First, we have set the damping force and driving frequency equal to those in Taylor’s book, to verify that the applet can reproduce the computer-generated graphs he shows,
which it does. But you can set the damping force over a wide range, and explore the differences.
The variable gamma is a measure of the relative importance of the driving force compared to the gravitational torque. For gamma less than one or so, gravity dominates.
The applet is initially set for gamma = 0.9. Run the applet and note it’s fairly close to sinusoidal, even though the pendulum is swinging above the horizontal.
Note the up and down amplitudes agree to five figures after six cycles with dt = 0.0005.
Now: gamma = 1, looks similar, but takes about 100 cycles for up and down amplitudes to agree to 5 figures. Gamma = 1.05: wild initial swing, then after four cycles or so,
settles but around 4 pi (it did two circles). Gamma = 1.06: takes 8 or so cycles to settle. 1.07: once it’s settled, track the heights of successive peaks (or dips) as
recorded in the amplitude, middle figure in bottom box. They alternate in height! The period has doubled. Try gamma = 1.073, compare with fig 12.5 in Taylor’s book.
These initial swings are very sensitive to gamma: try varying it by 0.001.
Try gamma =1.077 to find a period three final oscillating state. Now change the initial angle to 5 degrees.
Evidently there are (at least) two different final oscillating states possible, depending on
initial conditions. Such final states are called “attractors”.
Note the resemblance of the three cycle to the initial wanderings at other gamma values.
Use the amplitude readout to find further period doubling, at gamma = 1.081 there is period 4, at 1.0826 period 8, and at 1.0829 there is no periodicity: this is chaos.
Important! Near bifurcations, there are long-lived transients. Use the speed slider to speed things up: you can go to speed 50, run for hundreds or thousands of cycles in a short time, then pause, set speed to 1 or 2, play, and read out the amplitude, measuring the most recent extremum, to look for periodicity. The high speed doesn't affect accuracy, just prints a lot less dots. In fact, by looking at the pattern generated at speed, it's often easy to see the periodicity.
Next, try different values of damping, initial angle, gamma, etc.
Here's the relevant lecture, which includes more details.
Program by: Carter Hedinger