Re: [Discuss-gnuradio] tunnel.py distance problem

[Johnathan Corgan]

On 11/7/07, Marko Simunac <address@hidden> wrote:

> ...and am assigning a carrier treshold of 90(-c 90) to get rid the BBBBBBs. 
> Can someone help me to increase the distance between the two USRPs, and 
> theoretically what is the max distance between the two USRPs that tunnel can 
> work?

This carrier threshold is probably too high.

There is a sweet spot for this parameter, only determined by
experiment.  If it is too low, then the nodes won't transmit due to it
thinking background noise is someone transmitting (those are the BBBs
you're getting rid of.)

If it is too high, the nodes will "ignore" anyone else transmitting,
resulting in continuous collisions, and no actual data flow.

In my experience, a properly tuned carrier threshold results in one or
two Bs per transmitted packet on each side of a two node network, when
sending full duplex traffic.

The other knob to adjust is the receiver gain.  Depending on your
antenna, you need to adjust this for maximum signal to noise ratio.
But changing your gain changes your required carrier sense threshold
as well.

As Eric mentioned, a good way to judge this is to run the
benchmark_tx.py script on one node, then use the usrp_fft.py script on
the other to see the transmitted waveform.  You can adjust the gain
slider up and down to get the widest spread between the noise floor
and the received signal peak.  Then use this same gain in the
tunnel.py script.  Now you can adjust the carrier sense threshold to
be just high enough to be able to transmit with nothing being
transmitted from the other end.

All this /should/ be automated...

-- 
Johnathan Corgan
Corgan Enterprises LLC
http://corganenterprises.com/