FPGA Intellectual Property
PCI Express*, Networking and Connectivity, Memory Interfaces, DSP IP, and Video IP
6355 Discussions

help me with the fft ip core

Altera_Forum
Honored Contributor II
977 Views

I am doing a project on OFDM system which uses the FFT IP CORE, I use the ip core to do a 128 pionts transform . 

During the process,I met some problems which I will describe as follows: 

Firstly,I have a puzzle on the ip core .there are four kinds of I/O data flow architectural 

options supported by the FFT MegaCore function,they are : 

■ Streaming■ Variable Streaming■ Buffered Burst■ Burst 

My modelsim simulation waveform is: figure 1  

my stp wave is as follows figure2 

Rst is the reset_n of the ip core ,it is low first at the beginning of the system and after a while it keeps high .  

en_data: is a enable single of the system,which used to start the whole system . 

AddZero_slave : is a enable single of the IFFT module,it indicateds the input data of the ifft module . 

Data_real is the input data of the ip core. 

Sink_sop and sink_eop are the head and tail of in input OFDM symble. 

 

The input data is consists of numbers of frames, there are spaces between frames, during the frames the ena_data is high,and during the spaces the ena_data is low. 

 

Each of the frame is consists of OFDM symbles ,and each symble includes 128 data.(more than one time zeroes are padded in the middle). because of the clock,we design spaces between the symbles, you can see it from the picture one .the correct sop and eop of the input data can be got ,as the wave form shows .but we can’t get the correct output of the ip core . 

 

We use the streaming i/o data flow architectural. 

The fft ip core user guide shows that the streaming I/O data flow FFT architecture allows continuous processing of input data, and outputs a continuous complex data stream without the requirement to halt the data flow in or out of the FFT function. Our input data is not continuouce (there are spaces between the frames and OFDM symbles ),can we use the streaming mode ?(the modelsim simulation show the correct results ,but the stp wave doesn’t show correctly,so I am puzzled about this). 

 

I am much thankful for your help . 

0 Kudos
3 Replies
Altera_Forum
Honored Contributor II
270 Views

can you highlight where you are seeing the problem in the stp versus test bench?

0 Kudos
Altera_Forum
Honored Contributor II
270 Views

the differences between the both waveforms is the souce_sop and source_eop,in the stp file,the souce_sop and source_eop can't coincide with the output data.they are not the starting and the ending or the output data of the ip core. 

i hope you can help me .thank you
0 Kudos
Altera_Forum
Honored Contributor II
270 Views

Hi,aiwen 

I suggest you use the variable streaming mode.
0 Kudos
Reply