Trying to generate bode plot data for a 8196 points fft. it takes too long and never ends the calculation, what to do?

Hello,

i am trying to get the data for Bode and Nyquist plot using a FFt with 8196 points. I got the sine sweep move using a dc motor and got the data. Converted it to fft and calculated the Transfer Function. All i need is to plot the bode plot but it takes a lot time and never ends (tried for 30 minutes and then killed matlab). So is there a way to increase the speed or make it efficient to get data faster?

I tried the same using less number of points i.e. 512,1024,2196 and took 2 minutes, 5 minutes and 12 minutes.

I want to get faster results.

Answer 1

[Pratyush]

Hi Harshul,

To speed up Bode and Nyquist plot generation from FFT data in MATLAB, especially with a large dataset, consider the following strategies:

1. Optimize FFT Calculation

- Use fewer points if high resolution isn't critical.

- Ensure the number of points is a power of 2 for efficiency.

2. Efficient Transfer Function Calculation

- Vectorize calculations to avoid loops.

- Use sparse matrix operations if applicable.

3. Plotting Optimization

- Decimate data to plot fewer points without losing plot integrity.

- Use efficient plotting functions like 'semilogx' or 'semilogy'.

4. Parallel Computing

- Utilize MATLAB's Parallel Computing Toolbox for faster computations.

5. Efficient Data Handling

- Preallocate arrays to avoid dynamic resizing.

- Choose appropriate data types ('single' vs. 'double') based on precision needs.

6. Use MATLAB Profiler

- Identify and address bottlenecks in your code with MATLAB's profiler.

Hope this helps.