Note: Descriptions are shown in the official language in which they were submitted.
CA 02704277 2010-05-14
DRIVER CIRCUIT FOR THE DIRECT MODULATION OF A LASER DIODE
FIELD OF THE INVENTION
The present invention relates to the field of laser diodes and more
particularly
concerns a driver circuit for driving a laser diode based on a pulsing input
signal.
BACKGROUND OF THE INVENTION
Seeding a pulsed fiber laser oscillator with a directly modulated laser diode
is a
simple and cost-effective way of generating high energy, high peak power
optical
io pulses with high stability. In material processing applications where a
single pulse is
used to process a given structure, such as severing links for memory repair,
it is
important to keep the pulse energy within a predetermined range. If the pulse
energy
is too low, then the link could be only partially fused, whereas if the energy
is too high
the substrate can be damaged, leading to failure mechanisms. It is therefore
highly
desirable to have optical pulses that have very high energy and amplitude
stability. In
a MOPA configuration (Master Oscillator Power Amplifier), as any undesirable
features present at the seed level will be amplified, it is consequently of
very high
interest to have methods for pulsing a laser diode in a very predictable and
repeatable way when seeding a first fiber amplifier section, so as to form
what is
generally called a Master Oscillator. The light from the Master Oscillator can
be
further amplified in a power amplifier which can be based on a large mode area
doped fiber, or made of a solid state material such as a Nd:YVO4 rod.
Optical pulse shaping is also of great interest in material processing
applications as it
offers the ability to control how the energy is delivered to the target over
time.
Elaborate optical pulse shapes require a highly responsive optical shaping
mechanism. It is also of great interest to have the ability to generate
optical pulses
with fast rise time as it allows sophisticated material processing techniques.
CA 02704277 2010-05-14
2
Having a scheme for generating stable optical pulses that can be spectrally
tuned is
also highly desirable as it allows, for example the matching of a pulsed fiber
laser
oscillator with a Nd:YVO4 amplifier, which has a very narrow gain bandwidth
centered
around 1064.5 nm. Having the ability to tune continuously the seed laser
emission
wavelength on the peak gain of an amplifier or on peak efficiency of a
frequency
conversion apparatus improves significantly the stability and reliability of
such a
system.
DELADURANTAYE et al (see the international patent application published under
no.
io W02009/155712) teaches a versatile platform for generating pulse shaping
signals
adapted for such applications, as well as several laser oscillator
configurations
generating shaped laser pulses based on such pulse shaping signals. In some
configurations, the pulse shaping signal may be used to modulate the current
of a
seed laser diode directly, so that the seed laser diode outputs the shaped
laser
pulses. Driving a laser diode with high-speed pulsed signals however
represents a
real challenge to electronics designers. For example, difficulties arise if
one wants to
drive a laser diode with a relatively high current around 1 A or more with
very short
rise time and fall time, in the range of 1 ns or less. At these speeds,
difficulties such
as parasitic inductance and capacitance effects greatly limit the performances
of the
circuit.
There is therefore a need for a driver circuit which allows the driving of a
laser diode
with a pulse shaping signal at a high current and with short rise and fall
times.
SUMMARY OF THE INVENTION
According to one aspect of the invention, there is provided a driver circuit
for
providing a driving signal to a laser diode based on a pulsed input signal.
The driver
circuit includes, successively:
a signal input receiving the pulsed input signal;
CA 02704277 2010-05-14
3
an amplification stage amplifying the pulsed input signal;
- a buffering stage including a plurality of voltage buffers connected in
parallel
and a plurality of pre-emphasis circuits each provided at an output of a
corresponding one of the voltage buffers; and
- a biasing stage adding a biasing voltage to the pulsed input signal, thereby
obtaining the driving signal.
In accordance with another aspect of the invention, there is provided a driver
circuit
for providing respective driving signals to an anode and a cathode of a laser
diode
io based on a pulsed input signal.
The driver circuit first includes a signal input receiving the pulsed input
signal and
dividing the same into first and second input signal components. Inverting and
non-
inverting branches then each receive one of the first and second input signal
is components and output one of the driving signals.
Each branch includes, successively:
- an amplification stage amplifying the corresponding input signal component,
the amplification stage of the inverting branch inverting the corresponding
input
20 signal component;
- a buffering stage including a plurality of voltage buffers connected in
parallel
and a plurality of pre-emphasis circuits each provided at an output of a
corresponding one of the voltage buffers; and
- a biasing stage adding a biasing voltage to the corresponding input signal
25 component, thereby obtaining a corresponding one of the driving signals.
