Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: mode locking vs. filtering of broadband noise



At 03:08 PM 12/9/00 -0500, Tom McDonald wrote:
Fisher burner ... carpet roll ... Is this "mode locking"?

You can have some fun figuring out the various possibilities.

First, you need a way to select modes. Cut a hole in the side of the tube,
at the midpoint. This will ruin the fundamental (and all odd-numbered
modes) while leaving the #2 mode (and all the even-numbered modes)
more-or-less unaffected.

Fire up the system with the hole covered. This should produce the
fundamental in the standard way. Then open the hole. The fundamental
should quickly die out and the #2 mode will grow.

Note that at this point we can already rule out the notion of "broadband
noise plus resonant filter". The #2 mode was available all along. The
alleged broadband excitation was available all along. The BBN+RF notion is
allegedly linear. Linearity implies superposition. So why is the #2 mode
not heard until now?

Now gradually close the hole. Find the point, if any, at which the system
switches from one mode to the other. See if you can find a hole-size that
exhibits hysteresis --- that is, the observed frequency depends on whether
that hole-size is reached from the larger or smaller direction. Check out
the effect of removing/restoring the flame.

One thing that muddies the water in the foregoing experiment is that for
the typical carpet tube, the #2 mode frequency is a multiple of the #1 mode
frequency. This makes it hard for the obvious mode-locking mechanism to
defend mode #2 against encroachment by mode #1. You can lift this
degeneracy in various ways. One way is to put a stub on the tube:

. | |
. | |
. | |
. | |_____
. | _____|
. | |
. | |
. | |
. | |

This stub will shift the frequencies of odd-numbered modes while leaving
the even-numbered modes more-or-less unaffected. Then experiment with
various-sized holes in the end-cap of the stub.