For example, we can state when the temporal distance between
the updated timing error estimates is less than T/16, then the procedure terminates. A block diagram
depicting this procedure is shown in Fig. 6.29.
We have used ???circles??? and ???boxes??? to denote the time instances used in the interpolation. The timing
recovery begins with the circles and then updates the timing to produce the boxes. This procedure
is shown in Fig. 6.30 using the eye diagram.
In this figure, we show the iterative procedure operating on the eye diagram directly. Let us assume
the initial conditions are valid at time, t 0 with T/2 sampling creating two samples, A0 and B0.
The ML metric will select B0 as the better choice of the two. Next we interpolate near this timing
phase creating A1 and B1. The interpolation distance is a design parameter and chosen to be T/8 for
this example. As discussed above, the desired resolution should be obtained through system simulations.
The updated ML metric is recomputed and timing-phase instance A1 is chosen at t 1. Next
another iteration is performed near A1 to create A2 and B2. Once again a new ML metric is computed
and A2 is chosen. As we can see each iteration will bring the estimated timing closer and closer to the
desired point.
Pages:
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535