In a preferred embodiment, the amplification stage includes two serially
connected
high speed amplifiers.
CA 02704277 2010-05-14
4
Advantageously, embodiments of the present invention provide high-speed laser
diode driving circuits which are capable of adapting the electric signal
provided by a
pattern generator, such as a low voltage signal transmitted in a coaxial cable
with a
typical impedance of 50 0, to drive a laser diode. The output of the laser
diode is an
optical signal that reproduces the electric signal from the pattern generator
with high
fidelity and low amplitude noise.
Other features and advantages of the present invention will be better
understood
upon reading of preferred embodiments thereof with reference to the appended
io drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1A is a graph showing a pulse of an exemplary pulse shaping input signal
used
to drive a laser diode though a driving circuit according to an embodiment of
the
invention; FIG. 113 is a graph showing the resulting optical pulse outputted
by the
laser diode.
FIG. 2 is a schematic representation of a driver circuit according to an
embodiment of
the invention.
FIG. 3 illustrates the evolution of an exemplary pulse shaping signal through
the
driver circuit of FIG. 2.
FIG. 4 is a schematic representation of an optimized topology for the driver
circuit of
FIG. 2.
FIGs. 5A to 5D schematically illustrate the relationships between the pulsed
input
signal (FIG. 5A), the applied voltage (FIG. 5B) and output optical power (FIG.
5C) as
CA 02704277 2010-05-14
a function of circulating current and the resulting optical pulse shape (FIG.
5D) for a
typical laser diode.
DESCRIPTION OF PREFERRED EMBODIMENTS
5 In accordance with one aspect of the present invention, there is provided a
driver
circuit for providing driving signals to a laser diode, based on a pulsed
input signal.
Preferred embodiments of the invention are particularly adapted for the
driving of a
high-speed laser diode, such as used for microprocessing applications and the
like.
io The expression "laser diode" is understood to refer to a laser with a
semiconductor-
based gain media. The laser diode may for example be embodied by a 14-pin
butterfly packaged diode such as used in the telecommunication industry, but
other
type of diodes may of course be considered, such as DFB (distributed feedback)
laser
diodes, DBR (distributed Bragg reflector) laser diodes, external cavity laser
diodes
with a short cavity (< 1 cm) and a wavelength selective feedback element such
as a
small diffraction grating, for example, seeded Fabry-Perot short cavity laser
diodes,
Fabry-Perot short cavity laser diode with wavelength selective feedback, etc.
The expression "pulsed input signal" is understood to refer to an electrical
signal of
varying intensity and defining one or a succession of pulses. Each pulse
preferably
has a rise time, a fall time and a duration therebetween. In some embodiment,
the
pulsed input signal may be designed for so called "pulse shaping", that is,
the signal
amplitude may vary over its duration, defining any desired pulse shape. The
driving
circuit according to embodiments of the invention preferably adapts this
electric signal
to drive a laser diode. The output of the laser diode is an optical signal
that
reproduces the pulsed input signal, preferably with high fidelity and low
amplitude
noise.
CA 02704277 2010-05-14
6
Laser diodes are usually low impedance devices above their conductive
threshold.
Furthermore, typical packages of laser diodes (such as the 14-pin butterfly
package
used in the telecommunication industry) are not necessarily optimised for high-
speed
modulation. As a consequence, the driver circuit according to embodiments of
the
invention incorporates appropriate circuitry for optimal impedance matching.
The pulsed input signal may originate from any appropriate pattern generator.
For
example, the pulsed input signal may be generated by a digital pulse shaping
module
such as described in U.S. patent application no. 12/493.949 (DELADURANTAYE et
1o al.), the contents of which are incorporated herein by reference. The
pulsed input
signal may however originate from different pattern generators, such as the
Tektronix
AWG5000B or the Agilent 81180A. Typically, the pulsed input signal will have a
low
voltage, for example in the range of 0 to 2 V and be transmitted through a
coaxial
cable with a typical impedance of 50 0.
FIGs. 1A and 1B respectively show a pulse of an exemplary pulsed input signal
used
to drive a laser diode though a driving circuit according to an embodiment of
the
invention, and the resulting optical pulse outputted by the laser diode. In
this example,
the pulse repetition rate is 200 kHz and the pulse duration 32.5 ns. The
electrical
signal has a peak-to-peak amplitude of 500 mV into 50 0. The optical signal
was
detected with a 2-GHz InGaAs photodiode. The peak power of the optical pulse
is
250 mW and its center wavelength is 1064.5 nm.
