Re: [Discuss-gnuradio] GPS for USRP ideas

[Brian Whitaker]

Krys, Saber,

GPS receivers are a little strange in that the received signal is often
20-40dB below the thermal noise floor (yes, below -174dB/Hz). SNR in the
traditional sense is always negative. Like CDMA, what makes this work is the
spreading gain of the DSSS signal -- the brains of the baseband is able to
reconstruct the signal out from the noise.

Because  of this, GPS front-ends need to offer a couple of key things to
work well:
-> 2-5dB NF from the antenna to the low-IF (usually about 4MHz),
-> ultra-low phase noise, especially at 50-100Hz (as I understand, there is
a lot of information at about 50Hz in the signal -- I need to get a book,
though!)

The 2116/8's are satellite (DBS) receivers, and are pretty nifty in that
they have their own synthesizers and on-chip VCO. An obvious sacrifice in
the VCO being able to tune like crazy is that the tuning gain (Kvco) is very
high -- this means that the phase noise is high. In sat applications, the
BWs are so big that the extra phase noise isn't really an issue... But my
hunch is that it would corrupt the GPS signal to the point that you wouldn't
be able to use it.

Also, every GPS receiver I've seen uses a low-IF -- the receiver brings the
signal down to about 4MHz, where the BB ADC samples it from a very accurate
'GPS clock' (this clock is the low-IF center freq). Not that a GnuRadio GPS
implementation would have to do this, but we'd need to figure out why the
rest of the world does this so understand what we'd be giving up in going to
a direct conversion system.

For more general-purpose applications, a sat tuner like the 211x family, or
an IC TV tuner (Microtune has theirs out already), offers a lot of
flexibility. As soon as I think I can put together a USRP radio board and
use it for work, I'll do it... Until then, work and family will keep me from
committing to a venture like this. I'll try to serve a support role here for
the time being.

Brian. 

On 4/14/04 6:05 PM, "Krzysztof Kamieniecki" <address@hidden> wrote:

> There is also the max2116/max2118 which covers 925 to 2175 MHz. There
> are eval. kits for these parts
> which do not seem to need an external frequency reference.
> http://pdfserv.maxim-ic.com/en/ds/MAX2116EVKIT-MAX2118EVKIT.pdf
> 
> Brian Whitaker, an apps. engineer at MAXIM, was looking into using these
> parts on a USRP daughter
> card but I think he ran low on free time.
> 
> Any comments Brian?
> 
> tebinuma wrote:
> <snip>