Wavelength Options for my Thesis

[Trevor]

Although I'm a freshman (Virginia Tech), I'm beginning to work on my thesis as of this spring. It's part of the University Honors program here (not a specific college), therefore the thesis process is more laid-back than most.

I've settled on working with line-of-sight data transfer using laser, as this is a good marriage of my major (Business Information Technology), my interest in circuitry, my interest in laser, and my background in programming.

Before I can apply for grant money, I'd like to provide a smaller-scale proof-of-concept. I've worked with my EE friends to draw up bare-bones circuit diagrams that perform the system's most basic needs, and I've begun writing the firmware in what spare time I can scrape together.

The last main decision that needs to be made is wavelength. I've narrowed my options to 405nm, 532nm, 650nm, and 808nm. I've drawn up my list of perceived pros and cons, and I'm interested to see what you guys think.

1. 405nm
   • Pros
     • Easier collimation - lower divergence?
     • Medium affordability at medium to high power.
   • Cons
     • High susceptibility to Rayleigh Scattering - less light making it to the sensor?
     • Lab systems not easily obtainable?
2. 532nm
   • Pros
     • Medium collimation ability due to wavelength.
     • Less susceptible to divergence due to wavelength.
     • Lab units easier to come by.
   • Cons
     • Must be DPSS - diodes are prohibitively expensive.
     • Higher power requirement due to pump diode.
     • Higher heat output due to powerful pump diode.
3. 650nm
   • Pros
     • Fairly low cost to power ratio.
     • Widely available.
     • Lower susceptibility to Rayleigh scattering.
   • Cons
     • More susceptible to divergence due to wavelength.
4. 808nm
   • Pros
     • Very low cost to power ratio.
     • Widely available.
     • Lab units easy to come by.
   • Cons
     • Invisible, hard to aim!
     • More dangerous due to visibility.

This list was compiled from tidbits I've read around the forum. If I've misinterpreted anything, please call me out on it so I can rearrange my list and correct things. I'm eager to learn and want to make the most educated and thought-out decision possible.

Thanks,
Trevor

 

[jeffd]

All the 650nm lasers ive had, have great divergence and the higher powered 808nm diodes have bad beam profiles.

 

[Marcusss]

Green imo, will make it fun to write your thesis on that colour <3!

 

[jwc]

What kind of distances are you looking to transmit from? I remember seeing a thread where some guys used cheap 5mW red pointers to transmit data several feet in the KHz frequency.

If you're doing one-way communication, your system will probably have four main components: the encoder, laser, receiver, and decoder. The encoder and decoder will allow you to serialize your data stream into a digital signal. As for the optoelectronics to work with you need to consider the receiver as much as you consider the laser. If you intend to just use a photoresistor as the receiver, you'll need to create some harnessing optics and will need to tune the system to dismiss any noise (a more advanced encoder/decoder model will use checksumming in 2-way communication to ensure accurate data transfer).

As for the wavelength, it's cheap enough to try numerous wavelengths. But, you'll probably find that the best performance-for-price optics are typically for red diodes. 405nm diodes are really just taken out of blu-ray players nowadays and you'll need a good deal of correcting optics in order to get a nice beam that will really travel.

Green would be kinda fun to try; there are certainly plenty of optics out there that will allow you to get very good beam profiles (if it's good enough for laser shows, it's probably good enough for your purposes). But overall, I would recommend working with 632nm or 650nm red; it's fairly cheap overall.

If you can, try building a system that would allow you to swap out the lasers so you can test multiple wavelengths. I'd be interested in seeing how fast you can get the transfer rates for each wavelength.

Anyways, sorry if I was being too verbose there... those were just my preliminary thoughts on the project. It should be a fun one; good luck!

Edit: forgot to mention that you probably won't need very much power if you use filtering optics, etc. to let the photodetector see just the wavelength that you're transmitting.

