Re: [Discuss-gnuradio] [USRP-users] Simple correlator to detect synchronization timing.

[Marcus Müller]

Hi Jeon,

is this the mail you were referring to on discuss-gnuradio? I'm including it in my reply, because that seems sensible.

What you're looking at looks very much like a synchronization problem.
Manchester tells me we have a reliable rate of transitions, so that's handy :)
I'd go for oversampling (if that's possible) by a factor of let's say 4 or 5 of your symbol rate, and implementing a sensible filter for the signal; then, throw the Mueller & Müller clock sync (Look into the [Synchronizers] categorie in GRC) at the problem :)

Best regards,
Marcus

On 02/20/2015 03:02 PM, Jeon via USRP-users wrote:

I am trying to make something which:

* finds some synchronization timing position

* throws away what comes before sync code

* passes what comes after sync code

a sync code is '010110' (FYI, It is encoded with Manchester code, so the decoded code is 001)

I have tried:

* digital_correlate_access_code_bb with access code 010110
* fir_filter_xxx with taps of [0,1,0,1,1,0]
* interp_fir_filter_xxx with taps of [0,1,0,1,1,0],

where values of threshold, decimation and interpolation I cannot assure

But, I cannot see value 3 which is expected
if there is perfect timing synchronization for a given sample stream ......0101011001......

Am I using wrong blocks, or somewhat misconfigured threshodl, decimation and interpolation values?

If you have an idea for it, please give me some advice.

Regards,

Jeon.

_______________________________________________
USRP-users mailing list
address@hidden
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com