Typical patterns of interest in material processing applications are pulses of
duration
varying from 1 nanosecond (ns) or less, to a few hundreds of nanoseconds at
repetitions rates varying from single pulse to a few MHz. The pulses can have
complex shapes and it is an advantageous feature of the laser diode driving
circuit
according to a preferred embodiment of the invention that it can reproduce
complex
electrical shapes. However, the invention may also be used in contexts where
simpler
CA 02704277 2010-05-14
7
pulse shapes (square or the like) are desired. The electric duty cycle of the
signal to
be reproduced is usually low, typically less than 10% (for example a signal
composed
of square-shaped 10-ns pulses at a repetition rate of 100 kHz will have a duty
cycle of
0.1%).
For example, the bandwidth Fc [Hz] required for having a 1 ns rise time and
fall time
is given by the following formula:
Fc = 0.35 / rise time
A bandwidth of at least 350 MHz is therefore required. For this bandwidth, the
detrimental effects of parasitic inductance become significant when the
impedance
due to this parasitic inductance is of the order of the resistance associated
with the
laser diode above its conductive threshold, which is around 1 0 for typical
laser
diodes used as seed sources:
I Xp = RL
Where I Xp I is the amplitude of the parasitic impedance and RL [0] is the
laser diode
resistance.
By definition, we have :
I Xp 2xTrxFcxLp
Consequently :
Lp = RL / (2 x Tr x Fc)
where Lp is the equivalent parasitic inductance that could limit the
performances of a
conventional circuit driving an ideal laser diode.
Using the numerical example above, one obtains:
Lp = 0.45 nH
CA 02704277 2010-05-14
8
This inductance value is low compared to the realistic parasitic inductance of
4 nH
associated with a laser diode mounted in a 14-pin butterfly package, which is
a
common package for fiber-pigtailed, single-mode laser diodes. For a 4-nH
package
inductance, an equivalent inductive impedance of 8.8 0 is then combined in
series
with the intrinsic resistive impedance of 1 s2 of the laser diode.
In order to get a laser diode current of 1 A with a rise time of 1 ns or less,
this higher
impedance, and not only the laser diode resistance of 1 0, must be taken into
account. A higher voltage must therefore be applied to the laser diode very
quickly in
less than 1 ns. The embodiments of driver circuit described below allow this
goal to
io be achieved.
Referring to FIG. 2 there is shown a driver circuit 20 according to a
preferred
embodiment of the invention.
The driver circuit includes a signal input 22 which receives the pulse shaping
input
signal 24 from an appropriate pattern generator (not shown). As mentioned
above,
the pulsed input signal 24 is typically transmitted over an input coaxial
cable 25, and
the signal input 22 therefore preferably includes a coaxial connector 26 and
is
impedance matched to the input coaxial cable 25 (50 0 impedance in the
preferred
embodiment).
In the illustrated embodiment, the pulsed input signal 24 is divided into
first and
second input signal components 28 and 30, respectively received into one of an
inverting branch 32 and a non-inverting branch 34. Each branch 32, 34 outputs
one of
the driving signals 36 and 38 to a cathode 42 and an anode 40 of a laser diode
44.
However, it will be understood that in other embodiments the driver circuit 20
may
include only one branch, providing a same driving signal to the cathode 42 or
the
anode 40 of the laser diode 44 while the other electrode is maintained to
ground level.
CA 02704277 2010-05-14
9
Each of the inverting and non-inverting branches 32 and 34 include the
following
stages, successively:
1) an amplification stage 46 amplifying the corresponding input signal
component.
As their names entail, the amplification stage 46 of the inverting branch 32
inverts the
corresponding input signal component 28, while the amplification stage 46 of
the non-
inverting branch is non-inverting. Preferably, the amplification stage 46 of
each
1o branch includes first and second serially connected high speed amplifiers
48 and 50.
Providing two sub-stages of amplification may advantageously avoid a
degradation of
the rise and fall times caused by the gain-bandwidth limitation of the
amplifiers 48 and
50. In this manner, the fastest rise times and fall times achievable with
these high-
speed amplifiers can be obtained. In other embodiments, three or more
amplifiers per
1s branch may be provided. In a given branch 32 or 34, the amplifiers 48 and
50 may be
identical or dissimilar.
