thing is down and look down the optic output to see if the 'bulb is weak or somtin'

The laser diode is 5mW red laser. This will be a short range free-space optical communication. My first concern is to get this working between couple of meters. I want to achieve higher bit rates, that's why I thought I'd use FPGAs.

Thanks! Yeah, I understand the very dangers when it comes to lasers. One mistake with a high powered laser, you are blind next second. But still I've no intention of going for high power laser. I will stick to 5mW diode and see where it can take me. Even if it's couple of meters I'd be happy to have higher bit rates 

[SScattered]

Can you share some details?Although the fiber optic connection is good, it is a wired connection, which is very inconvenient.
I'm very interested in your project

I've already mentioned it in the first post. Anyway, I'll try to explain it again.

I currently have a system that I've already built. I used 2 ESP32s as sender and receiver. So the with the current system I send data from Micro SD to receiver which writes data to Micro SD on the receiver.
I want to do manchester encoding and hamming coding so I thought use of FPGA for this task would be quite good and give me better results and speed.

Please see the attached image. The cover and everything is designed and 3D printed.

[tggzzz]

So I think with this laser and the distance it's not much of an issue to do testings?

What I (and you) think is completely and utterly unimportant and irrelevant. What matters is what actually happens.

You need to be sure that it cannot output dangerous levels under any fault conditions, e.g. too high a current, or an inaccurate specification.

Feel free to test it on your eyeballs. Do not test it on other people's eyeballs.

I am very deaf. If I go blind for any reason whatsoever, then I will commit suicide. Seriously, no joke. Hence I'm sensitive to such "hopeful guesses".

[unseenninja]

I currently have a system that I've already built. I used 2 ESP32s as sender and receiver. So the with the current system I send data from Micro SD to receiver which writes data to Micro SD on the receiver.
I want to do manchester encoding and hamming coding so I thought use of FPGA for this task would be quite good and give me better results and speed.

The most important factor for the bandwidth you can achieve will be more down to how you are sending the data, not necessarily how many layers of redundancy and error correction you add.

How are you currently sending the data? If you are just using simple On Off Keying (OOK), then that will be your fundamental limitation. The ESP32 modules don't have very stable or accurate clocks.
 

[moffy]

As jonpaul mentions, fiber optics is easier because one can buy integrated packages with all of the optics and drivers built in. If you want to go free air then you will probably need to look at collimating lenses and your own driver and receiver. Optical communication starts with a fast LED/Laser and driver and a fast low noise receiver, possibly using a PIN diode. The optics and filtering are as important as the transmitter and receiver. Once you have the physical link established then you can look at the data transmission and reception protocols. So is 1mm the physical link distance you are looking for or do you plan to use it in some other application? And as tggzzz mentions eye safety is a critical factor and if you can stay for lasers at or below 1mw you are better off.

Yes, it will be short range free-space optical communication. The laser diode I have is 5mW (mentioned in the website tho)

I think you should start with a collimating lense, as the optical power coupled to the receiver without collimation will drop off very quickly. You should also practice the alignment process between transmitter and receiver, it gets trickier with distance.

[madires]

A 5mW laser would be laser class 3R which requires some safety measures. With appropriate optics (diverging, spreading) it could become class 1M (fairly safe). Commercial FSO systems are usually class 1M.

[cncjerry]

5mw is enough to cause serious retinal damage from a reflection, especially if you are aiming at a metal canned detector.   You dont need a laser to get distance, you just need an LED, think remote control,  but it has to be modulated for detection improvement.  Switching it on and off is not enough, you have to modulate the signal, think lock-in amplifier.  That will improve your distance tremendously.

[madires]

I agree, 5mW is way too much for a few meters. And there's also the LOS problem (line of sight). A better solution would be LiFi (https://en.wikipedia.org/wiki/Li-Fi) which can achieve surprisngly high data rates.

[MarkT]

The laser diode is 5mW red laser.

Just into class 3B - you should be thinking class 1, less than 0.39mW visible unless you have a really good reason for more power.  0.3mW is millions of photons per nanosecond, to put into context, quite capable of high-speed signalling with suitable emitter and receiver.

[jbb]

I can’t go into too many details, but some years ago I worked on a system which got some MBit/s over a 1 meter air gap using infrared LEDs and On-Off-Keying.

Encoding was 8b/10b without forward error correction.

Receiver used an IR photodiode and transimpedance amplifier. It used an FPGA LVDS input port as a comparator, I think…

I didn’t dig into the analog guts, but I suspect the on off keying was suboptimal, because LED turn off was likely a bit slower than its turn on (ie it was asymmetrical). Maaaybe it’d be better to switch the light source between a high (100%) and low (10% ??) current source so it’s always conducting?

[soldar]

Just as a curiosity I will mention that many years ago Elektor magazine had a project which modulated a halogen light (a car's headlight) and, if I remember correctly, could send voice over a mile. The filament is so hot that it has low enough thermal inertia and can be modulated.

[Echo88]

ESP32 -> RJ45 -> http://blog.svenbrauch.de/2017/02/19/homemade-10-mbits-laser-optical-ethernet-transceiver/
http://ronja.twibright.com/
I would assume that the ESP32 is quite capable of reaching better datarates than what youre achieving right now and a FPGA is overkill, but thats just a guess.

[themadhippy]

Just as a curiosity I will mention that many years ago Elektor magazine had a project which modulated a halogen light (a car's headlight)

different mag,same idea
https://www.worldradiohistory.com/UK/Hobby-Electronics/Hobby-Electronics-1978-12.pdf