Re: [Discuss-gnuradio] GRC and gr_fft_vcc function showing different signal power in dB

[Josh Blum]

On 10/17/2012 05:21 AM, Hemant Saggar wrote:
> Hi all,
> 
> I am new to USRP and I started by trying to calibrate the received power in
> USRP by giving a known sine signal. I gave a signal of *100.2 MHz at -30 dBm
> * to USRP which sampled it at 1MHz then passed it to a 1024 point FFT scope
> in Gnu Radio Companion.  I saw a level of about* -52 dB at 100.2 MHz*.
> 
> I also modified the *usrp_spectrum_sense.py* to find FFT of same signal
> using gr_fft_vcc function and then print the magnitude squared value to a
> csv file. When I took 10*log10(value) the plot showed me about* -30dB level
> at 100.2 MH*z.  Following are my queries
> 
> 1) Is the power level in GRC -52 dB or -52 dBm? Accordingly why is the -22
> dB loss? ( My antenna and channel gain were set to 0dB)
>

The WX gui FFT plotter is scaled for dBfs.

0 dBfs is equivalent 1.0 sample counts, which is fullscale.

> 2) Why are fft squared amplitude in dB different here and which one is
> correct?
> 
See ./gr-blocks/lib/nlog10_ff_impl.cc

This uses the gr_fft_vcc and does various scaling compensation

-josh