Re: [Discuss-gnuradio] OFDM packets

[Tom Rondeau]

Crespi Floriana wrote:
> Scrive Eric Blossom <address@hidden>:
>
>   
> > On Thu, Feb 28, 2008 at 12:51:18PM +0100, Crespi Floriana wrote:
> >     
> > > Hi,
> > > I have a problem with OFDM system.
> > > I transmit OFDM signal at 2.4 GHz, but I receive half of the total packets.
> > > Moreover, only half of the packets received are correct.
> > > CRC controller is perfect but i want a ber estimate, so I change the
> > > payload (binary file) and I  receive check bit by bit.
> > > The result of my alteration test is similar than CRC test.
> > > I have a lot of frequency offset, therefore I change manually the frequency
> > > receiver.
> > > I use revision 6845 and RFX 2400.
> > > Example with crc:
> > > The final packet I receive:
> > > ok: False      pktno 4602   n_rcvd 1300       n_right 83
> > >
> > > Why can't i get all the packets?
> > > Could you help me??
> > >
> > > I'm testing OFDM system without USRP and none of the packets are lost.
> > >
> > > Thanks a lot.
> > >
> > >       
> > Hi Crespi,
> >
> > You have neglected to tell us virtually anything about your problem.
> > Could you please tell us:
> >
> >   which version of GNU Radio?  trunk, tarball (which one), etc?
> >   what operating system, distribution and architecture?
> >   the command line you issued
> >   any modifications you may have made to distributed code?
> >   anything else that might be relevant
> >
> >
> > Thanks,
> > Eric
> >
> >
> >     
>
>
>
> Hi, thank you for your reply.
> I use:
> * Linux K-Ubuntu 7.04
> *Gnuradio Trunk 6845
> The command are :
> --TX--
> sudo python benchmark_ofdm_tx.py -f 2.442e9 -T A --tx-amplitude 10000
> --RX--> sudo python benchmark_ofdm_rx.py -f 2.4419899e9 -T A
> (occuped tones, size packet and fft are default value)
>   

That's a very high amplitude. The PAPR of OFDM will hit you at this 
amplitude. Stay between 1000 and 4000.

> Important: If I don't change Rx frequency it can't receive!!
>   

OFDM is notoriously sensitive to frequency offset. With the RFX2400 
boards, the frequency deviation is likely just outside the range we allow.

Tom


> I change the payload. I upload file to the code and I calculate its lenght.
> Number packets= (lenght file/ size packets)+1
> Before my modification I tested ofdm system and the result is similar then after
> modification.
> The trial simulation (without USRP and antennas) is perfect. I Receive all
> packets and they are correct.
> I tested with coax for estimate channel problem but the result is similar to
> real channel. I thought the problem was the channel but this experiment reject
> my theory.
> I don't find the problem because my alteration code don't change the core code.
>
>
> This is my result. final packet
> ===================================================================================
> Trial simulation
> ok: True         pktno: 4609     n_rcvd: 4610    n_right: 4610
>
> bit right 14750243
> bit wrong  0.0
> ===================================================================================
> All correct
>
> ===================================================================================
> Coax
>
> ok: False        pktno: 4000     n_rcvd: 1300    n_right: 830
>
> bit right 2109817
> bit wrong 2050183
> ===================================================================================
>
>
>
>
> Thank you a lot for your help.
> _________________________________________________________________
>
>
>
>
> _______________________________________________
> Discuss-gnuradio mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>
>