http://users.erols.com/mshaver /bikegen.htm Pictures of a floor mounted generator for an unmodified bike. Table of power outputs and results with different loads. Suggests that 100V @ 150 watts is possible!!
http://www.los-gatos.ca.us /davidbu/pedgen.html Pedal power generator; suggests that humans generate "peaky" power, and that a flywheel helps even out the load. (Not necessarily a disadvantage for lights - a bit of flaring and pulsing might be a good thing.) Again, suggests that 150 watts is practical for someone in good shape. "I could approximate the output of a 10 amp battery charger." Sounds like some kind of voltage control is important in order to not burn out the lights.
lubrication will be important for generator longevity, given the dusty playa environment. Sealed generator housings?
If we use LEDs instead of lightbulbs, this will be piercingly bright. (one hour of pedalling = 320 hours of glow??? one bike could drive hundreds of LEDs...)
Probably better to generate DC and run it directly than to use an inverter to drive 120VAC bulbs.
Can use a battery to smooth out voltage; that reduces feedback strength, though. It might be nice to cap the battery output at some low level, so the light always glows dimly, but only human input makes it bright.
possibility: get exercise bikes from a thrift shop instead of converting regular bikes. They are already set up for stationary operation.
http://www.econvergence.net/elect ro.htm floor generator for an unmodified bike - available in plans or assembled claims "150 to 200 watts at 12 to 20 VDC" - sounds high but plausible
http://ww2.green-trus t.org:8383/2000/humanpower.htm some discussion here says humans can only sustain about 75 watts
Personal experience with bicycle wattage: I spent 15 minutes on the stationary bike at the gym yesterday afternoon, about 95-100 RPM at difficulty level 9, and the bike claimed I was generating 149 watts. I am in reasonably good shape, but I couldn't have kept that pace up for more than 20 or 25 minutes no matter how motivated I was. The average playagoer probably won't be willing to sustain that pace for more than a couple of minutes. This suggests that we should shoot for something more like 60-75 watts per bike as a more normal figure. We also need to engineer the feedback system to reward people for putting out energy, but not reward them so much that they pump furiously and wear out after a minute or two. Perhaps there should be time-based feedback as well as output-based feedback: something happens after you've been pedalling for five minutes, something even cooler after ten minutes. Maybe this would be a good application for those penlight lasers: your light's brightness depends on the amount of power you put into it, but each light also has a "laser halo", and the longer you pedal the more lasers light up.
http://www.zetatalk3.com/ene rgy/tengy05g.htm idea: put an ammeter in line with the power from the bike - mount it on the handlebars or something, so the person can see how much power they are contributing.
this link describes a no-weld system using an automotive alternator; other links suggest that auto alts are not the best idea (require power to "excite" at low RPMs? designed for high-RPM, high-power applications)-
A quick, ugly "pedal generator" made from an old 3 speed bicycle. This page demonstrates the design I originally had in mind: the generator takes the place of the rear hub. Apparently "more poles" means it works at a lower RPM? These folks got 60 watts pedaling casually, 120 watts pedaling hard.