Screwy laser diode with monitor photodiode

[Ethan]

I am working on an instrument that uses a VCSEL light source for some
optical measurements. The current instrument has some problems with
the light output drifting with temperature. In order to correct this I
have been looking at using a VCSEL with a monitor photodiode to control
the light output. Presumably if I keep the signal at the monitor
photodiode constant, the output should be constant.

My problem is that as I change the current in the VCSEL the light
output as seen by an external photodiode does not track the light
output of the monitor photodiode packaged with the VCSEL. I have tried
several different devices, and they all have this problem, though
different devices have a different shaped relation between them, so
this is not a fixed transfer function that I can calibrate out. Since
the light out does not seem to track the monitor signal, if I control
the VCSEL current to keep the monitor current constant, the light
output is not nesicarily going to stay constant.

Does anybody have any suggestions as to what might be causing this? Is
there some mechanism for the monitor diode to pick up energy from the
VCSEL that is not from the light output?

My other thought is that it might be speckle on the two photodiodes
causing the inconsistancy. However this seems unlikely, since I think
the size of the photodiode is large relative to the size of a speck,
(what is signular for speckle?) that this would average out to
something insignificant.


Thanks for any suggestions.

Ethan Petersen

 

[John Larkin]

I am working on an instrument that uses a VCSEL light source for some
optical measurements. The current instrument has some problems with
the light output drifting with temperature. In order to correct this I
have been looking at using a VCSEL with a monitor photodiode to control
the light output. Presumably if I keep the signal at the monitor
photodiode constant, the output should be constant.

My problem is that as I change the current in the VCSEL the light
output as seen by an external photodiode does not track the light
output of the monitor photodiode packaged with the VCSEL. I have tried
several different devices, and they all have this problem, though
different devices have a different shaped relation between them, so
this is not a fixed transfer function that I can calibrate out. Since
the light out does not seem to track the monitor signal, if I control
the VCSEL current to keep the monitor current constant, the light
output is not nesicarily going to stay constant.

Does anybody have any suggestions as to what might be causing this? Is
there some mechanism for the monitor diode to pick up energy from the
VCSEL that is not from the light output?

My other thought is that it might be speckle on the two photodiodes
causing the inconsistancy. However this seems unlikely, since I think
the size of the photodiode is large relative to the size of a speck,
(what is signular for speckle?) that this would average out to
something insignificant.


Thanks for any suggestions.

Ethan Petersen

VCSELs are a mess. They have mode jumps, instabilities, sensitivity to
back reflections, all sorts of weirdness. What hits the monitor diode
is only loosely related to what exits the package.

And if you find a good supplier, they're almost guaranteed to change
the die or the process or discontinue the part in six months.

The old cleaved CD lasers were great, but you can't get them anymore.

Whose parts are you using? Lasermate never was very good; Optek was
good for a while.

John

 

[Ethan]

The parts I have tried are:
Advanced Optical Components HFE4083-322
Optec OPV310 and OPV314

I am just starting to set up an edge emmiting laser, US Lasers D850-5.
We'll see how that goes. I also have a couple more VCSEL samples to
try out, but don't have high hopes. One of which is from Lasermate.
For some reason the case is connected to the monitor diode pin rather
than the common, which just seems like a bad idea.

I am concerned about an an edge emmiting laser because of the oval
shaped beam, and how that is going to effect the instrument
calibration. I suppose this could be corrected with the right lenses,
but that will complicate the physical/mechanical design.

We'll see how it goes.

Ethan

 

[John Larkin]

The parts I have tried are:
Advanced Optical Components HFE4083-322
Optec OPV310 and OPV314

I am just starting to set up an edge emmiting laser, US Lasers D850-5.
We'll see how that goes. I also have a couple more VCSEL samples to
try out, but don't have high hopes. One of which is from Lasermate.
For some reason the case is connected to the monitor diode pin rather
than the common, which just seems like a bad idea.

I have quart-size ziploc bags full of reject Lasermate vcsels... bad
mode jumps, noise, rotten alignment. Our yield was under 50%.

And Optek keeps changing the chip - some are 8 pF, some are 35 - and
doesn't like to answer phone calls. Lasers suck.

Let me know if you find anything good.

John

 

[redbelly]

I am working on an instrument that uses a VCSEL light source for some
optical measurements. The current instrument has some problems with
the light output drifting with temperature. In order to correct this I
have been looking at using a VCSEL with a monitor photodiode to control
the light output. Presumably if I keep the signal at the monitor
photodiode constant, the output should be constant.

My problem is that as I change the current in the VCSEL the light
output as seen by an external photodiode does not track the light
output of the monitor photodiode packaged with the VCSEL. I have tried
several different devices, and they all have this problem, though
different devices have a different shaped relation between them, so
this is not a fixed transfer function that I can calibrate out. Since
the light out does not seem to track the monitor signal, if I control
the VCSEL current to keep the monitor current constant, the light
output is not nesicarily going to stay constant.

