FPU-FFT input data

[Phillip Jacobs]

Hi,

I am measuring a frequency using the ADC (200ksps) and saving the samples into an array. I save 128 samples then move that data to the fft module to check what the signal is composed of. Here is where I am unsure if I need to change my data. The samples are spaced 5us apart with varying amplitudes - So I have XY coordinates. Do I have to change these points to polar form or leave as is and dump the data into the fft function? Or am I missing something very important?

I am still using the FPU FFT example to try understand the process.

Thank you for your help!!!

Regards,

Phillip

[Amanda Hsieh]

Hi, 

Do I have to change these points to polar form or leave as is and dump the data into the fft function?

In the FPU FFT Example, it processes 64 pairs of complex data (real, img). It is important to use proper arm_cfft_sR_f32 structure associated with input/output data length. For example: 128 numbers in input array (64 complex pairs of samples) -> 64 output bins power data -> &arm_cfft_sR_f32_len64. The fft_process() will executes arm_cfft_f32() and arm_cmplx_mag_f32(). Please also see the input format of arm_cfft_f32() and arm_cmplx_mag_f32().

-Amanda H. 

[Phillip Jacobs]

Thank you Amanda. I think my problem is that I am just using the real number and not the imaginary number (making it 0) Will have to change that.