FPGA Intellectual Property
PCI Express*, Networking and Connectivity, Memory Interfaces, DSP IP, and Video IP
Announcements
Intel Support hours are Monday-Fridays, 8am-5pm PST, except Holidays. Thanks to our community members who provide support during our down time or before we get to your questions. We appreciate you!

Need Forum Guidance? Click here
Search our FPGA Knowledge Articles here.
5875 Discussions

FFT Megafunction v12.0 issues

Altera_Forum
Honored Contributor I
1,544 Views

Hi, 

 

I was using a FFT Megafunction v12.0 to calculate 512-point FFT, and evaluating it on a DE2-115 board, which has a Cyclone IV E FPGA. The issue I have with this Megafunction is that it always gives incorrect results in the first two run, and gives the correct results in the third run and after, given that I feed the same data into the FFT core. Have any of you met the same problem? I don't understand why this is the case? Is it something to do with the Megafunction itself or it is related to my FFT controller design. Any input is welcome. 

 

Thanks, 

 

-Roger
0 Kudos
4 Replies
Altera_Forum
Honored Contributor I
68 Views

 

--- Quote Start ---  

Hi, 

 

I was using a FFT Megafunction v12.0 to calculate 512-point FFT, and evaluating it on a DE2-115 board, which has a Cyclone IV E FPGA. The issue I have with this Megafunction is that it always gives incorrect results in the first two run, and gives the correct results in the third run and after, given that I feed the same data into the FFT core. Have any of you met the same problem? I don't understand why this is the case? Is it something to do with the Megafunction itself or it is related to my FFT controller design. Any input is welcome. 

 

Thanks, 

 

-Roger 

--- Quote End ---  

 

 

I have across same problem. I use reset per each frame and from start and it sorted out
Altera_Forum
Honored Contributor I
68 Views

Thank you for the reply. Can you be more specific how you reset the core? Are you reset it multiple times before you feed the data in? Would you be able to share more in this regard?

Altera_Forum
Honored Contributor I
68 Views

 

--- Quote Start ---  

Thank you for the reply. Can you be more specific how you reset the core? Are you reset it multiple times before you feed the data in? Would you be able to share more in this regard? 

--- Quote End ---  

 

 

In my case the first set of outputs was wrong (I used streaming block floating arch.) .  

Processing was continuous for some hundreds of frames but with some gaps at available input. So I put one pulse of reset at start and in the gaps. I mean fft was only allowed to work after a reset release. 

Check first your own logic of SOP,EOP,Vld, and any fft error output. if you are applying reset at start then my reply doesn't apply to you really
Altera_Forum
Honored Contributor I
68 Views

Thank you for the details. I think I didn't apply the reset_n before the start, after I apply it, it looks good. Thanks for the tips.

Reply