At telecom wavelengths GDD can be quite low. Laser source spectral widths can also be low. At visible wavelengths an fiber length imbalance of 1m with a 1nm bandwidth light source makes interferometry impossible/difficult without GDD compensation even if delays are matched. The moodulation bandwidth isnt an issue for this application but with a long enough fiber the source spectral bandwidth may be. Polarisation locked single transverse mode VCSELs are inexpensive and typically have spectral bandwidths of 100MHz or so.
Bruce > On 13 May 2018 at 11:31 Dana Whitlow <k8yumdoo...@gmail.com> wrote: > > > Indeed; however, with single mode fiber the limit is not too bad. At > Arecibo we routinely ran bandwidths in > excess of 1 GHz through fibers of about 1500 ft length with no problems. > For the science fair project a > bandwidth of a few MHz should suffice for lengths of, say, 500 ft. It's > just that I don't know how bad the > multimode dispersion problem would be when using shorter wavelengths, and > I'm sure not equipped to > make any measurements at home now that I'm retired and far away from the > observatory. > > Dana > > > On Sat, May 12, 2018 at 5:44 PM, Bruce Griffiths <bruce.griffi...@xtra.co.nz > > wrote: > > > Even with single mode fiber its finite group delay dispersion will likely > > restrict the usable light source bandwidth. > > > > Bruce > > > On 13 May 2018 at 03:38 Dana Whitlow <k8yumdoo...@gmail.com> wrote: > > > > > > > > > It may be that a nicely-written request to Corning could yield the loan > > of > > > a big spool of fiber > > > for the duration of a science fair project. > > > > > > Another alternative, perhaps easier to implement, might be an > > > electrically-driven light modulator > > > at the detector end. For the source, an LED or diode laser is easy to > > > modulate at respectable > > > rates. This approach should allow use of such high frequencies that an > > > open optical path using > > > mirrors might even suffice. > > > > > > Or here's an intermediate scheme: > > > If one were to use two modulated sources (or one with a beamsplitter), > > with > > > one path delayed > > > by the long(ish) fiber and the other by a minimal-length local fiber, > > > something resembling a streak > > > camera (implemented with a rotating mirror) might permit use of > > > substantially higher pulse rates > > > than with a rotating disk, without incurring the need for anything very > > > fancy in the way of mechanics. > > > Only the modulated source should require a reasonably accurate drive > > > frequency- the "detector" > > > would be essentially self-calibrating. A small mirror, say of cm size, > > > could probably be safely > > > rotated at Dremel speeds approaching 500 rev/s, and if 1 mrad angular > > > resolution is attained, > > > this would yield a resolution of ~160 ns. So a fiber length of 500 ft > > > (approx 750 ns one-way delay) > > > should yield an angular separation of nearly five "dots" between delayed > > > and undelayed dots. > > > And if the sources are modulated at a rate such that a few pulse > > > repetitions are visible in the > > > field of view, the scheme is self-calibrating as long as the PRF and the > > > velocity factor in the fiber > > > are known. Probably the only precision work would be the optics required > > > to focus a reasonable > > > amount of light from the source(s) onto the two fibers., and I believe > > this > > > requirement could be > > > adequately met with microscope objectives borrowed from one's school's > > > biology lab. > > > > > > A fly in the ointment is that if ordinary (read, inexpensive) IR fiber is > > > used at convenient visible > > > wavelengths, propagation will occur in more than one spatial mode, with > > > different modes propagating > > > at different speeds. I don't know how much of a problem this would > > > raise. But it may be that if > > > tweaking of the transmitting end illumination is done, both in angle and > > > transverse position, most > > > of the propagating light could be confined to a single mode. I speak of > > > visible wavelengths simply > > > because using these avoids the cost of electronic detectors, > > oscilloscopes, > > > etc, potentially saving > > > a lot on the cost of the experiment as well as making for a more > > satisfying > > > presentation. > > > > > > Dana > > > > > > > > > On Sat, May 12, 2018 at 9:45 AM, Bob kb8tq <kb...@n1k.org> wrote: > > > > > > > Hi > > > > > > > > > > > > > > > > > On May 12, 2018, at 7:01 AM, jimlux <jim...@earthlink.net> wrote: > > > > > > > > > > On 5/11/18 9:08 PM, Jeff Woolsey wrote: > > > > >> David.vanhorn wrote: > > > > >>> Measuring the speed of light (Fizeau or Michelson method? Other > > ways) > > > > >>> > > > > >>> > > > > >>> I saw a great demo of this at the Exploratorium in SF. They had a > > > > long spool of fiber optic, a disc with holes, and a light source. When > > > > static, if the light shines through the hole in the disc into the > > fiber, > > > > then you can see the light coming out the other end of the fiber > > through a > > > > different hole. When rotating, you increase speed and the fiber > > output > > > > gets dimmer and dimmer till it's gone. At that point, the light going > > > > into the fiber arrives when the other end is blocked, and vice versa. > > High > > > > tech, but simple. > > > > >>> > > > > >> My favorite exhibit that we never see anymore. IIRC it was a > > quarter > > > > >> mile of fiber and a green laser. And ISTR that the disc had one > > hole on > > > > >> one arm and two radially on the other, but I can't remember why. I > > > > >> thought that the light would pass through the same hole twice, once > > on > > > > >> the way in and on the way out when that same hole rotated 180 > > degrees to > > > > >> the other end of the fiber. The disk spun somewhere around 50 rps > > (60 > > > > >> with an AC motor?). > > > > > > > > > > > > > > > 1km in free space would be 6 microseconds round trip. I'm not sure a > > > > disk spinning at 3600 rpm would work. you'd need to have the "hole > > > > spacing" be on the order of 6 microseconds - and at 100 rps (6000 > > RPM), 10 > > > > ms/rev, you'd need the sending and receiving hole 6/10000 of a rev > > apart > > > > (about 0.2 degrees). > > > > > > > > > > if you had 10 km of fiber, it would be a bit easier. > > > > > > > > I think the term “long fiber” in this case should really be “very very > > > > long”. Exactly how the typical student > > > > funds the acquisition of something in the “many miles” range, I have no > > > > idea. > > > > > > > > You could use an optical grating of some sort as your “spinning disk”. > > The > > > > end of the fiber is going to be > > > > mighty small. The spacing on the grating could be quite tight. Where > > you > > > > get a circular part like that …. > > > > again no idea. > > > > > > > > Bob > > > > > > > > > > > > > _______________________________________________ > > > > > time-nuts mailing list -- time-nuts@febo.com > > > > > To unsubscribe, go to https://www.febo.com/cgi-bin/ > > > > mailman/listinfo/time-nuts > > > > > and follow the instructions there. > > > > > > > > _______________________________________________ > > > > time-nuts mailing list -- time-nuts@febo.com > > > > To unsubscribe, go to https://www.febo.com/cgi-bin/ > > > > mailman/listinfo/time-nuts > > > > and follow the instructions there. > > > > > > > _______________________________________________ > > > time-nuts mailing list -- time-nuts@febo.com > > > To unsubscribe, go to https://www.febo.com/cgi-bin/ > > mailman/listinfo/time-nuts > > > and follow the instructions there. > > _______________________________________________ > > time-nuts mailing list -- time-nuts@febo.com > > To unsubscribe, go to https://www.febo.com/cgi-bin/ > > mailman/listinfo/time-nuts > > and follow the instructions there. > > > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. _______________________________________________ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.