Does anybody have any suggestions as to what might be causing this? Is
there some mechanism for the monitor diode to pick up energy from the
VCSEL that is not from the light output?

My other thought is that it might be speckle on the two photodiodes
causing the inconsistancy. However this seems unlikely, since I think
the size of the photodiode is large relative to the size of a speck,
(what is signular for speckle?) that this would average out to
something insignificant.


Thanks for any suggestions.

Ethan Petersen

I haven't used vcsel's, just edge emitters. But I have never trusted
the internal photodiode on those. Any reflections that feed back to
the diode will also affect the internal photodiode.

If your setup is such that there are no reflections coming back to the
diode, then I don't know what else would cause this. I'm not sure if
vcsel's would behave any better or worse than edge emitters.

Mark

 

[Phil Hobbs]

redbelly wrote:
(snip)

I haven't used vcsel's, just edge emitters. But I have never trusted
the internal photodiode on those. Any reflections that feed back to
the diode will also affect the internal photodiode.

If your setup is such that there are no reflections coming back to the
diode, then I don't know what else would cause this. I'm not sure if
vcsel's would behave any better or worse than edge emitters.

Almost all VCSELs have multiple transverse modes. I saw one that ran in
an almost pure 7th order mode--12 off-axis peaks in the angular
spectrum, and a nice null in the centre. Multiple modes will give you
funny speckly patterns that are extremely noisy--often as much as 20 dB
noisier than the total integrated output, due to competition between modes.

Getting the monitor PD to track the output accurately will be very
difficult with a multimode device.

Cheers,

Phil Hobbs

 

[Ethan]

So it sounds like using the monitor diode is not going to work for me.
Does anybody have suggestions for how to get a tightly controlled
output from a laser?

I am using this to measure optical density of a fluid. I shine a laser
into the fluid and measure how much light comes out. I am modulating
the laser at a couple kHz, just to make the detection easier, and to
make it easier to cancel out external light sources.

This is pretty simple and for the most part this works pretty well. The
problem is the light source is drifting with temperature.

I have a few bad ideas, they are all physically cumbersome, such as a
beam splitter and an external monitor diode. Or a heater/thermister to
maintain a constant temperature just above ambient.

Ideas anyone?

Thanks

Ethan

P.S. I am realizing I don't know as much about VCSELs and other lasers
as I thought. Any suggestions on good books, or other references?

 

[John Larkin]

Almost all VCSELs have multiple transverse modes. I saw one that ran in
an almost pure 7th order mode--12 off-axis peaks in the angular
spectrum, and a nice null in the centre.

That sounds more like a muffin fan than a laser!

John

 

[redbelly]

So it sounds like using the monitor diode is not going to work for me.
Does anybody have suggestions for how to get a tightly controlled
output from a laser?

I am using this to measure optical density of a fluid. I shine a laser
into the fluid and measure how much light comes out. I am modulating
the laser at a couple kHz, just to make the detection easier, and to
make it easier to cancel out external light sources.

This is pretty simple and for the most part this works pretty well. The
problem is the light source is drifting with temperature.

I have a few bad ideas, they are all physically cumbersome, such as a
beam splitter and an external monitor diode. Or a heater/thermister to
maintain a constant temperature just above ambient.

Ideas anyone?

Thanks

Ethan

P.S. I am realizing I don't know as much about VCSELs and other lasers
as I thought. Any suggestions on good books, or other references?

It's fairly common to control the temperature of a diode laser.
Usually a TEC (thermoelectric cooler) is used to cool it. A heater
would be easier to set up, but you may be shortening the lifetime; if I
did that I'd run the diode at reduced (half?) power (if your
application can still work with less power). Of course, you'll need a
thermistor or other temperature sensor, plus additional circuitry. PID
or PI controllers are available from Wavelength Electronics,
http://www.wavelengthelectronics.com -- their site seems to be down
right now or I'd hunt down the temp controller part # I used to use at
my previous job.

HTH,

Mark

 

[John Devereux]

So it sounds like using the monitor diode is not going to work for me.
Does anybody have suggestions for how to get a tightly controlled
output from a laser?

You can eliminate mode jumps using strong UHF modulation (300+
MHz).The circuit can be quite simple; e.g. I did it with a single
transistor oscillator.

I am using this to measure optical density of a fluid. I shine a laser
into the fluid and measure how much light comes out. I am modulating
the laser at a couple kHz, just to make the detection easier, and to
make it easier to cancel out external light sources.

This is pretty simple and for the most part this works pretty well. The
problem is the light source is drifting with temperature.

I have a few bad ideas, they are all physically cumbersome, such as a
beam splitter and an external monitor diode. Or a heater/thermister to
maintain a constant temperature just above ambient.

Ideas anyone?

Could you use a LED instead? You can get quite narrow beam ones. You
can also get them with internal monitor photodiodes.