Forum Discussion

Altera_Forum's avatar
Altera_Forum
Icon for Honored Contributor rankHonored Contributor
12 years ago

Problem with using Inverse FFT Megacore

I have a megacore FFT (length=32768, Buffered Burst) on the fpga and sending it data from the ARM processor (Cyclone V SoC)

I can successfully perform FFT on data.

But when I use the same core to perform Inverse FFT, it fails.

I check output of the FFT core when inverse=1 and inverse=0. Both are the same.

I assert & deassert 'inverse' according to the documentation. That is, there is a pulse for inverse input exactly when there is a pulse at sink_sop. But for the rest of the time it is a '0'

There are no source_errors.

I do not understand why the FFTcore output is same regardless of the input pulse at inverse.

Is there else that I am supposed to do for the Inverse function?

Thanks,

Surabhi

2 Replies

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    12 years ago

    --- Quote Start ---

    I assert & deassert 'inverse' according to the documentation. That is, there is a pulse for inverse input exactly when there is a pulse at sink_sop. But for the rest of the time it is a '0'

    Surabhi

    --- Quote End ---

    are you sure about this? I thought inverse input should one value throught its frame
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    12 years ago

    Hi Kaz. Thanks for that. Keeping the inverse at '1' throughout the frame helped. Now it does calculate the Inverse.

    I just always thought it was a pulse because in the guide it said " Inverse FFT calculated if asserted. Only sampled at SOP" So I figured it checks for inverse input only at the sop. I guess that was a bad assumption to make.

    Thanks for the help.