[Discuss-gnuradio] Problems with DC offsets, leakage and other stuff

[frankist]

Hi,

This is a problem I have been having for a while, but I've never given it
enough importance until now.

Basically I have a USRP2 as a receiver and no transmitter. So basically, I
am just measuring noise.
I am using the 5GHz band. 

Here are the fft plots for freq. 5GHz and gain 30 and 50. 
http://old.nabble.com/file/p33690634/pic1.jpg 
http://old.nabble.com/file/p33690634/pic2.jpg 

As you can see the peak at center frequency is eliminated by changing the
gain. So I guess it was just some DC offset that wasn't removed by the USRP.
However, the other peak at -30kHz approximately remains. I tried to use
uhd.tune_request thinking it was some leakage but the peak doesn't move.

When I use a center freq. greater than 5GHz and lesser than 5.5GHz I get
this: http://old.nabble.com/file/p33690634/pic3.jpg 
http://old.nabble.com/file/p33690634/pic4.jpg 

In this case I have a peak at center frequency which isn't attenuated even
if I change the gain. The peak at -30kHz remains also. I tried using the
uhd.tune_request treating these peaks as leakage but they didn't move at all
once more.

At 5.5GHz, I don't have any peak (I don't think it is necessary to post a
image in this case) except for a small DC offset which is eliminated with a
increase in the gain.
The peak at the center frequency (but not the -30kHz peak), however, appears
again for frequencies greater than 5.5GHz but less than 6 GHz.

Someone answered me in another post about the leakage problem and how to
solve it. However, those tools are not working in this case.
-- 
View this message in context: 
http://old.nabble.com/Problems-with-DC-offsets%2C-leakage-and-other-stuff-tp33690634p33690634.html
Sent from the GnuRadio mailing list archive at Nabble.com.