Re: [Discuss-gnuradio] Number sink unit to dBm

[Colby Boyer]

Marcus,

Speaking on the topic of calibration. Has anyone characterized the
performance of the Ettus daughter cards, i.e. noise floor and
freq/volt measurements to card output?

--Colby

On Thu, Apr 28, 2011 at 6:41 AM, Marcus D. Leech <address@hidden> wrote:
> On 28/04/2011 7:38 AM, Patrik Tast wrote:
>
> Hi All,
>
> How do I convert the value from a number sink to dBm?
> GRC attempt
> http://www.poes-weather.com/~patrik/1.7GHz/HRPT/April-27th-2011/Screenshot-GRC.png
> Output
> http://www.poes-weather.com/~patrik/1.7GHz/HRPT/April-27th-2011/Screenshot-7.png
>
>
> Regards,
> Patrik
>
> First, received signal *power* is proportional to the square of the received
> *voltage*, and the received *voltage* (instantaneous) is
>   what comes out of a USRP source block.
>
> I usually feed into a complex-to-mag-squared block, followed by a
> single-pole IIR filter, then I decimate it with a keep-one-in-N block to
> reduce
>   the data rate.    Now, after this, you have an unscaled estimate of the
> signal strength across whatever bandwidth is "seen" by the
>   complex-to-mag-squared block.  If you want to scale it into dBm, then run
> it into a log10 block, and set 'n' to 10, and 'k' to whatever
>   calibration constant you have determined experimentally will map your
> power estimates into actual received dBm.
>
> Here's the thing.  None of the hardware involved here is intended/designed
> to be a precision measuring instrument.  So you have to calibrate
>   according to your own local setup so that you get dBm numbers that make
> sense.  Those calibration constants can, and usually will, change
>   with frequency, since most garden-variety amplifiers, mixers, etc, aren't
> perfectly flat across their operating frequency.
>
>
>
> _______________________________________________
> Discuss-gnuradio mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>
>