Re: [Discuss-gnuradio] Frequency Offsets in RFX 2400

[Eric Blossom]

On Wed, Jan 09, 2008 at 11:02:47AM -0800, Shyamnath wrote:
> 
> Hi,
> I am trying to run a few experiments to transmit a data  stream from a
> transmitter to receiver using a RFX 2400.  However, I am observing that the
> frequency offset is extremely high at 40-80kHz. In addition, it also changes
> significantly, nearly 2-8kHz, every time I restart the usrps.  I have used
> various ways of estimating the frequency offset and all of them converge on
> the above observations.
> 
> So I have a couple of questions
> 1) Is this high a frequency offset in the specs? 
> 2) Is this variation in the frequency offset  in the specs of these radio?
>  (I have been working on other costume made hardware and the frequency
> offset has been almost constant for the past couple of months)
> 3) Is there are way to prevent the frequency offset variation after every
> reboot? 
> 
> Thanks in advance for any replies
> Shyam

I'm not sure of the tolerance of the oscillators Matt's currently
using on the 2400's, but 40k/2.4G is about 17 ppm.  I wouldn't call
that "extremely high".  In general any receiver has to be able to
track out any frequency offset and symbol timing variation.  Expensive
test equipment has better specs (and costs more) because they use
things like ovenized crystal oscillators.

All xtal oscillators move around with temperature.  The question is
how much, and how much money is being spent on compensation and
environmental control.

Eric