hi raul, 1)
is your sync signal an integer multiple of the product of a) the order in the FFT's re-order block b) the fft length c) the pfb filter taps for info see casper memo 26 at: https://casper.berkeley.edu/wiki/Memos 2) is your operating system supported by Xilinx ? (RHEL or Windows). we had a problem with 2^15 point FFT's - the data output was wrong, even though the design compiled with no errors and simulated OK. when we switched operating systems to RHEL5, the problem went away. best wishes, dan On Wed, Jan 7, 2015 at 8:45 AM, Raul Sapunar Opazo <[email protected]> wrote: > Hello! > > I'm having problems with spectrometers in ROACH 2 with a high number of > channels (> 32k). > Everything works right but the tone (and every other spurius) of the > spectrum shifts in 8 channels every 14 seconds aprox. > > My first thought was that the problem was in the sync signal, but after some > changes I haven't get any good results. > > Anyone has encountered this problem or have any ideas of what could be > happening? > > Best Regards, > Raul Sapunar > >
