Re: Spurious Spikes Seen with usrp_fft.py

GNU Radio
1

Robert Fitzsimons wrote:
Date: Sun, 20 Jan 2008 12:57:57 +0000
From: Robert Fitzsimons
Subject: Re: [Discuss-gnuradio] Spurious Spikes Seen with usrp_fft.py
To: [email protected]
Message-ID: 20080120125757.GA10697@localhost
Content-Type: text/plain; charset=us-ascii

Hi Thomas

You email never made it into my mailbox, maybe it got caught as spam.

I am using a USRP with only a DBSRX daughterboard installed. The image
at the following link is a screen capture I took.
http://img340.imageshack.us/img340/3570/screenshotusrpfftfy8.jpg
Does anyone know why the spike at 921.6 MHz is there?

The spike doesn’t look like it’s from a GSM base station. What country
are you in? Have you tried tuning the usrp_fft.py to the 921.6 MHz
signal?

I have observed many of these spikes in other parts of the spectrum
using usrp_fft.py. I looked for them using a real spectrum analyzer
connected to an antenna and did not observe them.

Maybe your USRP is picking up some interference, can you try it with a
different computer and with a different power supply?

Robert

Thanks for you reply, Robert.

I am in the U.S.

I have used usrp_fft.py to tune to 921.6 MHz, and the spike is still
there at 921.6 MHz.

I tried powering the USRP from batteries, but the spike was still there.
It was also still there using a different computer. I tried using a USB
cable with a ferrite bead, but it didn’t change anything.

I positioned the USRP very close to the antenna connected to a real
spectrum analyzer, and I was able to observe the spike on the spectrum
analyzer. I guess the USRP is actually radiating this frequency. It
might go away if I place the USRP’s antenna farther away from the USRP
itself, which I haven’t been able to try yet.

Does anyone know what the source of this might be within the USRP? I
don’t believe this will cause me any problems. I’m really just curious.

Thanks,

Thomas

2

On Tue, 22 Jan 2008 10:24:53 -0800 (PST), Thomas wrote:

[…]

I positioned the USRP very close to the antenna connected to a real
spectrum analyzer, and I was able to observe the spike on the
spectrum analyzer. I guess the USRP is actually radiating this
frequency. It might go away if I place the USRP’s antenna farther
away from the USRP itself, which I haven’t been able to try yet.

Does anyone know what the source of this might be within the USRP? I
don’t believe this will cause me any problems. I’m really just
curious.

You might be re-discovering the basis of the Van Eck attack but on
the radio receptor itself. That would be quite funny/ironical IMHO :wink:

–
Best Regards,
ChoJin

3

Thomas schrieb:

I positioned the USRP very close to the antenna connected to a real
spectrum analyzer, and I was able to observe the spike on the spectrum
analyzer. I guess the USRP is actually radiating this frequency. It
might go away if I place the USRP’s antenna farther away from the USRP
itself, which I haven’t been able to try yet.

If you can adjust the band so that you can look at 64 Mhz up from the
‘current spike’ to see if there is another, you may be seeing the
digitization clock of the USRP.

In my activities, with an additional LNA in front of the USRP, I see
spikes at say, 512 Mhz, (the start of my region of interest…), and
then depending
on weather, cosmic rays, whatever, I will see spurs on 64 Mhz intervals
upward from there. The top of my range is about 700 Mhz.

Since I’m using the TVRX as well, I may be also seeing some spurs from
that as well.

However, the spurs I see are very narrow, on the order of 2 or 3 bins,
at 4k per bin, and again depending, sometimes the
spurs will ‘move’ a couple of bins one way or the other.

This of course is one of the reasons why I did not go into RF… and
chose ‘software’…

On discussion this with ‘real RF engineers’, I was informed that some
spectrum analyzer companies actually inject
‘noise’ into their displays to cover any ‘spurs’ generated by the
equipment itself. This giving the user the impression
that all that’s there is background noise…

John C…

4

John C. [email protected] wrote: Thomas schrieb:

I positioned the USRP very close to the antenna connected to a real
spectrum analyzer, and I was able to observe the spike on the spectrum
analyzer. I guess the USRP is actually radiating this frequency. It
might go away if I place the USRP’s antenna farther away from the USRP
itself, which I haven’t been able to try yet.

If you can adjust the band so that you can look at 64 Mhz up from the
‘current spike’ to see if there is another, you may be seeing the
digitization clock of the USRP.

In my activities, with an additional LNA in front of the USRP, I see
spikes at say, 512 Mhz, (the start of my region of interest…), and
then depending
on weather, cosmic rays, whatever, I will see spurs on 64 Mhz intervals
upward from there. The top of my range is about 700 Mhz.

Since I’m using the TVRX as well, I may be also seeing some spurs from
that as well.

However, the spurs I see are very narrow, on the order of 2 or 3 bins,
at 4k per bin, and again depending, sometimes the
spurs will ‘move’ a couple of bins one way or the other.

This of course is one of the reasons why I did not go into RF… and
chose ‘software’…

On discussion this with ‘real RF engineers’, I was informed that some
spectrum analyzer companies actually inject
‘noise’ into their displays to cover any ‘spurs’ generated by the
equipment itself. This giving the user the impression
that all that’s there is background noise…

John C…

There are other spikes at 64-MHz intervals. I guess that pretty much
answers my question.

Thanks!

Thomas

5

Hi Thomas

I am in the U.S.

Ok, that’s the problem. In the US the 900 MHz band isn’t used for GSM
traffic it is an ISM band. The relevant bands are GSM-850 and GSM-1900
in the U.S. I’ve updated the corresponding page to give details of the
different bands.

http://273k.net/gsm/find-a-gsm-base-station-manually-using-a-usrp/

Robert