Problems of lost of BINS

GNU Radio
1

Hi folks,

I have implemented a multi sub-carriers AM transmitter and here I still
have a strange problem: I plotted the FFT of the modulated signal at
baseband (as attached) and the number of the bins (spikes) I have on the
plot was exactly the same as the number of the sub-carriers as I set in
the program. This is correct so far as I know. But after the signal was
tuned to 2.45GHz by USRP, the number of the bins(spikes) on my spectrum
analyzer was not the same, which is only 4 or 6. While I have 20
sub-carriers at different frequencies and it seems that I lost most of
them after the signal was tuned to RF. Is there anybody who knows the
reason behind that? I have been stuck at here for a long time. Any helps
or suggestions will be greatly appreciated.

Thanks,

Milo

2

On Wed, Sep 16, 2009 at 11:34:38AM -0700, Milo W. wrote:

them after the signal was tuned to RF. Is there anybody who knows the
reason behind that? I have been stuck at here for a long time. Any helps
or suggestions will be greatly appreciated.

Thanks,
Milo

The USRP won’t be transmitting at exactly 2.45GHz because of the
tolerance in the oscillator. Is the span on your SA set wide enough
to see it, even if it’s not exactly on frequency?

Eric

3

Eric B. wrote:

analyzer was not the same, which is only 4 or 6. While I have 20
tolerance in the oscillator. Is the span on your SA set wide enough
to see it, even if it’s not exactly on frequency?

Eric

Thank you. Eric.

Yes. I think the span is wide enough(I searched from 2.3G to 2.7G). The
center frequency is just a little bit below 2.45G(aroud 2.446G). Is that
possible that amplitudes of the “lost” bins are too low to see and just
like random noise on SA? or something else problems?

Thanks,

Milo