Re: Problem with the PLL Carrier Tracking Block

[George Edwards]

Hi Daniel,

Thank you very much! 

Now, I can stop spinning my wheels. I have used the Costas loop and know that it works well, I was just trying to add a new tool to my quiver thinking that it functions in a similar manner, thanks for the clarification on how it works.

Regards,

George

On Wed, Mar 17, 2021 at 4:24 PM Daniel Estévez <daniel@destevez.net> wrote:

El 16/3/21 a las 19:17, George Edwards escribió:
> Hello GNURadio Community,
>
> I am having trouble getting the PLL Carrier Tracking Block to work
> properly. I designed the GRC with a qpsk constellation modulator at a
> sample rate of 40 k Hz. I designed the PLL Carrier Track block to
> correct incoming waveforms that are off centered from DC by +/-500 Hz
> (setting max/min freq with the formula +/-2*pi*f(=500)/samp_rate). The
> eye test at the PLL outputs seems to show that the input signal freq is
> recentered around DC, however, the signal level coming out of the PLL is
> greatly reduced and appears to be smeared. So obviously, when I follow
> this block with a Symbol Sync block it does not lock and the
> constellation points continue to rotate.

Hi George,

The PLL Carrier Tracking block expects an unmodulated carrier at its
input. It won't work with a QPSK signal. For a QPSK signal you need to
use the Costas Loop block.

In more detail: The PLL Carrier Tracking block uses what is sometimes
called as a pure (or four-quadrant arctan()) discriminator, which only
locks to an unmodulated carrier. You would use that for residual carrier
signals. For a usual suppressed-carrier QPSK signal you should use a PLL
with a Costas discriminator (of order 4, in the case of QPSK). In GNU
Radio, this is implemented by the Costas Loop block.

Best,

Dani.