Further preferably, the second amplifier 50 of the inverting branch 32
performs the
signal inversion. By "inversion" of the signal, it is understood that the
voltages in both
20 branches 32 and 34 have the same magnitude but opposed signs.
The amplifiers 48 and 50 may for example be embodied by Texas Instruments
THS3202 chips. In the illustrated embodiment, the voltage gains are
respectively 5
and 5.5 for the two stages.
2) a buffering stage 52.
The buffering stage 52 of each branch 32, 34 includes a plurality of voltage
buffers 54
connected in parallel, in order to increase the current drive capacity, for
example up
CA 02704277 2010-05-14
to 1 A in the preferred embodiment. Voltage buffers are used to transfer a
voltage
from a high output impedance circuitry to a low input impedance circuitry.
Devices
suitable for this use are for example the buffers model BUF602 available from
Texas
Instruments. The number of voltage buffers 54 set in parallel depends on the
targeted
5 injection current to the laser diode 44. A significant degradation of the
pulse rise and
fall times can occur if the output current per buffer is over 200 mA for this
particular
buffer model. In such an embodiment, obtaining the desired current of 1 A
within 1 ns
may requires six or more buffers 54 per branch 32 and 34.
1o A plurality of pre-emphasis circuits 56 is also provided, each pre-emphasis
circuit 56
being connected serially at the output of a corresponding voltage buffer 54.
The role
of each pre-emphasis circuit 56 is to allow momentarily the application of the
full
voltage available from the output of each voltage buffer 54 to the laser diode
electrodes (anode 40 and cathode 42). This `pre-emphasis' circuit is
preferably a
differentiating RC circuit that produces positive or negative overshoots at
each
transition in the pulse shaping input signal. Each pre-emphasis circuit 56
therefore
includes a capacitor 58 and a resistor 60 connected in parallel. The values of
the
resistor 60 and capacitor 58 are trimmed as functions of the overall parasitic
inductance of the laser diode 44, the number of voltage buffers 54 and the
targeted
current value injected in the laser diode 44. Typical values for the
resistance and the
capacitance are 62 0 and 5.6 pF for a two banks of six buffers configuration.
The outputs of all the pre-emphasis circuits 56 of a given branch 32 or 34 are
then
combined and the resulting signal forwarded to the next stage.
3) a biasing stage 62 adding a biasing voltage 64 to the corresponding input
signal
component.
CA 02704277 2010-05-14
11
In the preferred embodiment, the driver circuit 20 includes a bias input 66
receiving
the biasing voltage 64 and dividing this bias voltage 64 into first and second
biasing
voltage components 68 and 70. The bias voltage is selected to properly bias
the laser
diode 44, and is typically adjusted around the conductive threshold of the
laser diode
44 and below the lasing threshold voltage of the laser diode 44. In the
preferred
embodiment, the biasing voltage 64 is embodied by a DC voltage of the order of
1 V.
The biasing stage 62 of each branch 32 and 34 is in turn provided with a bias-
T circuit
72 coupling one of the first and second biasing voltage components 68 and 70
to the
to corresponding branch 32 or 34. As one skilled in the art will readily
understand, a
bias-T circuit is typically composed of a low-inductance capacitor 74
connected in
parallel with an inductor 76. The Bias-T circuits 72 mainly serve two
purposes: they
are used as high-pass filters for the application of the high-speed input
signal
components 28 and 30 as outputted by the buffering stage 52, and they are used
as
low-pass filters for the application of the DC biasing voltage 64.
As one skilled in the art will also understand, the simultaneous application
of the
pulsed input signal 24 to the anode 40 of the laser diode and of an inverted
version of
this same signal to the cathode 42 provides for a differential driving of the
laser diode
44. The main advantage of using a differential driving circuit is to double
the
maximum voltage range of the voltage buffers 54. However, in embodiments where
this doubling of the maximum voltage range is not necessary, a single branch
may be
provided without departing from the scope of the present invention.
FIG. 3 illustrates the evolution of an exemplary pulse shaping signal at each
of the
stage of the driver circuit according to the illustrated embodiment of FIG. 2.
In the first
stage of voltage amplification, each component of the pulsed input signal is
amplified
in a linear fashion in both branches of the circuit. In the second stage of
voltage
amplification, the input signal components are further amplified; in the non-
inverting
CA 02704277 2010-05-14
12
branch the voltage sign is identical to that of the pulsed input signal,
whereas in the
inverting branch, the second stage of amplification inverts the sign of the
corresponding input signal component.
