fft function in matlab and fft block in simulink dont match

Started by lolo_mk1 June 2, 2005

thanks to everyone who helped me with my last question, lots of
responses and fast too :)

I have a simulink model running where I am performing FM and
measuring adjacent channel power. I can't really quantitise this
inside simulink so I send the output of my FM modulator to my
workspace where I perform the FFT.

During the simulation i can watch the output of the modulator with a
FFT block and things look good. When I modulate with a square wave
there is alot of power in the adjacent channel (around -50dB) and if
I remove the modulating signal so I just have a carrier, the
adjacent channel is around -120dB. So I think all of this works fine.

Back in standard matlab I wrote a small function to look at 50%
overlapping frames (pretty much what the simulink fft block does)
and calculates the average adjacent channel power ratio...which
always ended up being around 35dB no matter what kind of modulating
signal was being used...so something is wrong here.

On closer inspection it seems that the PSD plot I get from the fft
doesn't go any lower than about -50dB (in my frequency range of
interest) compared to my simulink model which shows it around the -
80 to -100dB region.

Has anyone experienced this before? or knows a work around? It could
be caused by something that I'm doing wrong...but I've checked my
working many times and it's really not that complicated:/ I'm just
stumped for directions to go at this point.

I'm happy to send plots/files etc to better explain the problem