In general, the laser hazard depends on the laser power, the output beam diameter, and the laser wavelength. For a class IIIa or 3R laser (the "IIIa" designation is basically obsolete, although it remains in use for products certified before the new classes were defined), you're at low risk if you don't force yourself to stare into the beam. If the beam just happens to stray into your eye, you'll generally have a reflex response to look away from the painfully bright light. (Don't fight this reflex -- keep yourself safe). Note that the high end of the class 3R power limits is defined by the power where 50% of people will have an "aversion response" sufficient to avoid injury --- and the other 50% won't.
Reducing the current to a diode laser will reduce the output power, and thus make the output safer. However, most laser accidents don't happen when the laser is operated as intended or as planned. You also need to consider all the possible "fault conditions" or ways that things can go wrong.
Say you design a control circuit that regulates the laser output to always be less than 1 mW (in most cases, a "safe" level) using a feedback photodiode. For real safety, you should also consider things like
- What if the optical feedback path to the photodiode is blocked (by dirt getting in your box)?
- What if the electrical path from the feedback diode to its amplifier is broken?
- What if there is a spike or drift in the power supply to the laser or laser drive circuit?
- What if some bit of metal junk gets in your box and short-circuits the laser to the power supply, bypassing the power controller?
- How are you calibrating the limit value for the photodiode current? If you're setting it with a pot, could the pot drift or be mis-adjusted after calibration? If you're using a digital circuit, could the EEPROM with the cal data be erased or damaged?
- etc.
These are the kind of conditions that marketable laser products need to consider before they can pass regulatory requirements. Before you risk your eyesight with your laser system, you should at least consider doing this kind of analysis for yourself.
If you used a laser with less power capability, you would know that before any of these kind of hazards could lead to the laser producing a dangerous beam, the laser itself would burn itself up. If you use a laser capable of producing 5 mW before burning out, you'd be wise to treat it with a proportional level of respect.
1) Is it really as simple as grabbing a laser drive chip and a
three-terminal diode from digikey and hooking them up according to the
datasheet? Should the laser drive chip be able to handle all
protection mechanisms necessary, or is there typically another device
that's needed to handle some other form of protection?
The laser drive chips I'm familiar with are more about applying rapid modulation to the laser than providing DC power. Usually there's an additional power circuit required; and that power circuit is where the protection is normally implemented.
If you have a different type of drive chip in mind, please link the datasheet in your question.
2) Is there a central regulatory body that does any testing to
determine what class of laser you have, and whether your product
follows all the necessary regulations?
In the US, it's up to the laser manufacturer to self-certify their product. You may be able to find a consultant to assist you with that process if you don't have the expertise.
3) Are there any known issues using lasers with 1mm core plastic
optical fiber? I know that POF has very different transmission windows
vs. glass fiber, and I know that one of these optimal windows is
650nm.
Would the beam stay narrow inside the fiber, or would it begin
to disperse?
The fiber is a waveguide, and the laser power will remain confined within the fiber core. It will attenuate (lose power over distance). There is also a process called dispersion which means different components of the laser power taking different amounts of time to traverse the fiber---but if you're not switching the signal quickly that's not likely to affect you.
Edit: A major difference between POF and glass fiber is that even in its transmission window, POF has much higher attenuation than glass. Attenuation in glass fiber is measured in tenths of dB per km. Attenuation in POF (last time I worked with it, several years ago) is measured in tenths or whole dB per meter.
Would it still be coherent and collimated after going through, say, 15 meters > of POF?
The signal will still be coherent, but the dispersion effect I mentioned above may reduce the coherence length if you've gone through a very long fiber.
The output beam will diverge at a substantial angle (not strictly collimated) when it exits the fiber. The divergence is a diffraction effect and the angle is inversely related to the fiber core diameter --- meaning POF will have a lower divergence angle than smaller-core fibers. In multi-mode fiber like POF the output divergence angle also depends on details of the fiber construction. In general the output divergence angle will be similar to the input acceptance angle.
I am investigating the laser approach, because it seems like most LEDs aren't even capable of 500 uW.
It doesn't matter much what most LEDs can do --- if you can find one LED that meets your needs, that is enough. And I think you should be able to find an LED to produce 1 mW and couple into POF, if you look long enough. But a laser should be able to do it more efficiently (but maybe more expensively).
Edit: Be aware that using an LED does not reduce your safety concerns. 1 mW is still 1 mW and can still be dangerous. You will want the same safety precautions (you mentioned open-fiber control) whether you use a laser or LED. Regulations have not all kept up with the improved capabilities of LEDs in recent years, but that doesn't mean you shouldn't protect yourself and your users.
Best Answer
I would consider a small servo to move your laser right and left. Move it from the center of the angled spinning reflector to the edge. Draw out some angles and see what it looks like.