I don't have a part number

Hi Abhishek,

I've forwarded this to the design team. Their feedback should be posted here. But could you give more detail about the algorithm which you are using?

BR
Tsvetolin Shulev

Abhishek,

Please stop using the "I don't have a part number" subject line. You do have a part number, the TMS320C6713. It will help us and will help other users if you use a subject line that is related to your question.

There is not enough information in your post to determine whether what you have as output is correct or not. The output of an FFT is not a time-observable waveform. The function transforms the input from the time domain to the frequency domain, and frequency domain data is not time-observable unless other factors are considered.

The output shown does not look like a good signal, but it does have some components of a sine wave. Is the apparent oscillation period the same as your input sine wave?

Regards,
RandyP

No sir my input is a pure sine wave signal of frequency 730 Hz as  shown below. My output oscillation period is not same as the input.

Using FFT algorithm I got the o/p as shown below:

Now can you tell me Whether this o/p is correct or not? And If not what is the correct o/p for the given signal?

Abhishek,

As I explained above, the output of an FFT is not time-observable. So your output is not correct. It is not clear what your display shows. If it is intended to represent the output of an FFT, then more information is needed.

I must assume that you understand what an FFT is supposed to calculate. The output should be mostly small numbers (noise) and two large pulses representing the frequency of the input.

Your display appears to be mostly noise, but there is no way for me to interpret what you are showing. A more meaningful output display would be a CCS graph showing the values in your FFT output buffer.

Why did your first output display have a smooth but unusual waveform, while your second output display looks more like an ungrounded scope probe's noise?

Have you written your own FFT algorithm or are you using a library function from TI?

Regards,
RandyP

Sir, I got same o/p in both the cases but the difference is that in first case I kept volts/div as  20mV/div but in 2nd case I kept volts/div as 5V/div. I am using the library that is been provided by TI only and also I observed the o/p in CCS.

Now I have modified the program and I took real time i/p from the function generator as sine wave but I have not got the o/p as it is in the CCS(non real time).

Abhishek,

Please go to the dsplib_C674x folder where your library is installed, and find the examples folder. In there should be an example program or project that shows how to use the fft library function. Please load and run that example. Set the constant NUM_SIN_WAVES to 1 to best match your desired waveform.

Use the CCS graph tool to view the input buffer. The input buffer graph will show you how the data should appear for the function to operate properly. You can compare that to what your own program's input buffer looks like when you run it.

Use the CCS graph tool to view the output buffer. The output buffer graph will show you how the data should appear when the function is operating properly. You can compare that to what your own program's output buffer looks like when you run it.

After you have the example project running correctly and can observe its input and output within CCS, then you will be ready to try to capture an external signal and run the FFT on it.

Regards,
RandyP

Sir I have done the program of FFT in CCS and  also got the o/p of  which u are talking about.

Do you have any program of FFT using real time signal?

Abhishek,

I do not have any program of FFT using real time signal. I believe your best benefit will be from the archived C6x1x DSP Integration Workshop. Please go to TI.com and search for "c6713 workshop" (no quotes) and look for that course material. You will find student guides and labs that do include the use of realtime signal, although not your specific application using an FFT.

Since I am not in a position to give you what you want, and we have passed the original intent of this thread, I will drop off so others more helpful will have the chance to discuss this. If that is the case.

Regards,
RandyP