Multiplex communications – Communication techniques for information carried in plural... – Combining or distributing information via time channels
Reexamination Certificate
1998-11-20
2002-10-22
Ton, Dang (Department: 2661)
Multiplex communications
Communication techniques for information carried in plural...
Combining or distributing information via time channels
C370S516000
Reexamination Certificate
active
06470033
ABSTRACT:
TECHNICAL AREA
The invention concerns a stuffing procedure for plesiochronous data transmission in connection with which:
a) stuff bits are input into an elastic memory on the basis of a clock pulse and
b) are retrieved on the basis of a second clock signal which runs asynchronously with the first, whereby during reading out
c) they are inserted at the end of a data frame as needed, and to be sure based upon a phase difference measured between the clock signals at least exceeding or falling short of a specified threshold.
Furthermore, the invention concerns a data transmission system and a terminal for implementing the procedure.
STATE OF THE ART
If two asynchronized pulse ranges in a data transmission system follow each other with approximately equal pulse rates, an elastic memory (FIFO memory) should be provided at the interface into which data should be input at the one rate and read out at the other. In practice, the frequencies of the two clock pulses deviate by some ppm as a rule. In order to assure that the elastic memory cannot be overfilled, the frequency of the retrieving clock pulse is selected slightly but sufficiently greater than that of the storing signal. Correspondingly, now and then stuffing bits are inserted into the reading out data flow. The insertion takes place (if need be) at the end of a data frame in any given case. This consequently leads to the length of the data frame (or in connection with firmly specified frame length, the number of useful bits) varying. If now a clock pulse is derived from the frame length (or the block length of the useful bits) in a subsequent component of the transmission system, then a low frequency “wander” or phase jitter occurs.
It is known that the jitter amplitudes can be reduced by a suitable choice of the stuffing procedure. It is known from the article “Measured pulse-stuffing jitter in asynchronous DS-1/sonet multiplexing with and without stuff-threshold modulation circuit, ” by W. D. Grover et al. Electronics Letters, Aug. 27, 1987, 23(18): 959-961, for example, that a stuff threshold modulation with a sawtooth curve can deliver better results than a constant threshold value.
REPRESENTATION OF THE INVENTION
The object of the present invention is to present a procedure of the type mentioned earlier that permits a reduction of jitter amplitudes which surpass the current state of the art.
The solution in accordance with the invention is defined by the characteristics of claim
1
. Accordingly, a difference frequency existing between the clock pulses (for example, in the form of a ppm deviation) is measured, and an appropriately predefined modulation curve is set as the threshold as a function of which of several specified (difference frequency) segments it lies. The (threshold value) modulation curve can also vary from segment to segment. In accordance with the invention, at least two different segments with different modulation curves are specified.
Preferably a modulation curve optimized with respect to minimal jitter amplitudes is established in each segment. It has been shown that it is much simpler to avoid strong amplitude maximums for only a section of the overall maximum difference frequency allowable than for the entire area. There are, for example, modulation curves which permit high amplitudes to arise directly alongside the null point (which is defined in that the stuff rate &rgr;=0.5). Within the framework of the invention, for example, the region adjacent to the null point can be defined as an independent segment which operates with its own modulation curve. That is, outside the segment mentioned, a modulation curve “A” could be used, while within the same, another modulation curve “B” is adduced.
Preferably the segments vary in size. This must not, however, apply for all. Some can be of equal size without further ado.
For creating a numerically optimized modulation curve for a specific segment (or rather generally of a specified area), the invention proceeds as follows:
a) determination of an area of the stuff rate &rgr; (for example, &rgr;e∈[0.53, 0.57]) corresponding to the selected segment, that is, the limiting value of the admissible &rgr; area;
b) fixing a desired number N of supporting values T
O,
. . . , T
N-1
and a numerical resolution ∈ for each supporting value;
c) establishing a desired maximal index K
max
and determination of all index pairs (K, m) for which the stuff rate &rgr;=m/KN lies in the admissible (in accordance with step a) area of the stuff rate &rgr; (m=0, 1, 2 . . . K=1, 2, 3);
d) determination of those support values T
O
, . . . , T
N-1
, which lead to the smallest values of the greatest amplitudes D
K,m
. The amplitudes mentioned are defined here as follows:
D
K
,
m
=
1
π
·
K
(
d
K
,
m
r
)
2
+
(
d
K
,
m
1
)
2
whereby
:
d
K
,
m
r
=
1
N
·
∑
n
=
0
N
-
1
cos
[
2
·
π
(
K
·
T
n
-
m
·
n
N
)
]
d
K
,
m
1
=
1
N
·
∑
n
=
0
N
-
1
sin
[
2
·
π
(
K
·
T
n
-
m
·
n
N
)
]
In this way, a numerically optimal modulation curve can be calculated for each segment (or any arbitrary ppm area). It is evident that this optimization must be carried out in advance. The modulation curves determined are stored in the form of truth tables in a memory in the data transmission system. The stuffing procedure then simply selects the right table and reads out the values varying from data frame to data frame.
It is apparent that the preferred numerical optimizing method basically can be used independently of the segmentation of the difference frequency area.
Preferably, the periodicity is N≦10. Good results can be obtained with 5≦N≦7. The maximal value of the (whole number) index K is preferably not greater that 2
5
. Specifically, it is clear that the amplitudes D
K,m
are proportional to 1/K. The maximal amplitude values to be minimized are present chiefly in connection with small index values.
The resolution ∈ should be better than ⅕. Good results were obtained with, for example, a resolution of {fraction (1/20)}. With an all too fine resolution (for example <{fraction (1/100)}), the numerical optimization becomes very protracted and is not capable of bringing forth any appropriate improvement of the result despite this.
In accordance with a preferred embodiment, 9 segments of, for example, −80 ppm to +80 ppm, are provided. The segmentation in accordance with the invention may unfold its gauges typically beginning with 4 segments. The optimal number will, however, depend upon the individual case. In a concrete application, operations were successful with, for example, 9 or 14 segments.
A terminal which has a circuit for determining a difference frequency existing between the clock pulses and a correspondingly relevant segment, and which disposes over a memory which contains a modulation curve for each segment is basic for a data transmission system of the invention. With unequally large segments, a table with the respective segment limitations should furthermore be provided.
The following advantageous embodiments and feature combinations result from the detailed description and the totality of patent claims.
REFERENCES:
patent: 5563891 (1996-10-01), Wang
patent: 0248551 (1987-12-01), None
patent: 0549125 (1993-06-01), None
patent: WO91/18457 (1991-11-01), None
“Measured Pulse-Stuffing Jitter In Asynchronous DS-1/Sonet Multiplexing With and Without Stuff-Threshold Modulation Circuit”, W. D. Glover et al.,Electronic Letters,vol. 23 No. 18, pp. 959-961, Aug. 27, 1987.
Saman S. Abeysekera et al., Australian Telecommunications Research Institute, pp. 825-828.
Saman S. Abeysekera et al., Australian Telecommunications Research Institute, pp. 169-172.
Menzi Ulrich
Zahnd Hans
Ascom Transmission AG
Ton Dang
LandOfFree
Padding process for plesiochronous data transmission (II) does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Padding process for plesiochronous data transmission (II), we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Padding process for plesiochronous data transmission (II) will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2939578