Re: [Discuss-gnuradio] Scaling of Noise amplitude (was: Re: Changing Bandwidth USRP)

[Marcus Müller]

Hi!
So three things I care about first:
1. Always ask a question to the mailing list, not single members of that, 
unless it's a direct followup. You're not paying anyone to help you, so it'd be 
only fair that anything you learn here can be learned by everyone.
2. Try to give your mail a meaningful subject. If the issue changes, please 
alter the subject (see the subject of this mail, using (was: XY) is a good idea 
IMHO)
3. Please, at least have an identifiable pseudonym, giving out your real name shouldn't 
hurt you, either. It's good style to offer people a name to talk to, sign your emails if 
they are long or include a request. Just everything you would do in "offline" 
life if you went somewhere asking for some advice. You don't have to be overly polite, 
but really, putting your name below an email is kind of basic etiquette.

So, now to your question:
Of course, your noise values will be multiplied by the same number, nothing 
wrong about that.
Your noise amplitudes, however, might be in the nonlinear region of the receiver. Again, 
this is not inherently wrong, because usually you don't care about thermal noise. You can 
do nothing about the thermal noise introduced at the receiver ADC and analog components, 
you can just crank up the amplification for your desired signal. See "noise 
figure" on wikipedia, if you don't know the concept. To combat nonlinearity, there 
are several approaches, but to be completely frank, I don't think you want to implement 
any of those, because they imply that you do things directly after ADC (so in the FPGA) 
or replace hardware elements (which isn't quite feasible), and I also don't think you 
should care about your noise being represented nonlinear.
Frankly, your questions show that you still need to get accustomed to digital 
signal processing and signal theory first before you should dive into advanced 
topics.

Greetings,
Marcus



> > Hi Marcus,
> >
> >                      Thanks for earlier, but there is a problem. After applying 
> > the multiply_const Operator, the 'noise floor' also changes which does not 
> > represent the real 'floor' of the signal. How can I change that?
> >
> >
> > _____________________________________
> > Sent from http://gnuradio.4.n7.nabble.com

On 10/09/2013 09:27 AM, address@hidden wrote:
> <quote author='Marcus Müller'>
> The factor, however, depends on your ADC/daughterboard combination, on the
> gain you set for the RF frontend, a little bit on temperature, the tuner,
> the impedance matching of your RF input to the Daughterboard etc.  So if you
> already know what you're putting in, just use this as a reference and
> "calibrate" yourself.
>
> Hope that helped clear things up,
>
> Marcus
>
> On 10/04/2013 01:05 PM, Antmrt wrote:
> > Thanks Marcus, my problem is how to calibrate the scale on the FFT plot.
> > e.g:
> > I'm feeding a signal of -30 dBm to the usrp and on the plot the value
> > becomes about -10 dBm