Can you construct a function where small oscillations have equal amplitude to larger ones?

10 replies · 4K views · last reply
Emilijo
Messages
35
Reaction score
0
In Fourier series we have small waves on the top of big waves (the function seems like that),
but the small waves do not have the same amplitude. Does somebody know how to get a function with waves and small waves on the top but with the same amplitude.
 
What is the definition of a small wave if it doesn't have anything to do with amplitude?
 
Line graph of a square-like periodic wave with x-axis ticks 2, 4, 6, 8, 10, 12 and y-axis ticks 500 to 3500; plateaus near 3000 and 750 show small ripples on top.


Can you see now, small waves on the top of big wave are not the same (equal amplitude)
{click on the picture to see it better}
 
Attachments
  • Plot of a periodic step-like function with a y-axis from 500 to 3500 and an x-axis from 0 to about 12, showing large square-wave levels near 3000 and 750 with small ripples riding on each plateau.
    images.jpg
    5.4 KB · Views: 508
If you're talking about the changing amplitude of the Fourier Series approximation then the answer is that you really can't. Due to the Gibbs Phenomenon, you'll have an overshoot at any discontinuity, of which the amplitude doesn't diminish.
 
Do you know how to get a function (any kind of function) with "small" waves on the top of "big" waves, but for the same amplitude of all small waves?
 
The Fourier series is of the form
[tex]\sum A_n cos(nx)+ B_n sin(nx)[/tex]

It looks to me like your series happens to have only two non-zero terms, one with a period of about 6 and amplitude 3000, the other with period about .6 and amplitude about 100. In other words, something like
[tex]3000 cos(2\pi x/60)+ 100 cos(10(2\pi/60))[/tex]
 
HallsofIvy said:
The Fourier series is of the form
[tex]\sum A_n cos(nx)+ B_n sin(nx)[/tex]

It looks to me like your series happens to have only two non-zero terms, one with a period of about 6 and amplitude 3000, the other with period about .6 and amplitude about 100. In other words, something like
[tex]3000 cos(2\pi x/60)+ 100 cos(10(2\pi/60))[/tex]

-Your function is not like on the atachment,
do you have better idea?
 
I found a function: sin(1-cos(x))
But there are only 2 "small" waves on every wave (put the function in wolfram)
How to get 3, 4, 5, ... or n "small" waves?