Optical: systems and elements – Optical modulator – Light wave temporal modulation
Reexamination Certificate
2001-10-24
2003-07-01
Mack, Ricky (Department: 2873)
Optical: systems and elements
Optical modulator
Light wave temporal modulation
C359S315000
Reexamination Certificate
active
06587249
ABSTRACT:
STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT
Not Applicable
The invention relates to apparatus comprising a light emitter (typically a laser diode) operating under steady conditions, and means for modulating the light emitted by the emitter as a function of an electrical signal representing information to be transmitted, typically digital information. Applications for sources of this type lie in particular in the field of optical fiber digital transmission at high bit rates (at greater than 1 gigabit per second (Gbit/s)).
Known modulators present a characteristic of transmitted light power as a function of control voltage which tends to vary over time, thereby affecting the quality of the emitted light signal.
More precisely, this characteristic tends to shift relative to the control voltage so that it becomes necessary to adapt the control voltage regularly as a function of these shifts.
The invention relates to means for compensating such drifts, both for an optical signal of conventional non-return-to-zero (NRZ) format and for a signal of return-to-zero (RZ) format, which means enable the propagation characteristics of optical fibers to be used advantageously. (An example of an RZ format signal is known under the term “soliton”.)
The modulators that are used most often, in particular interferometer type modulators, possess a sinusoidal characteristic of the kind shown in FIG.
1
. They can be made on lithium niobate (the usual case) or on a III-V semiconductor suitable for monolithic integration with a laser diode.
In conventional apparatuses, of the kind shown in
FIG. 2
, an electrical information signal S
1
is amplified by a driver
3
and is capacitively coupled to a modulator
5
. The control signal is positioned relative to the characteristic of the modulator
5
by means of a DC voltage referred to as a “bias” voltage which is applied to a line
22
and which is added to the amplified control signal. The excursion required in the control voltage to go from transparency (point B of the characteristic in
FIG. 1
) to extinction (point A of the characteristic) is a few volts. The lightwave at the output from the modulator then has “1” levels at the power PH of point B and “0” levels at the power PL of point A. The ratio PH/PL is the extinction ratio of the light signal, and it is advantageous for this ratio to be as large as possible.
The information signal must therefore have peak-to-peak amplitude equal to the amplitude Vpi of the sinewave. It is therefore desirable to add a bias voltage such that the sum (signal voltage+bias voltage) enables both points A and B of the characteristic to be reached.
Because of the way the control circuit is built, the bias voltage is the mean value of the signal actually applied to the modulator
5
.
It should therefore be observed that the optimum bias point depends on the nature of the electrical information signal.
The most usual case is that of a digital signal in which “1s” are represented by pulses having a duration equal to the bit time (100 picoseconds (ps) at 10 Gbit/s), and the signal is then said to have the non-return-to-zero (NRZ) format. An example of such a signal is represented by the upper curve in FIG.
3
. If it is assumed that the peak-to-peak amplitude of the control signal is Vpi, and that “0s” and “1s” are equally probable, then the optimum bias point lies in the center of the characteristic (point O in
FIG. 1
) and the optimum bias voltage is (VA+VB)/2.
When “1s” are represented by a pulse of duration shorter than the bit time, then the signal is said to be of return-to-zero (RZ) format. An example of such a signal is shown by the lower curve in FIG.
3
. Under such circumstances, the ratio between the peak-to-peak voltage of the total applied signal over its mean value is different, and the optimum bias point is no longer some remarkable point of the characteristic. For example, for a rectangular RZ pulse of duration equal to half a bit time associated with “1s” and “0s” of equal probability, this voltage would be: (3/4 VA+1/4 VB).
It is thus known that modulators are subject to slow drift having the effect of shifting the characteristic of
FIG. 1
in translation parallel to the horizontal axis. The causes of such drift are various and depend on the technology used to make the modulator. For modulators using lithium niobate, drift comes mainly from temperature variations, from charge trapped in the electrode-insulating dielectrics, and, at high light injection, from photorefractive effects.
The bias voltage must be servo-controlled in such a manner that the electrical control signal is constantly positioned optimally relative to the real characteristic.
Various solutions have been proposed to solve this problem of servo-controlling the bias voltage.
A first known method consists in superposing on the bias voltage a small amplitude sinewave signal at a frequency f that is situated below the working spectrum of the information signal. At the output from the modulator, a fraction of the light signal is taken using a coupler and is converted back into an electrical signal by means of a photodetector. Thereafter, the component at the frequency 2f in the detected signal is analyzed. If the modulator is biased to the center of the characteristic (the optimum point for an NRZ signal), then this component is zero since the bias point is a point of inflection. Otherwise, the non-linearity of the characteristic will generate a component at said frequency. Servo-control is thus performed so as to eliminate this component at the frequency 2f. The main drawback of this method is that it is applicable only to NRZ format.
Another solution proposed in U.S. Pat. No. 4,306,142 of 1981 consists in using a coupler which takes a fraction of the optical signal input to the modulator and a second coupler which takes light from its output. The signals are converted into electrical signals by photodetectors and they are analyzed by an electronic system including, in particular, duty ratio detectors (an equal probability RZ signal in which a pulse lasts for half a bit time has a duty ratio of 1/4). Servo-control is applied to ensure that the input and output duty ratios are equal. This principle can be used regardless of the format of the signal, but it nevertheless presents the main drawback of being expensive to implement for an application to transmission at high bit rates: photodetectors and duty ratio detectors are circuits which need to operate at the information rate, and at 10 Gbit/s, this requires expensive technology to be used.
Another solution which is also applicable independently of format is proposed in document U.S. Pat. No. 5,805,328. In that solution, two couplers and two photodetectors are used to measure the mean light powers at the input and at the output of the modulator. Servo-control then consists in keeping the ratio between these two powers constant. In this case the measured magnitudes are mean values and the components used can therefore be made out of low cost technology. The drawbacks of the method are firstly the use of two couplers, giving rise to losses of light power and secondly the fact that the servo-control can become defective, particularly in the event of the information signal having its peak-to-peak amplitude drop below Vpi (as can happen due to a small loss of driver performance). Under such circumstances, in order to compensate for the decrease in output power, the servo-control will tend to increase power during optical “0s” which quickly degrades the performance of the system. It can also be observed that the servo-control has the potential of being inaccurate since it depends on a ratio of two measured magnitudes, so measurement inaccuracies and variations in the measuring electronics as a function of temperature or of component aging will affect the overall behavior of the servo-control.
Servo-control techniques relying on a signal which has a physical reason for becoming zero at the optimum operating point are not subject to this defect.
Proposals have
Corvis Algety SA
Mack Ricky
Thomas Brandi N
LandOfFree
Apparatus and method for adjusting the control signal of an... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and method for adjusting the control signal of an..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and method for adjusting the control signal of an... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3001423