Optical: systems and elements – Deflection using a moving element – Using a periodically moving element
Reexamination Certificate
1999-05-06
2003-03-11
Pascal, Leslie (Department: 2633)
Optical: systems and elements
Deflection using a moving element
Using a periodically moving element
C359S199200, C359S199200, C359S199200
Reexamination Certificate
active
06532087
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to a system for providing a measure of the clarity of a signal on a multiplexed fiber optic transmission line. More particularly, the present invention relates to a system for continually scanning for the Q-factor of multiple signals, individually and sequentially, being transmitted along a single fiber optic line.
In any transmission system, including optically transmitted systems, it is desirable to know the accuracy of the transmitted data at the receiver, i.e., the end of the system. In a digital system, the transmitted signal comprises a plurality of 1's and 0's, i.e., a plurality of high and low signals. Thus, errors in transmission occur when these 1's and 0's are not properly identified by the receiving circuit.
The digital values “1” and “0” each have an ideal voltage associated with it depending upon the parameters of the transmission circuit. Since no system is completely ideal, however, the actual 1's and 0's being transmitted will run through a range of voltages around the ideal voltages. For this reason, the ideal voltage can also be referred to as a mean voltage &mgr;, since it is the average voltage transmitted for a given digital value. The distribution of these voltages will either be Gaussian or at least a close approximation of Gaussian for the range of interest.
This means that the distribution of voltages transmitted as a particular value, i.e. “1” or “0,” will fall into at least an approximation of a Bell curve, as shown in FIG.
1
. As shown in
FIG. 1
, for a given digital value, i.e. “1” or “0,” half of the transmitted voltages will fall above the mean voltage &mgr;, half will fall below the mean voltage &mgr;, and a majority will fall within three standard deviations
3
&sgr; of the mean voltage &mgr;.
In a given system, therefore, it is necessary to set some threshold voltages for the 1's and 0's, i.e., HIGH and LOW thresholds, to allow a decision circuit to determine what will be identified as received 1's and 0's. The closer these thresholds are to the mean voltage &mgr;, the greater the error rate will be, and the farther away, the lower the error rate will become.
The accuracy of the transmitted signals can be determined by a statistic called the bit error rate (BER). The BER is the number of errors per bit transmitted, and depends upon the decision threshold. Another indicator of the accuracy of transmission can be given by the transmission's Q-factor. The Q-factor is an indicator of the signal quality at the decision circuit.
While the BER is easy to understand, the Q-factor is generally considered a more useful indicator of the accuracy of a transmission circuit, because it can be used to characterize the signal quality under conditions in which it is not practical to measure the BER. For this reason, it is preferable to determine a circuit's Q-factor rather than its BER. The Q-factor is related to the BER at the optimal threshold setting by the following formula:
BER
=
1
2
erfc
(
Q
2
)
≈
1
2
[
1
(
2
π
)
Q
×
ⅇ
-
Q
2
2
]
(
1
)
BER
≈
1
2
[
1
2
π
×
1
Q
×
ⅇ
-
Q
2
2
]
(
2
)
As a result, it is possible to determine the Q-factor of a signal by first measuring BER versus threshold for both “1s” and “0s”, and then fitting the results to extract the Q-factor.
This relationship is helpful, since the BER is more readily measured than the Q-factor. To measure the BER of a signal a measuring circuit need only monitor an incoming circuit for errors and determine how frequent the errors are. The accuracy of an error count is roughly {square root over (N)}, so a rule of thumb is that 10 errors causes an uncertainty of 3%.
In conventional optical transmission systems, BERs of 10
−15
, i.e., one error per 1,000,000,000,000,000 bits transmitted, are typical. These low BERs lead to one significant problem. Given the small number of errors, it is extremely difficult to actually measure the BER of an optical system in an efficient manner. Since an accurate BER measurement (3% uncertainty) requires the measuring circuit to detect ten individual errors it is necessary to run the measuring circuit for a sufficient period of time for ten errors to pass through. This means that with a BER of 10
−15
, the detection circuit would have to actually detect 1016 data bits before it detected the ten errors required for an accurate BER measurement. For an optical system that can transmit 2.488×10
9
bits per second (i.e., OC48), it would take nearly 4×10
6
seconds, or 46 days, for ten errors to be detected, and thus for the BER to be accurately determined.
In addition, many times optical fibers transmit more than one signal along a single multiplexed fiber. Just as it is desirable to determine the Q-factor of a single signal, and thus its BER, it is just as important to determine the Q-factor of each of the multiple signals transmitted over a multiplexed fiber optic line. However, the problems noted above associated with BER measurements are increased as the number of multiplexed signals grows.
If the signals are to be monitored sequentially, the time required for testing the BER will be multiplied by the number of signals multiplexed on the fiber optic line, further exaggerating the unreasonably long time required for an accurate measurement of BER. And even if the measurements were performed in parallel, the time required for determining the BER would remain unreasonably long, while the equipment costs for the testing device would increase dramatically. Thus, even using the most efficient and expensive approach, it would still require a month and a half to accurately obtain an accurate measurement of the BER of all of the multiplexed signals being transmitted over the fiber optic line.
This is too long a time for any effective testing circuit to employ such a method. As a result, it is extremely difficult to measure the true BERs for the threshold voltages used in optical transmission systems, and thus similarly difficult to determine the systems' Q-factors. It is therefore desirable to have a way of easily determining the BER or Q-factor for single-signal or multiplexed fiber optic lines without having to wait over a month for each test sample.
One possible method of estimating BER was suggested in detail in “Margin Measurements in Optical Amplifier Systems,” by Neal S. Bergano, et al., IEEE Photonics Technology Letters, Vol. 5, No. 3, March, 1993, (“Bergano et al.”) the contents of which are herein incorporated by reference. Bergano et al. observes that high values of BER can be easily measured and plotted against their respective threshold voltages. If several measurements are taken of high BER values and plotted against the median voltage on a logarithmic scale, the resulting curve is a close approximation of a straight line. Bergano et al. then suggests plotting one line for the threshold for 1's transmitted through the optical system and another line for 0's transmitted through the optical system.
The point at which these two lines intersect will be the point where the optimal threshold voltage is for marking the difference between 1's and 0's, and will show the BER for that threshold voltage. In this way, the BER can be quickly determined for an ideal threshold voltage, even if the time required to actually confirm that BER would be great. Using Bergano et al.'s method, measurements need only be taken for several larger BER's, which will take a dramatically shorter amount of time.
However, Bergano et al. does not suggest any circuitry for implementing this method, nor does it address the problems inherent in implementing the method into a physical circuit. In addition, Bergano et al. concerns itself with a single signal transmission, and so fails to address the problems associated with multiplexed fiber optic lines.
It is therefore desirable to pro
Bhatnagar Vipul
Livas Jeffrey C.
Nguyen Minh T.
Ransford Michael J.
Taylor Michael G.
Cammarata Michael R.
Ciena Corporation
Daisak Daniel N.
Pascal Leslie
LandOfFree
Multiple signal Q-tester does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Multiple signal Q-tester, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Multiple signal Q-tester will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3067452