need help with revising a laser diode APC circuit
Ray Xu
Accompanying file: http://www.mediafire.com/download.php?wyz0wmnnxc4
I just started designing my version of a laser diode driver circuit;
and I was hoping you could look over it and give me some feedback. I
designed it to be similar to the 2-transistor versions; except using
op amps (it should allow for faster modulation; and I find using op
amps very simple). The theory is that the current going into the
laser diode is a inverse linear function to the photodiode current.
The circuit should be able to control a common-anode laser diode (yes,
I made a mistake in the laser diode schematic symbol), with monitor
currents of around 0.3 to 0.5mA, and a laser diode operating current
of how ever much the precision current sink can sink. The precision
voltage bias is to provide a stable operating voltage with respect to
the virtual ground (4.5V).
The modulation controller circuit block uses a switch such as using
the CD4066 to switch between the signal from the monitor photodiode or
ground. The op amp acts as a buffer to compensate for any signal
attenuation caused by the switch (because they have relatively large
resistances). For a logic HIGH signal, the switch would be switched
to the output of the op amp, and it would be switched to ground for a
logic LOW.
The inverting amplifier block is to provide the inverse function and
to amplify the signal form the photodiode monitor I-V converter so
that the ratio between photodiode and laser diode is 1:1 (inverted).
I have yet to include the small specific details (including the
resistor and part values); this is just an idea. I also lacked the
filtering capacitors in the diagram for simplicity.
The voltage supplies would be equivelant to +/- 4.5V; VCC=+9V, VEE=0V,
and a +4.5V as virtual ground.
I'm specifically worried about the modulation controller block; where
it switches back and forth between ground and the output of the op
amp. Also please provide feedback.
I'm pretty new to lasers; I only know how not to damage them by static
and how to drive one using a simple 2-transistor circuit ;)
Many Thanks
Ray Xu
rayx...@gmail.com
Kai-Martin Knaak
> I just started designing my version of a laser diode driver circuit; and
> I was hoping you could look over it and give me some feedback.
Some questions spring to mind:
* What exactly do you mean by "precision power supply"? This black-box
does the real work in your circuit. So unless you give more information
it is not possible to say whether your controller might stand a chance to
work properly.
* Are you aware of the subtilities of the transimpedance amplifier sub
circuit you use to pump up the photo diode signal?
* What will the supply?
* How will you produce the intermediate potential 4.5 V?
> I
> designed it to be similar to the 2-transistor versions; except using op
> amps (it should allow for faster modulation;
Are you aware of Sams excellent laser-FAQ? There is a pretty long chapter
on diode laserdrivers, complete with circuits.
http://www.repairfaq.org/sam/laserdps.htm#dpstoc
If I were you, I'd start with the circuit based on the voltage regulator
LM317 ("Simple Laser diode Power supply"). Yes, there is no feedback from
the photo diode. But I doubt, you really need this complicating feedback.
For extra safety I'd add another LM317 in standard voltage regulating
configuration at the input of the circuit.
For fast modulation I'd add a capacitor/resistor directly at the diode.
The result would be a robust and simple constant current laser driver
with modulation. Yes, I have successfully operated laser diodes with such
a driver. Modulation frequency was 100 kHz to 200 MHz.
---<(kaimartin)>---
--
Kai-Martin Knaak
Öffentlicher PGP-Schlüssel:
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x6C0B9F53
Ray Xu
Precision voltage bias is to provide the 2.5V difference, such as with
the second LM317 in your description. The power will be supplied by
the 9V voltage regulators, and the 4.5V would be produced by dividing
the 9V in half (resistor divider).
And yes I have heard of sam's laser FAQ; it is where I built my first
diode driver (the 2-transistor version) like about 1 week ago.
And WOW, its amazing (to me) that you've got the diode to modulate at
200MHz, but I am not taking any risks running the laser without
feedback; my circuit has to be fail-safe as much as possible. I only
need it to modulate at at least 200KHz with feedback that drives a
common-anode laser diode. Got any ideas for this (excluding a laser
module)? I would personally prefer op amps, since I have more
experience with them than transistors.
I personally don't like the 2-transistor versions with a potentiometer
for brightness, because I don't have a laser intensity meter, and my
last experiment with that, the laser diode over heated.
Also, I have a few questions about the http://www.repairfaq.org/sam/laserdps.htm#dpshp1
if you dont mind. How can I modify that to support a common-anode
laser diode? And what do those 2 potentiometers at the bottom do? If
they are for brightness, can I just short them (as if they were not
there) to get the brightness listed in the specs of the specific laser
diode?
Thanks
Ray Xu
Kai-Martin Knaak
> Precision voltage bias is to provide the 2.5V difference, such as with
> the second LM317 in your description.
Sorry, I wasn't as clear as I wanted to be: The unspecified black box
that has to do all the work in your circuit, is the "precision current
sink". How do you intend to achieve this function in terms of actual
electronic parts?
