USRP2 noise floor peculiarities

[Jordan J Riggs]

Folks,

I have been trying to get a better understanding the noise characteristics of the USRP2/RFX2400 platform. To get an idea of the noise floor at various bandwidths, I connected the SMA's on the RFX2400 to a 50ohm load and averaged out the FFT at various bandwidths. Naturally, I would expect the floor to drop at lower bandwidths (higher decimation), but this is not always the case. For example, the floor at decim 100 is slightly lower than it is at 128. This is counterintuitive. The only explanation I can come up with is some filtering quirk in the FPGA. Does anyone have any ideas?

I've attached screenshots at decim 100 and 128.

Thanks,
Jordan


_______________________________________________
Discuss-gnuradio mailing list
Discuss-gnuradio@...
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio

[Douglas Geiger-2]

I believe the filtering used does change with the decimation - per an
earlier posting by Matt Ettus: if the decimation is a factor of four
there are two half-band filters + CIC filter, if it is a multiple of
two (i.e. even, but not a factor of four) then there is only one
half-band filter + CIC, and if it is an odd number, just a CIC filter.
 Doug

On Mon, Oct 26, 2009 at 4:20 PM, Jordan J Riggs <jjriggs06@...> wrote:

> Folks,
>
> I have been trying to get a better understanding the noise characteristics
> of the USRP2/RFX2400 platform. To get an idea of the noise floor at various
> bandwidths, I connected the SMA's on the RFX2400 to a 50ohm load and
> averaged out the FFT at various bandwidths. Naturally, I would expect the
> floor to drop at lower bandwidths (higher decimation), but this is not
> always the case. For example, the floor at decim 100 is slightly lower than
> it is at 128. This is counterintuitive. The only explanation I can come up
> with is some filtering quirk in the FPGA. Does anyone have any ideas?
>
> I've attached screenshots at decim 100 and 128.
>
> Thanks,
> Jordan
> _______________________________________________
> Discuss-gnuradio mailing list
> Discuss-gnuradio@...
> http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>
>

--
Doug Geiger
doug.geiger@...


_______________________________________________
Discuss-gnuradio mailing list
Discuss-gnuradio@...
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio

[RH200]

Hi we to have been getting interesting stuff at low power levels. Some one is trying to implement and RFI scanning system using a small dish to scan for RFI between 800 MHz and 1800 MHz. Using the USRP2 and DBRX and its a bit hard to tell the difference at times between the birdies in the system and RFI.

Also does anyone have a reliable estimate of the noise tempreture for the USRP2 DBRX combination?

Attached is three images each image has two plots one is a linear plot the other is a logarithmic scale. One plot is the dish and LNA coupled with the log perodic antenna. The other plot is the LNA terminated, and the last one is the USRP2 terminated. The plots are just lots of little data captures stiched together.rfi_logper_20091020b.pngrfi_termamp_20091020.pngrfi_termcoax_20091020.png

Jordan J Riggs wrote:

Folks,

I have been trying to get a better understanding the noise characteristics
of the USRP2/RFX2400 platform. To get an idea of the noise floor at various
bandwidths, I connected the SMA's on the RFX2400 to a 50ohm load and
averaged out the FFT at various bandwidths. Naturally, I would expect the
floor to drop at lower bandwidths (higher decimation), but this is not
always the case. For example, the floor at decim 100 is slightly lower than
it is at 128. This is counterintuitive. The only explanation I can come up
with is some filtering quirk in the FPGA. Does anyone have any ideas?

I've attached screenshots at decim 100 and 128.

Thanks,
Jordan

 
 
_______________________________________________
Discuss-gnuradio mailing list
[e-mail]
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio