Timing recovery and Carrier recovery

There are many ways to implement a "digital receiver", and there are many definitions of that term. You need to be more specific. What kind of modulation are you dealing with?

Carrier recovery does not bandlimit the signal, it removes residual frequency and phase offset so the constellation stabilizes. Timing recovery is a synchronous decimation process where the signal is downsampled to one sample per baud, and that sample is taken at the instant where the SNR is maximized, ie at the the peak time of the eye pattern. This normally involves interpolation between adjacent samples.

Timing recovery can be and often is performed before carrier recovery, because it is often possible to determine the ideal sampling instants in the presense of a residual carrier.

There are many references on this. Try searching on google.

John

Hi Steve,

Thank you for your help :-) and I really find it a little weird.

I searched "Gardner timing recovery" in scholar.google.com and find "Interpolation in digital modems-Part I: Fundamentals". Is it the paper talk about timing recovery without having tied down the carrier?

And can you recommend a book talk about "digital receiver" thoroughly? Thanks!

Best regards, Davy

Keep studying.

I think carrier recovery should be done in front of

No. The carrier recovery should be done jointly with the timing recovery.

There is a special newsgroop: help.moro

That's a great paper to read, and so is Part II.

Hi Davy,

Carrier recovery, or tracking, does not necessarily mean eliminating the carrier, although that is generally the goal. It means working out exactly what the carrier is - its precise frequency and phase - and usually creating a precise copy of the original.

Timing recovery means working out the exact positions of the symbols within the signal stream. Various algorithms allow this to be done ahead of precise carrier recovery. This means a lot of digital receivers look something like:

use the carrier to get to baseband -> perform timing recovery -> equalise the signal -> perform carrier tracking and feed that back to tune the first stage -> demodulate the symbols

while others look like:

use a rough approximation of the carrier to get to baseband -> perform timing recovery -> equalise the signal -> perform carrier tracking and remove the last traces of the carrier from the signal -> demodulate the symbols

The first time you approach this it might seem a little weird that you can track the symbol timing without having tied down the carrier precisely. However, a number of really simple algorithms - e.g. Gardner

- allow this.

Regards, Steve

Davy wrote:

...

...

You often have useful tidbits to impart. When you have nothing useful to say, it is best to just say nothing. Regularly sneering at those who know less than you gives the impression that you're insecure.

Jerry

When I will need your advice, I will certainly ask for it.

VLV

At comp.dsp, we give advice gratis.

Jerry

It looks like cheap advice very often.

VLV

I agree. It also tends to decrease the signal to noise of the newsgroup.

Sometimes too gratis ;)

Here, one gets at least what one pays for! :-)

Jerry

Vladimir Vassilevsky writes: [MUCH]

Dear Sir:

If you posted with a valid email address, I would have my say privately.

Please "wake up and smell the roses".

You may have technical skills, but your interpersonal skills resemble mine in the midst of the last century ( PS that aint no complement ).

My email is valid if you would like a *PRIVATE* discussion.

Richard Owlett Springfield, MO USA

I tried. VLV's interpersonal skills

I have to suggest that you may have wonderful interpersonal skills but your technical skills are likely not to improve since the midst of the last century. Did you mean throwing some kind of challenge or it is all just the idle talk?

VLV

Randy,

I wish I had put it as well as you. Thank you.

Jerry

Vladimir,

I think your knowledge in DSP and communications is extremely extensive and impressive.

However, ... regarding your posture towards some posters, I think you sometimes make assumptions that may not be true and/or aren't empathetic to the posters' place in their DSP/communications journey.

In my opinion, this forum is for ALL who seek information on this topic, not just the most advanced people. You too, at some point, had to ask basic questions just like some of the folks here have asked recently.

Regarding the comment you made to tommy about 32-bit division, I think you were way out of line. You have no idea who this person is. It could be a 5th-grader, for all you know, who is trying to learn how to program. Try not to make assumptions, because not only does it reflect poorly on you when you are wrong, but it also can damage the heart of a person, and that is a despicable thing to do.