FPGA Central

Hi,


I would like to knwow if the QPSK Gardner algorithm for symbol
recovery works too for QAM or MPSK signals ?

Thanks

On Wed, 19 Sep 2007 06:44:20 -0700, [email protected]
wrote:

>Hi,
>
>
>I would like to knwow if the QPSK Gardner algorithm for symbol
>recovery works too for QAM or MPSK signals ?
>
>Thanks

As number of levels increase the self-noise of the ted becomes worse.
Gardner is mainly useful for a binary signal but can be used for lower
M levels with not too low SNR.

if gardner can't works for QAM what kind of symbol synchronisation
could we use for QAM ?

On Sep 20, 9:16 am, fpga.vhdl.desig...@gmail.com wrote:
> if gardner can't works for QAM what kind of symbol synchronisation
> could we use for QAM ?

You could try applying a nonlinearity such as abs(), sqrt() or half
bit delay-conjugate multiply to the data. This produces a discrete
component at the symbol rate, which can be bandpass filtered to
provide a symbol clock reference.

John

[email protected] wrote:
> if gardner can't works for QAM what kind of symbol synchronisation
> could we use for QAM ?

Who said it can't work? As long as the excess bandwidth is more than
something like 25%, it works quite well. Problems occur with narrowing
bandwidth, so it certainly can't work for some people.

Where Gardner it won't work various things are possible, You could try
something like cross-correlating the outputs from two narrow bandpass
filters centred at the band edges.

Steve