Application of FFT Fast Fourier Transform on MM32F013 Single Chip

yaoquan5201314

Application of FFT Fast Fourier Transform on MM32F013 Single Chip [Copy link]

Hello everyone, I am now using FFT fast Fourier transform on MM32F013 to perform frequency division processing on the mixed data read by ADC. The following is the FFT code written by other colleagues before. Because it is a long time ago and the handover is not clear, and the comments are garbled, I wonder if there are any netizens who can understand this part of the code?

#ifdef FAST_FFT

for (i = 0; i < 5; i++) //first, clear all
{
dft_temp1[i] = 0;
dft_temp2[i] = 0;
}

for (i = t_start; i < t_end; i++)
{
p = data + (i * TN); //μ??·??ò?
dft_temp1[0] += (*(p + 1) + *(p + 9) - *(p + 11) - *(p + 19));
dft_temp1[1] += (*(p + 2) + *(p + 8) - *(p + 12) - *(p + 18));
dft_temp1[2] += (*(p + 3) + *(p + 7) - *(p + 13) - *(p + 17));
dft_temp1[3] += (*( p + 4) + *(p + 6) - *(p + 14) - *(p + 16));
dft_temp1[4] += (*(p + 5) - *(p + 15)); //sin

dft_temp2[0] += (*(p + 4) + *(p + 16) - *(p + 6) - *(p + 14));
dft_temp2[1] += (*(p + 3) + *(p + 17) - *(p + 7) - *(p + 13));
dft_temp2[2] += (*(p + 2) + *(p + 18) - *(p + 8) - *(p + 12));
dft_temp2[3] += (*(p + 1) + *(p + 19) - *(p + 9) - *(p + 11));
dft_temp2[4] += (*(p + 0) - *(p + 10)); //cos
//dft_temp2[5] += *(p+15);
}
//sumim = 3164*dft_temp1[0]+6019*dft_temp1[1]+8284*dft_temp1[2]+9739*dft_temp1[3]+10240*dft_temp1[4];
//sumre = 3164*dft_temp2[0]+6019*dft_temp2[ 1]+8284*dft_temp2[2]+9739*dft_temp2[3]+10240*dft_temp2[4];//-dft_temp2[5];

sumim = 316 * dft_temp1[0] + 602 * dft_temp1[1] + 828 * dft_temp1[2] + 974 * dft_temp1[3] + 1024 * dft_temp1[4];
sumre = 316 * dft_temp2[0] + 602 * dft_temp2[ 1] + 828 * dft_temp2[2] + 974 * dft_temp2[3] + 1024 * dft_temp2[4]; //-dft_temp2[5];

#else
p = data; //+ (s16)(t_start*TN); //×￠òaμ??·??ò?
for (i = t_start * TN; i < t_end * TN; i++)
{
sumre += *(p + i) * costab[i % TN];
sumim += *(p + i) * sintab[i % TN];
}
#endif

My problem now is to figure out the Fourier series of this part of the code, and then get the data processing effect I want by modifying the appropriate series. I hope someone familiar with FFT fast Fourier transform can help analyze and solve this problem. Of course, you can also pay for consultation. Thank you!

huo_hu

This section has no beginning or end, and even the data type is unknown. It is part of the function and is useless.

You should check if there is any library available on the official website. If that doesn't work, you can modify the DSP of stm32

秦天qintian0303

Is it a bit difficult for a single-chip microcomputer to perform Fourier transform?

huo_hu

Qintianqintian0303 posted on 2023-3-22 16:44 Is it a bit difficult to use a single-chip microcomputer to perform Fourier transform?

Same, now TI is embarrassed to be called DSP

freebsder

I don't understand and can't help you.

Gen_X

Directly use the routines in DMK. Different functions correspond to different depths.

Just pick one for yourself, it's easy.

Pay attention to the correspondence between the operation cycle and sampling. Many "computer enthusiasts" do not understand this, which leads to strange calculation results!