 

[charliebruce]

Green probably isn't the best for this - I've heard that over a certain frequency, modulating green creates an unreliable output, unless you use an AOM (Acousto-optical Modulators (AOM)). You could also try experimenting sending at 4 seperate intensities instead of binary, so doubling your data throughput but maintaining a pretty high reliability. The colour and relatively good divergence of 532nm lab units would make it easily aimable as well.

 

[Cyparagon]

AOMs are pricey. It's a little over-complicated anyway unless you need a ridiculous bandwidth. It really should be direct injection. The lower power (<150mW I believe) IR diodes have good divergence. I would ask the distance required as well, but for now I would suggest a basic <5mW red diode.

 

[aXit]

Out of curiosity, is there any problem with just leaving the laser on continuously, and placing an LCD in front of the beam. It would fix the modulation issues with green...

I have a project planned out for a mircocontroller based spirograph, with some modulation capabilities, I was thinking of using a LCD.

Any thoughts?

 

[maxh]

If you really want a lot of bandwidth maybe you could invent a laser transistor! Scientists break light modulation speed record--twice

Out of curiosity, is there any problem with just leaving the laser on continuously, and placing an LCD in front of the beam. It would fix the modulation issues with green...

I have a project planned out for a mircocontroller based spirograph, with some modulation capabilities, I was thinking of using a LCD.

Any thoughts?

I don't know off the top of my head what kind of modulation speed you can get with LCDs, but I don't think it's very fast. Maybe a couple hundred Hz? If you were to try, PWM probably wouldn't be the best route. But perhaps you could go for an analog type modulation instead. Differing electric field strength can give you shades, rather than being fully transparent or fully opaque. Or if you have a high res lcd panel, perhaps you could control how much light gets through by having some pixels open and some closed. Imagine a checkerboard for cutting the intensity of your beam in half.

 

[ProjMan]

To be frank, the fact that you're asking questions about wavelength indicates that you're in over your head and don't have the background to have a reasonable chance of pulling this project off. Laser communications is more a matter of electronics and optics hardware than programming. Your major likely won't give you the background to build a device like this--this kind of project is unlikely to be appropriate for a thesis unless you're majoring in EE or optical engineering. If you have strong interests in this area, then look very seriously at switching majors to engineering.


If you do want to go ahead with this, then red or 405nm are good choices. DPSS (e.g. green) lasers can't readily be modulated at speeds useful for data transmission and IR will be a pain to work with. Power will need to be actively limited to <5mw--as demonstrated with a calibrated LPM--for legal and safety reasons.

 

[Things]

Green would be out, as it is too hard to modulate, which means you would have to use diodes. IR would be boring and hard to align. I would say either 405 or 650.

405 scatters much more than 650 in air, so IMO red would be the best wavelength.

 

[jwc]

Power will need to be actively limited to <5mw--as demonstrated with a calibrated LPM--for legal and safety reasons.

Not really. Colleges typically have regulations with regard to Class 3b and 4 lasers where they set specific standards that turnkey systems must have to be registered on their campus (I know that mine does). Prototype systems are typically given a little more leeway, but are usually required to have certain safety features such as external signs integrated into the power system.

However, you don't need very much power to be able to transmit a decent distance. With appropriate noise reduction and aiming, you should be able to go for miles. Of course, this depends on where you're planning to set up tests; you won't get the same kind of range in a city that you will in the middle of nowhere, obviously.

One of the advantages of IR is that you can block out the visible spectrum or use a detector that works only in some specified bandwidth. Typically, there's far less IR noise than visible noise.

 

[Trevor]

What kind of distances are you looking to transmit from? I remember seeing a thread where some guys used cheap 5mW red pointers to transmit data several feet in the KHz frequency.

I'm looking to start with around 30m, and I'm probably going to culminate in 250-300m. Hmm, red. Okay.

Green probably isn't the best for this - I've heard that over a certain frequency, modulating green creates an unreliable output, unless you use an AOM (Acousto-optical Modulators (AOM)). You could also try experimenting sending at 4 seperate intensities instead of binary, so doubling your data throughput but maintaining a pretty high reliability. The colour and relatively good divergence of 532nm lab units would make it easily aimable as well.

