[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Discuss-gnuradio] FPGA Gain?

From: Eric Cottrell
Subject: Re: [Discuss-gnuradio] FPGA Gain?
Date: Wed, 11 Jul 2007 18:01:01 -0400 (EDT)

----- Start Original Message -----
Sent: Wed, 11 Jul 2007 12:38:41 +0200
From: "Trond Danielsen" <address@hidden>
To: address@hidden
Subject: Re: [Discuss-gnuradio] FPGA Gain?

> 2007/7/10, Eric Cottrell <address@hidden>:
> > Hello,
> >
> > The LFRX daughterboard on the USRP has 20 dB of gain.  I seem to remember 
> > that 10 dB of that gain is from the FPGA.  I assume that this gain is after 
> > the ADC?  If so, is it really useful?
> >
> > I relate digital gain to zoom on a digital camera.  The analog gain is the 
> > optical zoom and the digital gain is the digital zoom.  Digital cameras 
> > have 10x or more digital zoom but it is not really useful.  At high digital 
> > zooms you get lower resolution and the picture is blocky.  Optical zoom is 
> > more useful as it comes before the digitizer.
> >
> > So if I have noise values of 35 and signal values of 40 then I get values 
> > of 350 and 400 after the 10 dB digital amplifier.  I may see a higher 
> > amplitude signal but how do I gain anything?  Am I missing something?
> >
> > 73 Eric
> Hi Eric,
> The ADC has a programmable gain stage prior to sampling. Could it be
> that one you are thinking of?
> Regards,
> -- 
> Trond Danielsen


It must be.  I thought there was a digital amplifier implemented in the FPGA.  
Also I realized that in my example that the problem would be the lsb error 
would grow from +/- 1 bit to +/- 10 bits and the resolution would decrease.  
The values would be the same for analog or digital.

With the oscilloscope and fft programs I can experiment with gain settings and 
see the practical effect.  I tend to set the gain to where the noise floor 
rises up slightly.  I put in a 20 dB amplifier inline to use the AR5000 with 
the LFRX so I could keep the gain at 10.  Now I can try a USRP Gain Setting of 
11 or even 20 to see what happens.  Nigel would be pleased.

To you RF Engineers, Noise Figure and Gain Distribution are common mundae 
design issues.  I am still figuring them out.

73 Eric

reply via email to

[Prev in Thread] Current Thread [Next in Thread]