Forum: GNU Radio Simple QAM mod/demod

Posted by "Fabián Kozynski" <fabikw@gmail.com> (Guest)
on 2012-10-01 22:28
(Received via mailing list)
As sanity check I'm trying a simple flowgraph in GRC with a QAM 
mod/demod:

Vector Source ==> Throttle ==> QAM Mod ==> QAM Demod ==> Unpacked to 
Packed
==> File Sink

The parameters in the mod/demod are as default. I also have a File Sink
before the mod.

In the output I see first a bunch of zeros, then a bunch of random 
numbers
and after 20 or so bytes I get a sequence with the same period as the
original one, but different numbers. I suppose this is a problem of 
phase
sync.

What's the right way to simulate a simple mod/demod? How do I add sync
between mod and demod?

Thanks,
Fabin
Posted by Ben Reynwar (Guest)
on 2012-10-03 06:24
(Received via mailing list)
On Mon, Oct 1, 2012 at 1:27 PM, Fabin Kozynski <fabikw@gmail.com> wrote:
> original one, but different numbers. I suppose this is a problem of phase
> sync.
>
> What's the right way to simulate a simple mod/demod? How do I add sync
> between mod and demod?
>
> Thanks,
> Fabin

It sounds like you're comparing bytes in the two files.  First thing
that I'd check is that the two files aren't offset from each other by
a few bits so that the bytes don't line up in the same way.

If that's not the problem then email me the grc file and I'll have a 
look.

Cheers,
Ben
Posted by "Fabián Kozynski" <fabikw@gmail.com> (Guest)
on 2012-10-03 16:13
Attachment: QAMMod.grc (10,5 KB)
(Received via mailing list)
It's not just an offset problem. Even though in the output file, there 
is a
delayed sequence, it doesn't have the right bits. As it has exactly the
same period it leads me to believe that it's just a problem of phase 
sync.

Attached is a minimal grc file.

Cheers,
Fabin
Posted by "Fabián Kozynski" <fabikw@gmail.com> (Guest)
on 2012-10-03 19:33
(Received via mailing list)
I figured out the problem. There were two things I wasn't taking into
account:


   - The modulator reads bits in MSB order, which means that I should 
pack
   them the same way.
   - There is a bit framing problem (can be solved using packet
   encoder/decoder). Checked converting both sequences to bits (8 per 
number)
   and realizing one's a shifted version of the other.

Cheers,
Fabin

On Wed, Oct 3, 2012 at 10:12 AM, Fabin Kozynski <fabikw@gmail.com> 
wrote:

It's not just an offset problem. Even though in the output file, there 
is a
Please log in before posting. Registration is free and takes only a minute.
Existing account (Switch to SSL-encrypted connection)
NEW: Do you have a Google/GoogleMail or Yahoo account? No registration required!
Log in with Google account | Log in with Yahoo account
No account? Register here.