Working With Light
By Bryan Bergeron
A good friend asked if I could lend a hand developing a remote optical sensing device for his robot. The challenge was that communications between the remote environment and the local electronic sensors had to be carried via a single pair of optical fibers over several meters. Intrigued, I quickly agreed. After all — I asked myself — how hard could it be? I’ve worked with IR rangefinders, light radars, and even a pair of high voltage HeNe lasers. Plus, I had just finished experimenting with the IR LIDAR from a Neato vacuum teardown.
Well, the project that I assumed could be handled in a weekend has now stretched to over two months. It’s one thing to work with off-the shelf IR communications components and cables, but quite another to design an optical sensing system from the ground up. For example, even though most of my audio gear is connected via TOSLink optical cable, I had never tried to cut or splice the cable, or alter the end-cap lenses. I discovered that such an operation is non-trivial.
A challenge in working with fiber optics in general is the need for an infrastructure significantly different from that found in an electronics shop. Lacking the equivalent of a DMM or oscilloscope, I was virtually testing for battery voltage by placing the terminals on my tongue. That is, I was initially limited to using my visual perception of light intensity and area of illumination in place of quantitative tools.
I quickly learned that my drawers of standard electronic tools were useless when working with glass fiber. Not only is it fragile, but snapped-off ends have a tendency to seek out corneas like a heat-seeking missile. And once you break glass fiber, you might as well toss the entire fiber. That is, unless you have a $300 repair kit and know how to use it.
After wasting several hours polishing a few glass ends – the final step in a repair – I moved to plastic fiber. Plastic fiber is generally less efficient at carrying light than glass fiber, but is easier to work with. You can cut it with a sharp blade and then fire-treat the end to smooth it and reduce internal reflection. Of course, it helps to have a good pocket or even desktop microscope to examine the ends.
I love working with lasers, and this project was an excuse to purchase some hefty 100 mW IR red, green, and blue laser diodes; 100 mW lasers are “scary bright” – at least to me. Actually, 100 mW IR lasers are just plain scary. So, before powering up any of the laser diodes, I went online and ordered two pairs of laser goggles. It took two pairs to cover the four wavelengths of the lasers in my set.
As with glass and plastic fibers, laser diodes, and phototransistors, goggles are designed to work at specific wavelengths. Don’t even think of using a pair of generic sunglasses for protection if you experiment with lasers of any significant intensity.
Although I found cheap no-name goggles on the Web that cater to the laser light show industry, I didn’t trust the intensity reduction figures. Instead, I went to Thorlabs (http://www.thorlabs.com) for certified goggles. You might gasp at spending $150 or more for a pair of plastic goggles, but you’d probably gasp even more if you burned out your retinas because you trusted a $10 pair of no-name goggles. Another source for quality certified goggles is Phillips Safety Products (http://www.phillipssafety.com).
A major infrastructure technology is the perfboard equivalent in optical work: the optical bench. Think of a typical perfboard on steroids, with regularly spaced holes about 1/4” in diameter. Assuming you own standard mounts for your lasers, targets, lenses, mirrors, and other components, an optical bench makes setup quick and painless. I picked up a used 2 x 3 foot table-top optical bench on eBay. This isn’t a must-have, but it beat mounting modules on mounds of putty or magnets tie-wrapped to the laser cylinders.
If you work with optics and lasers, you should know about Edmond Optics (http://www.EdmundOptics.com). It’s a one-stop source for mirrors, lenses, and anything else you probably need to work with optics. I spent $30 on a “singlet” lens to concentrate the 3 mm beam from my red laser to about 1 mm. I wouldn’t even think about working with a concentrating lens without safety goggles and precautions to minimize accidental reflections. There’s a reason most of the hardware designed for laser and optics work is finished in dull black. Cover exposed stainless steel hardware with tape if you don’t have time to paint it.
As demonstrated by the latest generation of robotic vacuum cleaners, semi-autonomous cars, and, of course, the military drones, optics and optical sensing are definitely part of the future of robotics. With a modest investment in infrastructure and safety equipment, you can make optics a part of your future experiments, as well. SV