> The power will be supplied by the
> 9V voltage regulators,
ok.
> and the 4.5V would be produced by dividing the 9V
> in half (resistor divider).
not ok.
The intermediate potential of the voltage divider will depend on the
current drawn from this point. That is, your common potential will depend
on the current current. This is asking for trouble.
> And WOW, its amazing (to me) that you've got the diode to modulate at
> 200MHz,
No big deal -- Just a little capacitor tied to one pin of the laser diode
and a 50 Ohm terminated cable on the other side. It was the photo diode
amplifier, that made me scratch my head...
> but I am not taking any risks running the laser without
> feedback;
The simple laser diode driver based on LM317 actually relies on
feedback -- feedback of the amount of current flowing through the diode
at any given instant.
It is true, that laser diodes are destroyed by too much light. However,
this is an extremely fast, self accelerating process. If started, the
laser diode turns into a LED in nanoseconds. Any feedback circuit you can
design, will be way too slow to stop the current fast enough. Thus, the
only chance is to stay away from dangerous levels of light by a
sufficient margin. This can be done by monitoring the output power, or by
making sure, the current never exceeds a certain level.
There is a reason, why data sheets list maximum current.
Feedback of optical power involves many more electronic parts than
feedback of current. That's why it is more at risk to show dangerous
transients and is much slower than the purely electric feedback. This is
why I'd avoid using the photo diode as a means of primary feedback. If
the application requires constant power, this may be an additional loop
to regulate the set point of the fast current regulation.
> my circuit has to be fail-safe as much as possible.
Add a power-on retarder for the supplies to avoid dangerous transients at
switch-on. This can be achieved with the LM317 (again). See the data
sheet by national.com for a circuit.
> I only
> need it to modulate at at least 200KHz with feedback that drives a
> common-anode laser diode. Got any ideas for this (excluding a laser
> module)?
As I said: I'd vote for the LM317 based circuit. Three LM317, that is.
Only reason for something more complicated would be special noise
requirements or a need for constant power.
> I would personally prefer op amps, since I have more
> experience with them than transistors.
I felt the same for quite some time...
Note, that the vast majority of opamps fails to delver more than about 20
mA. So they may just not cut it.
> I personally don't like the 2-transistor versions with a potentiometer
> for brightness, because I don't have a laser intensity meter, and my
> last experiment with that, the laser diode over heated.
Data sheets specify maximum current (for a reason). DC current can be
measured reliably with cheap digital multimeter.
>
> Also, I have a few questions about the
> http://www.repairfaq.org/sam/laserdps.htm#dpshp1 if you dont mind. How
> can I modify that to support a common-anode laser diode?
Why do you want to use such a complicated circuit?
> And what do those 2 potentiometers at the bottom do?
Calibration to achieve some defined relation of output (3) to actual
optical power of the laser diode. You'd need an optical powermeter to
find the correct position of the trimmers.
> If they are for brightness,
> can I just short them (as if they were not there) to get the brightness
> listed in the specs of the specific laser diode?
No.
Ray Xu
> > can I modify that to support a common-anode laser diode?
>
> Why do you want to use such a complicated circuit?
>
> > And what do those 2 potentiometers at the bottom do?
>
> Calibration to achieve some defined relation of output (3) to actual
> optical power of the laser diode. You'd need an optical powermeter to
> find the correct position of the trimmers.
>
> > If they are for brightness,
> > can I just short them (as if they were not there) to get the brightness
> > listed in the specs of the specific laser diode?
>
> No.
>
> ---<(kaimartin)>---
> --
> Kai-Martin Knaak
> Öffentlicher PGP-Schlüssel:http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x6C0B9F53
For the precision current sink: http://www.national.com/an/AN/AN-31.pdf
page 13. I had to leave that a "black box" since I already made a
bunch of mistakes there already. I may need to think about the
precision voltage bias and the precision current sink some more before
it goes on my final design.
For the resistor divider to achieve the 4.5V, are you saying that it
will not be 4.5V if current is drawn from there (just to make sure)?
I can substitute the voltage divider with another voltage regulator.
Hhmmm; now since you mentioned that the LM317 incorporates electrical
feedback (and is much faster), I'm starting to think that its actually
a good idea.
If optical feedback is slow, would it be slow for 200Khz and below of
modulation? Is there a higher risk of frying up the laser diode when
using optical feedback than electrical feedback, with modulation?
If I'm going to be operating a high-powered laser diode (which I
eventually will) with electrical feedback, just to be safe, I'll
prefer to operate it at 50% to 75% of the maximum current, but above
the threshold current (which means I'll need a higher-powered diode to
achieve my wanted power output, which means more $$$).
Thanks
Ray Xu
Kai-Martin Knaak
> For the precision current sink: http://www.national.com/an/AN/AN-31.pdf
> page 13.
This is essentially a darlington cascaded transistor stage, linearized
with an opamp. I'd say it will do the job.
Speed of the circuit is limited by the opamp.
> For the resistor divider to achieve the 4.5V, are you saying that it
> will not be 4.5V if current is drawn from there (just to make sure)?
yes.
> I can substitute the voltage divider with another voltage regulator.
no.
A voltage regulator acts like a valve. So its output can source current,
but it cannot sink current. However, your driver uses 4.5 V as common
potential. So current will flow from 9V to 4.5V. This current would need
to be drawn from the 4.5V line.
> Hhmmm; now since you mentioned that the LM317 incorporates electrical
> feedback (and is much faster), I'm starting to think that its actually a
> good idea.
:-)
> If optical feedback is slow, would it be slow for 200Khz
Feedback of optical power can be fast. But the same speed is much harder
to implement. For starters, you obviously need a fast photo detection
circuit.
> and below of
> modulation? Is there a higher risk of frying up the laser diode when
> using optical feedback than electrical feedback, with modulation?
More components, more complicated feedback loop ==> more risk
Kai-Martin Knaak
>> For the precision current sink: http://www.national.com/an/AN/AN-31.pdf
>> page 13.
>
> This is essentially a darlington cascaded transistor stage, linearized
> with an opamp. I'd say it will do the job.
^^^^^^^^^^^^^^^^^^^^^^^^^^^
On second thought, I revoke that statement! :-|
The amount of current the circuit sinks is determined by the Resistor R1
which you replace by the laser diode in your driver. However, the
resistance of the diode is not constant in any way. It depends on the
amount of current flowing. I addition, it depends on the temperature of
the diode. So this current sink would not be be "precision" as intended.
A proper transistor-opamp stage for the job employs a separate shunt
resistor to detect the momentary current. John Hall and K. Libbrecht
published a classic, well working design in the nineties:
http://www.its.caltech.edu/~atomic/publist/diodecontroller.pdf
A complete design based on this principle can be retrieved from the
webpages of the university of melbourne:
http://optics.ph.unimelb.edu.au/atomopt/electronics/index_claudius.html#Current
<shameless plug>
And from the webpage of my day-job:
http://bibo.iqo.uni-hannover.de/dokuwiki/doku.php?id=eigenbau:lasertreiber
</shameless plug>
However, these circuits are certainly overkill for your application.
They are designed to provide state of the art minimum noise current and
excellent temperature stabilization.
Ray Xu
OK, I've made my decision to use the LM317 constant-current method, or
something similar. Thanks for saving me from a step closer to
destroying a very expensive laser in the future ;-)
Since I may sell my laser driver circuit, I'd want it to be adaptive
to every kind of laser diodes (common-anode) as much as possible. To
solve that, I've developed a calibration procedure where at startup, a
circuit will slowly step (increase) the laser diode current until the
current at the photodiode matches the monitor current, then everything
is handed over to the LM317-based ACC circuit. The problem is that
how do I deal with stuff like environmental changes that will affect
the performance of the laser diode without having to calibrate often?
Thanks
Ray Xu
Lostgallifreyan
2ed8c1...@s16g2000vbp.googlegroups.com:
> The problem is that
> how do I deal with stuff like environmental changes that will affect
> the performance of the laser diode without having to calibrate often?
>
Not much you can do unless you control them. So either you need to calculate
the monitor photodiode current for worst case, or leave a margin for safe
operating above normal conditions, the second being best, because martgins
allow all sorts of safety, like exclusion of the effects of tiny surges that
mean nothing if the filtering is good and your diode is given some breathing
space, but can mean destruction if the tiniest little bump in current is
enough to push it over the edge.
Also, once you start using high power red diodes you'll find they usually
don't have photodiodes so this safe operating margin is something you'll have
to plan for anyway. Also, diodes vary between each other even from the same
batch, and widely enough that statistical analysis of lifetime performance
and failure is the only way to really know anything about the safe margins,
calulations mean very little, and the data sheet specs serve only to reduce
the amount of guessing you have to do. Lased diodes are notoriously weak on
firm specifications as it is.
If you search through past posts by me on this group you'll find several
posts all relating to an efficient LM317 based driver that has to-zero linear
current adjust, and a safe upper limit set by a hardwired fixed resistor. It
uses a small voltage reference to dupe the regulator into seeing the sense
voltage it needs so it will control the current at an arbitrary set level up
to maximum. LM417's are very cool. They are an op-amp and a pass transistor,
basically, so the power, accuracy, etc, are all there. It has very good
regulation if you take care with capacitors in appropriate places.. I've even
managed to add an op-amp (comparator based on differential amp) to give it
clean egded modulation up to around 100 KHz with no overshoot, but I won't go
into details because the last I saw of it it was still on two boards and
built out of whatever was to hand, it was a mess. Some time I might work on
it again but I haven't had a lot of reason to lately.