AOMs are pricey. It's a little over-complicated anyway unless you need a ridiculous bandwidth. It really should be direct injection. The lower power (<150mW I believe) IR diodes have good divergence. I would ask the distance required as well, but for now I would suggest a basic <5mW red diode.

AOM sounds good... just with a research grant. Chances are I can get it.

Out of curiosity, is there any problem with just leaving the laser on continuously, and placing an LCD in front of the beam. It would fix the modulation issues with green...

I have a project planned out for a mircocontroller based spirograph, with some modulation capabilities, I was thinking of using a LCD.

Any thoughts?

That's a very interesting idea.

To be frank, the fact that you're asking questions about wavelength indicates that you're in over your head and don't have the background to have a reasonable chance of pulling this project off. Laser communications is more a matter of electronics and optics hardware than programming. Your major likely won't give you the background to build a device like this--this kind of project is unlikely to be appropriate for a thesis unless you're majoring in EE or optical engineering. If you have strong interests in this area, then look very seriously at switching majors to engineering.

The whole purpose of the exercise is cross-discipline collaboration. Not writing a long paper on things you knew before you came to school. Diving in headfirst is the way to go here - since it's supposed to span all of our years in University Honors. I'm working with two EE's in my building; they have the hard electronics knowledge that we're marrying with my original drawings, firmware, and laser addiction. Without reposting the whole set of criteria, this is the objective in a nutshell.

Green would be out, as it is too hard to modulate, which means you would have to use diodes. IR would be boring and hard to align. I would say either 405 or 650.

405 scatters much more than 650 in air, so IMO red would be the best wavelength.

Seems like the critics agree - red is the way to go.

One of the advantages of IR is that you can block out the visible spectrum or use a detector that works only in some specified bandwidth. Typically, there's far less IR noise than visible noise.

I'll put 808nm definitively on the list.

-Trevor

 

[ossumguywill]

I'll put 808nm definitively on the list.

-Trevor

you might take it off your list. You can filter ANY wavelength, EG with red or blue plastic, and there is a very small noise difference, also you can use a squelch filter to get rid of light noise problems. Also, I would recommend that you simply use an arduino on both ends, simply take the data, send it as binary through one of the analog ports, or through a digital port with an amp between the arduino and the diode. On the other end, simply program the arduino to read at the same baud rate as the transmission, slow baud rates will be easier to sync, otherwise the easy way 2 xbee transceiver modules to sync them above 100 hertz, at which point it becomes pointless to use a laser. If you can figure out a way to get the modules to sync by laser, which will be tricky, your thesis will actually have practical uses.

 

[Cyparagon]

Out of curiosity, is there any problem with just leaving the laser on continuously, and placing an LCD in front of the beam. It would fix the modulation issues with green...

The typical LCD takes 5ms to refresh, which means your modulation speed would only be 200Hz. That is a throughput of 25 bytes a second. So if you combined two thousand of the systems together, you would have dial-up.

 

[ossumguywill]

The typical LCD takes 5ms to refresh, which means your modulation speed would only be 200Hz. That is a throughput of 25 bytes a second. So if you combined two thousand of the systems together, you would have dial-up.

line of sight systems have TERRIBLE modulation like this anyway. Unless you go analog for voice or whatever.

 

[steve001]

Although I'm a freshman (Virginia Tech), I'm beginning to work on my thesis as of this spring. It's part of the University Honors program here (not a specific college), therefore the thesis process is more laid-back than most.

I've settled on working with line-of-sight data transfer using laser, as this is a good marriage of my major (Business Information Technology), my interest in circuitry, my interest in laser, and my background in programming.

Before I can apply for grant money, I'd like to provide a smaller-scale proof-of-concept. I've worked with my EE friends to draw up bare-bones circuit diagrams that perform the system's most basic needs, and I've begun writing the firmware in what spare time I can scrape together.


Thanks,
Trevor

Red I think would be your best option for long distances. See what these guys have done.
laser contact to france