The buffering stage has a voltage gain of 1 and does not affect the input
signal
components. The purpose of this stage is to increase the current driving
capability of
the circuit for driving a low impedance component such as a laser diode. After
the
pre-emphasis portion of the buffering stage, however the input signal
components are
slightly distorted, the transitions therein are more pronounced as positive
and
1o negatives overshoots having been generated by the differentiating RC
circuits 56 for
each of these transition.
FIGs. 5A to 5D illustrates the relations between the bias voltage, the pulse
shaping
signal applied to the laser diode and how this voltage varying signal is
converted to a
shaped optical pulse in the laser diode. Referring to FIGs. 5A and 5B, the
bias
voltage is adjusted around the conductive threshold of the laser diode. The
purpose
of this bias is to bring the laser diode close but under its lasing threshold
so as to
operate the laser diode in a gain-switched type of operation. Without the bias
voltage
the high-speed driving circuit would have to provide the full voltage to bring
the laser
diode to its conductive threshold and above for optical emission, which may
have a
detrimental effect on the performances of the system. When the pulsed input
signal is
applied to the laser diode, the laser diode is brought rapidly above its laser
threshold.
If the laser cavity is short enough, the laser oscillation will quickly
establish itself and
provide fast rise time. Above the conductive threshold, the relation between
the
applied voltage and the associated circulating current in the laser diode is
linear (see
FIG. 5B). In the same manner, above the laser threshold, the relation between
the
circulating current and the emitted optical power is linear (see FIG. 5C). As
a
consequence, above the laser threshold, the emitted optical power is in linear
relation
with the applied voltage on the laser diode chip, the optical pulse (FIG. 5D)
will follow
CA 02704277 2010-05-14
13
the pulse shaping signal. Operating the laser diode in such a fashion, in a
gain-
switched regime, ensures that significant optical power will only be emitted
when the
pulsed input signal is present at the input of the circuit, and therefore
limits the optical
power emitted by the laser diode between successive pulses.
As one skilled in the art will readily understand, the transmission of the
electrical
pulsed input signal from the driving circuit to the laser diode chip may
impact heavily
on the performance of a tailored pulse shaping signal during fast transients.
As
described previously, the trimming of the 'pre-emphasis' circuit will help in
to compensating the parasitic inductance and capacitance related to the signal
transmission. Further design elements which may help in that matter are:
minimizing the physical length over which the signal is transmitted between
the
drive circuit and the laser diode chip;
- increasing the number of wires or the wire diameter for the wire bonding
inside
the laser diode package;
using stripline technology for transmission of the pulsed input signal to the
laser diode chip;
reducing metal contacts areas on the laser diode chip for decreasing the laser
chip capacitance; and
decreasing the length of the laser chip for decreasing its capacitance.
The parasitic inductance of the layout of the printed circuit board embodying
the
invention should preferably be reduced to a minimum for the electronic driver
and the
mounting of the laser diode. For those who are skilled in the art, the
parasitic
inductance of a printed circuit board trace is function of the trace length,
width and the
height of the trace over the ground plane. For example, a printed circuit
board trace
with a length of one inch (25.4 mm), a width of 10 mils of an inch (254 pm)
and a
height of 50 mils of an inch (1.27 mm) will have a parasitic inductance of
20.4 nH.
CA 02704277 2010-05-14
14
This value is five times the parasitic inductance of the laser diode case. In
order to
reduce this parasitic inductance under 1 nH, the parts placement and the
printed
circuit board trace layout are preferably optimized. An example of an
optimized
topology is shown in FIG. 4, where the first amplifier, the buffering stage
and the
biasing stage of the first and second branches, respectively, are superposed.
All
similar components of the inverting and non inverting branches are preferably
top-
and-bottom mounted, with the exception of the second amplifiers 50 (which is
either
inverting or non-inverting).
1o It will be readily understood by one skilled in the art that the present
invention may be
embodied by different devices, hardware, elements, combinations of elements or
the
like in different arrangements than shown in the appended drawings and
described
herein. Additional hardware may also be provided depending on the need of a
particular application as is well known in the art.
Of course, numerous modifications could be made to the embodiments described
above without departing from the scope of the present invention as defined in
the
appended claims.
