Re: [Discuss-gnuradio] trouble with gr_multiply_const_ff

[Ian Buckley]

Lee, 
What Marcus and Tom are telling you is that when a signal reaches a USRP Sink 
in GNURadio it is transformed into an integer representation that is weighted 
so that the integers represent values -1 <= x < 1.

Input values with magnitudes larger than these will be saturated to these 
limits, which of course alters (destroys) the signals spectrum. 

Ideally your flow graph should try to deliver signals to a USRP sink that have 
been scaled so that they are slightly within these constraints to maximize the 
dynamic range offered by this (16Bit) integer representation, hence Marcus's 
suggestion that you target 0.8 as a maximum amplitude of your signal.

The various USRP daughter card products have different available programable 
analog gain ranges for the TX amplifiers, and these are controlled by the Gain 
parameter of the USRP sink. Generally the highest gain values will completely 
saturate the analog output stages of the Daughter card, given a full range 
digital input signal. You will need to empirically determine a gain value that 
obtains the maximum rated power from the daughter card, given your input 
signal, without driving the output stage into saturation.

-Ian

On Dec 1, 2013, at 8:24 PM, atools_cook <address@hidden> wrote:

> Marcus，
> 
> Thanks for your advice and we do use the "very old" code exactly.
> 
> Would you please elaborate the "backing things off" things because I can't
> your point of it.
> 
> Best
> Lee
> 
> 
> 
> --
> View this message in context: 
> http://gnuradio.4.n7.nabble.com/trouble-with-gr-multiply-const-ff-tp45018p45039.html
> Sent from the GnuRadio mailing list archive at Nabble.com.
> 
> _______________________________________________
> Discuss-gnuradio mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio