Clock Recovery MM: Difference between revisions

From GNU Radio
Jump to navigation Jump to search
No edit summary
(Note symbol_sync instead, since Deprecated in 3.9)
 
(One intermediate revision by one other user not shown)
Line 1: Line 1:
[[Category:Block Docs]]
[[Category:Block Docs]]
This block is meant to act as a clock recovery, to synchronize to a signal's frequency and phase, so that symbols can be extracted.   
This block is meant to act as a clock recovery, to synchronize to a signal's frequency and phase, so that symbols can be extracted.   
<br><code>Deprecated in 3.9</code> (Favor [[Symbol_Sync]] instead.)


Specifically, this implements the Mueller and Mueller (M&M) discrete-time error-tracking synchronizer.  The peak to peak input signal amplitude must be symmetrical about zero, as the M&M timing error detector (TED) is a decision directed TED, and this block uses a symbol decision slicer referenced at zero.  The input signal peak amplitude should be controlled to a consistent level (e.g. +/- 1.0) before this block to achieve consistent results for given gain settings; as the TED's output error signal is directly affected by the input amplitude.  The input signal must have peaks in order for the TED to output a correct error signal. If the input signal pulses do not have peaks (e.g. rectangular pulses) the input signal should be conditioned with a matched pulse filter or other appropriate filter to peak the input pulses. For a rectangular base pulse that is N samples wide, the matched filter taps would be [1.0/float(N)]*N, or in other words a moving average over N samples. This block will output samples at a rate of one sample per recovered symbol, and is thus not outputting at a constant rate.  Output symbols are not a subset of input, but may be interpolated.   
Specifically, this implements the Mueller and Mueller (M&M) discrete-time error-tracking synchronizer.  The peak to peak input signal amplitude must be symmetrical about zero, as the M&M timing error detector (TED) is a decision directed TED, and this block uses a symbol decision slicer referenced at zero.  The input signal peak amplitude should be controlled to a consistent level (e.g. +/- 1.0) before this block to achieve consistent results for given gain settings; as the TED's output error signal is directly affected by the input amplitude.  The input signal must have peaks in order for the TED to output a correct error signal. If the input signal pulses do not have peaks (e.g. rectangular pulses) the input signal should be conditioned with a matched pulse filter or other appropriate filter to peak the input pulses. For a rectangular base pulse that is N samples wide, the matched filter taps would be [1.0/float(N)]*N, or in other words a moving average over N samples. This block will output samples at a rate of one sample per recovered symbol, and is thus not outputting at a constant rate.  Output symbols are not a subset of input, but may be interpolated.   

Latest revision as of 20:31, 24 March 2021

This block is meant to act as a clock recovery, to synchronize to a signal's frequency and phase, so that symbols can be extracted.
Deprecated in 3.9 (Favor Symbol_Sync instead.)

Specifically, this implements the Mueller and Mueller (M&M) discrete-time error-tracking synchronizer. The peak to peak input signal amplitude must be symmetrical about zero, as the M&M timing error detector (TED) is a decision directed TED, and this block uses a symbol decision slicer referenced at zero. The input signal peak amplitude should be controlled to a consistent level (e.g. +/- 1.0) before this block to achieve consistent results for given gain settings; as the TED's output error signal is directly affected by the input amplitude. The input signal must have peaks in order for the TED to output a correct error signal. If the input signal pulses do not have peaks (e.g. rectangular pulses) the input signal should be conditioned with a matched pulse filter or other appropriate filter to peak the input pulses. For a rectangular base pulse that is N samples wide, the matched filter taps would be [1.0/float(N)]*N, or in other words a moving average over N samples. This block will output samples at a rate of one sample per recovered symbol, and is thus not outputting at a constant rate. Output symbols are not a subset of input, but may be interpolated.

The complex version here is based on: Modified Mueller and Muller clock recovery circuit: G. R. Danesfahani, T.G. Jeans, "Optimisation of modified Mueller and Muller algorithm," Electronics Letters, Vol. 31, no. 13, 22 June 1995, pp. 1032 - 1033.

For more info see [1]

Parameters

(R): Run-time adjustable

Omega (R)
Initial estimate of samples per symbol
Gain Omega (R)
Gain setting for omega update loop
Mu (R)
Initial estimate of phase of sample
Gain Mu (R)
Gain setting for mu update loop
Omega Relative Limit
Limit on omega

Example Flowgraph

Source Files