Re: [Discuss-gnuradio] USRP2 and external RF tuner

[Luca Pascale]

Hi Matt,

I have connected the RF tuner that provides (nominal) a gain of 30 dB. So the chain is (SIGGEN)--->(TUNER)--->(USRP2).
The tuner itself introduces spurious signal (below -70dBm).
The output spectrum is:

a) without tuner (description is above the picture)
http://imagebin.ca/view/ZTARvXd.html

b) with tuner
http://imagebin.ca/view/KOYDSsHI.html


Do I need more amplification on the IF signal ? 
Or does exist another way to obtain a clear spectrum with the injected tone only ?

Is it possible that some spurious are due to the downconvertion process ?

Regards,
Luca

2010/9/2 Matt Ettus <address@hidden>

The BasicRX has no amplification.  Its full scale amplitude is +10dBm. The ADC in the USRP2 has a rated SFDR of 88dBFS.  That means you may see spurs at levels high as -78dBm equivalent when using a BasicRX.  In the plots you sent me, your signal is -50 dBm and the spur is more than 40 dB lower, putting it at better than -90dBm, or more than 12dB better than the spec.  That is what you need to expect from a real-world ADC.

Normally, if you want to look at a -60dBm signal, you will amplify it a lot before you put it into the ADC.  That will make it much stronger than the spurs.  This is why boards like the RFX, DBSRX, and WBX all have amplifiers.  I suggest amplifying your signal by at least 30dB.

Matt

On 09/01/2010 06:10 AM, Luca Pascale wrote:

Hi all,

I am working with a usrp2+BasicRx and an external tuner with a 21.4 MHz
IF output and a bandwidth of about 20 MHz.
The same setup but with usrp1 (instead of usrp2) it's working well.

Now I am injecting a tone at 21.45 MHz (-60 dBm) at the input of the
BasicRX and set the usrp2 freq at 21.4MHz (with "set_rx_freq()" UHD API).
The problem is that:

a) with sample rate > of about 3.125M I have 2 strong spuriouses. The
first in the middle of the FFT (FFTDIM/2) and the second one at about
the middle of the second half of the FFT

b) with a sample rate <=3.125 the second spurious disappear but the
first at FFTDIM/2 is still present.

In both case I can also see the tone at 21.45MHz.
To compute the spectrum I use matlab
[10*log10(|fftshift(fft(signal*window))|^2)].


Am I doing mistakes or someone have the same problem ?
Suggestions ?

Thanks in advance

Luca

_______________________________________________
Discuss-gnuradio mailing list
address@hidden
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio