Hi all,

I am working on the 5510 DSK. For some strange reason, my output is
being inverted. I am implementing an FFT so I can tell it's being
inverted because the spike showing up is negative. I've checked the
values in memory that is being read and they are still positive. Even
the trigger pulse I have set is being inverted. What I mean by
inverted is what shows up on the oscilloscope is negative of what's in
the memory.=20

I was thinking it could be the AIC23 A/D D/A causing this problem, but
looking at the manual http://focus.ti.com/lit/ds/slws106d/slws106d.pdf
didn't get me anything. Does anyone have any ideas?

Any help appreciated. Thanks,

Thomas Shen

=20

=20

Thomas-

> I am working on the 5510 DSK. For some strange reason, my output is
> being inverted. I am implementing an FFT so I can tell it's being
> inverted because the spike showing up is negative.

Are you looking at a magnitude result of the FFT?

-Jeff

> I've checked the
> values in memory that is being read and they are still positive. Even
> the trigger pulse I have set is being inverted. What I mean by
> inverted is what shows up on the oscilloscope is negative of what's in
> the memory.
>=20
> I was thinking it could be the AIC23 A/D D/A causing this problem, but
> looking at the manual http://focus.ti.com/lit/ds/slws106d/slws106d.pdf
> didn't get me anything. Does anyone have any ideas?
>=20
> Any help appreciated. Thanks,
>=20
> Thomas Shen
=20

=20

Hi Jeff,

Yes, I am calculating/outputting the magnitude squared by individually
squaring and adding the real and imaginary parts. 

In memory, the result is a positive 16-bit signed integer as expected,
but it comes out negative on the oscilloscope.

Thanks for your reply,

Thomas

--- In c...@yahoogroups.com, Jeff Brower <jbrower@...> wrote:
>
> Thomas-
> 
> > I am working on the 5510 DSK. For some strange reason, my output is
> > being inverted. I am implementing an FFT so I can tell it's being
> > inverted because the spike showing up is negative.
> 
> Are you looking at a magnitude result of the FFT?
> 
> -Jeff
> 
> > I've checked the
> > values in memory that is being read and they are still positive. Even
> > the trigger pulse I have set is being inverted. What I mean by
> > inverted is what shows up on the oscilloscope is negative of what's in
> > the memory.
> > 
> > I was thinking it could be the AIC23 A/D D/A causing this problem, but
> > looking at the manual http://focus.ti.com/lit/ds/slws106d/slws106d.pdf
> > didn't get me anything. Does anyone have any ideas?
> > 
> > Any help appreciated. Thanks,
> > 
> > Thomas Shen
>

Thomas-

> Yes, I am calculating/outputting the magnitude squared by individually
> squaring and adding the real and imaginary parts.
> 
> In memory, the result is a positive 16-bit signed integer as expected,
> but it comes out negative on the oscilloscope.

Please explain how you are measuring FFT magnitude using a dig scope?  Are you
letting the scope internally calculate using time domain input?  Or are you
outputting the DSP memory contents in some way, for example using CCS GUI window?

-Jeff

> --- In c...@yahoogroups.com, Jeff Brower <jbrower@...> wrote:
> >
> > Thomas-
> >
> > > I am working on the 5510 DSK. For some strange reason, my output is
> > > being inverted. I am implementing an FFT so I can tell it's being
> > > inverted because the spike showing up is negative.
> >
> > Are you looking at a magnitude result of the FFT?
> >
> > -Jeff
> >
> > > I've checked the
> > > values in memory that is being read and they are still positive. Even
> > > the trigger pulse I have set is being inverted. What I mean by
> > > inverted is what shows up on the oscilloscope is negative of what's in
> > > the memory.
> > >
> > > I was thinking it could be the AIC23 A/D D/A causing this problem, but
> > > looking at the manual http://focus.ti.com/lit/ds/slws106d/slws106d.pdf
> > > didn't get me anything. Does anyone have any ideas?
> > >
> > > Any help appreciated. Thanks,
> > >
> > > Thomas Shen

Thomas-

> I am only using the oscilloscope to display the output of the dsp. The
> fft calculation is being done by the dsp chip.
> 
> I have the 5510 DSK, which includes analog input and output and A/D
> D/A converters. For input, I have connected a function generator with
> a sine wave. The DSP is supposed to calculate the fft, square the
> magnitude, and output the memory contents, which shows up on the
> oscilloscope.

Ok -- I see now.  In that case output may be inverted simply due to audio circuitry
on the DSK 5510 board.  Maybe the circuitry has a 1-stage amplifier that inverts.  In
any case, if you can invert the data prior to DAC output and it matches your mem
result, then I wouldn't worry about it.

> What do you mean by CCS gui? If you mean the "Graph" command in CCS to
> display memory contents graphically, I have tried that as well. The
> values are correct there.

Ok that's good.

-Jeff

> --- In c...@yahoogroups.com, Jeff Brower <jbrower@...> wrote:
> >
> > Thomas-
> >
> > > Yes, I am calculating/outputting the magnitude squared by individually
> > > squaring and adding the real and imaginary parts.
> > >
> > > In memory, the result is a positive 16-bit signed integer as expected,
> > > but it comes out negative on the oscilloscope.
> >
> > Please explain how you are measuring FFT magnitude using a dig
> scope?  Are you
> > letting the scope internally calculate using time domain input?  Or
> are you
> > outputting the DSP memory contents in some way, for example using
> CCS GUI window?
> >
> > -Jeff
> >