Re: [Discuss-gnuradio] USRP spike

[Vincent W]

On 08/16/2010 11:22 PM, Marcus D. Leech wrote:
>> On Thu, Aug 12, 2010 at 1:56 PM, naveen nischal
>> >> <address@hidden> wrote:
>> >>     
>>>> >>>> problem though is a spurious spike of about 17db which appears at
>>>> >>>> whatever center frequency we tune to in the spectrum. we think that 
>>>> >>>> this
>>>> >>>> might be the problem as that might be jamming the signal which was
>>>> >>>> supposed be about the same db level. The point of notice for us is 
>>>> >>>> that
>>>> >>>> this spike is always there even without the antenna connected.
>>>> >>>>         
>> >>
>> >> Chances are you have a DC offset at the ADC that needs to be removed.
>> >> This will happen if the DDC in the FPGA isn't required to resolve any
>> >> frequency offset due to the limitations of the LO in the RF chain.
>> >>
>> >> One way to mitigate this is to tune a little bit away from your signal
>> >> of interest, then mix your signal of interest to baseband, and filter
>> >> off the DC component.
>> >>
>> >>     
>>> >>> Thanks,
>>> >>> Naveen
>>> >>>       
>> Hi,
>>
>> I think I am encountering a similar problem, and would very much appreciate 
>> your
>> feedback, bearing in mind have an experimental physics background. I'm using 
>> the
>> WBX board on the USRP2, and notice unusual signal spikes when I tune to a
>> specific frequency. In some cases, there are multiple spikes that show 
>> symmetry
>> about a given frequency, but disappear when tuning to that frequency. In 
>> others,
>> there is a single spike that seems to follow the centre signal.

Answering your questions:

> For the 100MHz signals, what does your receive chain look like?

I'm not quite sure what you mean. My initial experiments use the 400-1000MHz log
periodic antenna hooked up to the USRP2, with the u2_rev3-20100603.bin FPGA
image and txrx_wbx_raw_eth_20100608.bin firmware. I see these anomalous signals
even if I write the raw usrp2 output to a file, with no processing, read it in
octave using read_complex_binary, and plot it. I opted to use the usrp2_fft
program for ease of use.

> Keep in mind that the FM radio band exists between 88MHz and 108MHz,
>   so if you're trying to do "off air" experiments, you'll run into
> trouble.

One of the first things I did, and where I noticed this, was when building the
FM receiver. That exercise is what gave me the first inkling that those peaks
aren't FM bands or stations, and, when centred at 100MHz, may not even exist.

>  If you're not doing "off air" experiments, what kind of shielding do you
>  have for your equipment, to protect it from the FM broadcasting band?

Ultimately, one of my goals are to build a low frequency dielectric spectroscope
(DC-2.2GHz for now). My initial target is more modest - I would like to build a
wideband spectrum scanner. To this end, I'm quite interested in the relative
amplitude across a large frequency range. However, it's been very difficult
because I keep having to work around these artifacts when taking the individual
samples that eventually make up the wideband composite plot.

I don't have a Faraday cage or RF anechoic chamber. I tried, unsuccessfully, to
build the former using aluminum foil and wire mesh, but judging from the
wideband spectrum output of the usrp2, it didn't do much. I can show you
pictures of the set up, and the graphs of the spectrum before and after (there's
not much difference).

As evidence for the other kind of spikes, the ones that follow you around,
please see the attached pictures, centered around frequencies of 1134 and
1135MHz. These are the spikes that follow the centre frequency around.

The problem around 100MHz is very similar to what has been previously reported
here: http://lists.gnu.org/archive/html/discuss-gnuradio/2010-03/msg00412.html,
but I'm not sure exactly what the correct resolution is.

Thank you very much for all your help, I look forward to figuring this out.

Vincent

usrp2_1134M_spike.jpeg
Description: JPEG image

usrp2_1135M_spike.jpeg
Description: JPEG image