[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Discuss-gnuradio] [USRP-users] Switching and high spike in spectrum

From: bob wole
Subject: Re: [Discuss-gnuradio] [USRP-users] Switching and high spike in spectrum
Date: Fri, 31 Oct 2014 11:31:23 +0500

On 10/29/2014 01:54 PM, bob wole via USRP-users wrote:
> USRPN210r4 with SBX
> I am observing a strong spike at the center of the receive spectrum
> when I start burst transmission.
> My top flowgraph contains following two hierarchical blocks
>  1) A transmitter flow graph with (tx_time, tx_sob, tx_eob)
>  2) A receiver flow graph
> When I run top flowgrpah (without transmitting anything) and observe
> the FFT of the received signal the spectrum does not contain high
> spike in the center.
> But as soon as I start transmitting in burst mode I see a very high
> spike in center of the received signal FFT spectrum. It looks like LO
> (transmitter or receiver ) is being received? Which one is it ? And
> why is it happening?  How can I avoid it because it is affecting my
> packets.
> When I apply the offset in digital using DDC/DUC, the spike moves out
> of the band.
> --
> Bob
> _______________________________________________
> USRP-users mailing list
> address@hidden
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
That spike in the middle is a consequence of using direct conversion in
both the RX and TX paths--it'll be there in both to some degree.

You can use offset-tuning to move the DC offset outside your passband:


In built-for-a-particular-purpose radios, there will also be undesired
LO leakage and mixing products--those are generally dealt with using an
   application/band-specific filter to eliminate them.  For
general-purpose SDRs, that isn't possible to do "as manufactured", you
have to deal
   with RF hygiene and plumbing issues yourself.

So, moving the LO leakage outside your passband is part of the
picture--use offset tuning for that.  Then, if you have "this won't meet
   our hygiene requirements", you have to look at filtering.

Another thing you really should do is to run the calibration utilities,
which will attempt to balance I/Q amplitude and phase, which can improve
   some of these issues, but not, usually, eliminate them entirely.

Yes, I know that LO leakage/DC offset is an issue present in direct conversion receiver. But as I mentioned earlier, the received spectrum looks fine (a very little spike at DC around -70dB) while the burst transmission is not running. The spike becomes much more significant ( high spike at DC -20dB) when burst transmission (tx_time,tx_eob, tx_sob ) starts  and all the spectrum just shifts up and down with it. I am using TX/RX antenna in both usrp source and usrp sink. I want to know why the burst transmission is affecting the received spectrum on the same node. 

reply via email to

[Prev in Thread] Current Thread [Next in Thread]