# [USRP2] Some problems by sending sinusoidal signals. :-(

Dear all,

the last discussion on sending and receiving the sinusoidal signal
raised to me a lot of questions.
I am also having some problems with this basic example.

I have searched over the whole archive without founding an answer.
I know that such questions may not be appreciated by a list of expert
people,
but I hope that some of you can enlight me on this problem.

I have two USRP2 with XCVR2450 (last firmware version):

• USRP A: sends and receives a sinusoidal signal.
• USRP B: receives the sinusoidal signal from USRP A.

This is my setup for USRP A (USRP B is equal but without transmit path):
http://img168.imageshack.us/img168/8176/grcedit.png

The fft and the scope of the sending and receiving path in USRP A seem
to be correct:
TX: http://img216.imageshack.us/img216/5880/txfftscopeusrp2.png
RX: http://img256.imageshack.us/img256/7669/rxsameusrp2.png

This is instead the output of fft and scope in USRP B:
http://img216.imageshack.us/img216/1213/rxdifferentusrp2.png

I really don’t understand why I am seeing such a signal on USRP B.
It seems that the sinusoid here is centered on a frequency of ~ -115KHz.

Also note the amplitude of the sinusoid: in USRP B is larger than in
USRP A, although A is actually the sender.

Best Regards,
Andrea Costantini

On Tue, Mar 17, 2009 at 12:32:57PM +0100, Costantini, Andrea wrote:

The fft and the scope of the sending and receiving path in USRP A seem to
be correct:
TX: http://img216.imageshack.us/img216/5880/txfftscopeusrp2.png
RX: http://img256.imageshack.us/img256/7669/rxsameusrp2.png

This is instead the output of fft and scope in USRP B:
http://img216.imageshack.us/img216/1213/rxdifferentusrp2.png

I really don’t understand why I am seeing such a signal on USRP B.
It seems that the sinusoid here is centered on a frequency of ~ -115KHz.

What exactly is your problem? Is it the frequency offset? If so, at what
frequency are you transmitting and have you sync’d the clocks? If no,
isn’t this what you’d expect?
Btw, how do you get these fantastic SNR values, did you connect the
USRPs
directly?

Also note the amplitude of the sinusoid: in USRP B is larger than in USRP
A, although A is actually the sender.

You are confusing values behind the ADC with power levels. This was
discussed in another thread not too long ago:

Don’t forget there’s a lot of stuff between GNU Radio and the SMD
connector.

Cheerio
MB

Costantini, Andrea wrote:

I have two USRP2 with XCVR2450 (last firmware version):

1. lower the amplitude of your sine wave, 1.0 is the max, you will have
clipping.

2. There was a small update to xcvr (after march 6th) that may fix your
problem. svn up your 3.2 branch, build the firmware, and reflash your sd
cards.

-Josh

On Tue, Mar 17, 2009 at 02:37:07PM +0100, Costantini, Andrea wrote:

Frequency is set to 2.4GHz. But I didn’t sync the clocks.
I am at the beginning of my research with USRP2 and I didn’t expect I would
have need to sync the clocks for such a preliminary experiment with sinusoidal
signals.
Can you please point me to some reference on how and why we do it? In the
meanwhile I’ll continue to search in the archive.

I’m afraid you misunderstand, although probably I wasn’t clear enough
and I must admit your pictures show the centre frequency - my apology.
At 2.4 GHz, a frequency offset of a couple of kHz is normal, i.e. you
don’t need to sync your clocks if all you want to do is check if the
sine wave was transmitted. If you wanted to build a real communication
link, you would need a frequency sync at the receiver. This is something
you can read up on in any good book on wireless communications.

MB

Hi Martin!

What exactly is your problem? Is it the frequency offset? If so, at what
frequency are you transmitting and have you sync’d the clocks? If no,
isn’t this what you’d expect?
Btw, how do you get these fantastic SNR values, did you connect the USRPs
directly?

No, actually they are around 50 cm apart from each other and connected
to two different PCs.

Frequency is set to 2.4GHz. But I didn’t sync the clocks.
I am at the beginning of my research with USRP2 and I didn’t expect I
would have need to sync the clocks for such a preliminary experiment
with sinusoidal signals.
Can you please point me to some reference on how and why we do it? In
the meanwhile I’ll continue to search in the archive.

You are confusing values behind the ADC with power levels. This was
discussed in another thread not too long ago:

Thank you very much.

Cheerio
MB

Andrea

On Tue, Mar 17, 2009 at 12:32:57PM +0100, Costantini, Andrea wrote:

This RX picture doesn’t look right to me. There should be a single
peak. Possible causes include pushing the transmitter into a
nonlinear region, overdriving the receiver input, having the gain set
too high, or seeing transmitter underruns.

Try setting the Signal Source amplitude to 0.1 instead of 1.0 and see
if that doesn’t clean it up.

Best Regards,
Andrea Costantini

Eric

This forum is not affiliated to the Ruby language, Ruby on Rails framework, nor any Ruby applications discussed here.