Hey Brian,
Running through a descrambler and then trying to correlate with 128 ones, I
get the following for the 11 packet trace:
http://cyprus.cmcl.cs.cmu.edu/~gnychis/mfilter/descrambled.png
There are clearly 11 spikes, but the closest we get in a window of 128 bits,
is roughly 82 ones. Here are the descrambled bits when closest to a
correlation of 128 ones:
0 1 1 1 0 1 1 0 0 1 1
1 0 1 1 0
1 1 1 1 1 1 1 1 1 0 1
1 1 0 1 1
0 1 1 1 1 1 1 1 1 1 1
0 1 1 1 0
1 1 0 1 1 1 1 1 1 1 1
1 0 1 1 1
0 1 1 0 1 1 1 1 1 1 1
1 1 1 0 1
1 1 0 1 1 0 1 1 1 1 1
1 1 1 1 1
0 1 1 1 0 1 1 0 1 1 1
1 1 1 1 1
1 0 1 1 1 0 1 1 0 1 1
1 1 1 1 1
Timing/frequency error? If I use these bits as my preamble (instead of 128
ones) and pass it through our full modulator (scramble -> DBPSK -> barker),
we correlate very nicely with all 11 transmissions:
http://cyprus.cmcl.cs.cmu.edu/~gnychis/mfilter/remodulated.png
- George
On Mon, Dec 7, 2009 at 3:28 PM, Brian Padalino <[email protected]> wrote:
> On Mon, Dec 7, 2009 at 12:44 PM, George Nychis <[email protected]> wrote:
> > I double checked the scrambler code and I really do not see anything
> wrong
> > with it, it seems to meet the 802.11 spec for DSSS. Then I found
> something
> > in the spec that says that the transmitter need not start with the seed,
> > since the receiver's descrambler is self-synchronizing. So then I
> figured,
> > maybe the transmitter just started with a different seed... so I
> generated
> > 128 different signatures, for all 128 different possible scrambler seeds,
> > and NONE correlate.
> >
> > Maybe it's using a different scrambling algorithm? I'm running out of
> > ideas.
>
> It might transmit the opposite of what it is expecting to flush the
> receiver scrambler and get it synchronized. So if you need a long
> length of 1's, maybe it starts sending a couple symbols worth of 0's
> first - gets synchronized, then starts the sending of 1's. Same with
> the 0's for the preamble.
>
> Maybe despread your symbols and run through the derandomizer, then
> check to look at the pattern. You should see your long string of 1's
> or 0's (whatever you're looking for). Then check to see what they
> send beforehand?
>
> Brian
>
_______________________________________________
Discuss-gnuradio mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
