Re: [Discuss-gnuradio] Why is Barker Spreading difficult to usrp?

[Greg Troxel]

DiX <address@hidden> writes:

>        With BBN codes and FLEX2400 d'borads, now I can received 802.11b
> packets from standard wireless cards, and communicate between two USRPs with
> barker spreading off.  These make me really excited  :) 

That's good to hear - that matches where we got to.

>        However, I failed trying to send any packet from USRP to PC.  I
> believe it is because the barker spreading of thej tx path was not working
> well and PC could not interpret the packets correctly. I also sent packets
> between USRP with barker on. That did not work too.  

I am not sure if we ever thought that we had the barker transmit working.

>        It seems barker demodulator works but modulator does not. Why is the
> barker spreading difficult for usrp? Is it because the 1MSps symbol rate is
> too tough for the USRP board?  ( USRP is not that fast to deal with 1
> million 11-bit batter sequence?)

It's because the bandwidth of the signal from 1 Msample/s times 11
chip/symbol is 11 Mchip/s.  I think we used 8-bit i/q and 8 Msample/s to
get 16 MB/s across the USB.  For the barker receiver, the code computes
pre-distorted waveforms of the barker code.  I don't know that the
transmitter does, but really I think it just has to hope for the best in
terms of the receiver receiving a distorted signal.

You may be able to up the rate to 32 MB/s (16 Msample/sec) and tx may
work better.

Or, you could implement the barker code in the fpga.

The BBN code is assigned to FSF, so you are welcome to integrate it into
GNU Radio proper.