Laser Diode Control

[Pack34]

I've been working on a problem on the side for a while. It works but it's not quite as good as I would have liked.

I'm working with a low-power laser diode (120mW max) and I'm attempting to make the power output as stable as possible. What tends to happen is that the output power can fluctuate a couple mW over about a week. This is with the LD on 24/7.

I've been working with an ic-Haus driver ( http://www.ichaus.de/product/iC-NZP ) that's specifically designed for driving laser diodes. But I'm starting to wonder if this is actually the best approach. To my understanding, most of these types of lasers and drivers are geared at high speeds for use in communication. I'm using the laser in a different application that requires a 100% duty cycle for up to a month at a time.

Would a discrete precision current source be a better solution for something like this?

[mazurov]

Temperature variation could be a factor as well as quality of the current source. Have a look at Linear Tech application notes 89,90 for ideas:

http://cds.linear.com/docs/en/application-note/an89.pdf
http://cds.linear.com/docs/en/application-note/an90f.pdf
 

[Psi]

Yep, i also suspect temperature.

A lot of quality laser modules use a peltier block to regulate the temp.

[T3sl4co1l]

AFAIK, much better to servo current based on a measurement of the emission itself.  It's not a co-pack with a photodiode, is it?

Tim

[ajb]

Are you using photodiode feedback, or just constant current?  Temperature of the diode is important for sure, especially if you're not using optical feedback.  Also note that the reference in the IC you're using has a drift of about 0.5%/C, so its temperature matters as well.  And of course, there's the question of drift and thermal coefficient (and consistency of technique) of whatever you're using to measure the output power.  A relatively low percentage of laser applications really care about a variations of ~1%, so I wouldn't expect to find an integrated solution intended to achieve that.  Building a driver for un-modulated CW operation wouldn't be too difficult, but will take a bit of work to achieve the stability you're after.  Depending on whether you have more time or money to throw at the problem it may be worth looking at scientific laser diode drivers.  Some will incorporate TEC drivers for thermal control as well. 

[rs20]

What tends to happen is that the output power can fluctuate a couple mW over about a week.

How are you measuring this? If you're able to make this measurement, surely you can just hook that measurement into an op-amp driving a FET driving the laser and control the laser accordingly (servo current as mentioned by T3sl4co1l.)

[mathias]

What kind of laser are you using? FP, ECDL, Bragg-grating, ...? (I am now assuming a run of the mill red FP LD)

To get the power stability you want, take care of the following:
1) Limit the temperature, small heatsink (with or without a peltier) should suffice. (Absolute) temperature control of 0.1 deg C is irrelevant, unless you want to prevent mode-hopping etc. Since I assume a FP LD, you have many modes already.
2) Measure the output optical power, use the in-package photodiode (PD) or sample the beam with a beamsplitter and place a PD there. The latter has the advantage of being insensitive to the temperature of the LD, the former suffer from that effect.
3) Adjust the current through the LD with the signal coming from the PD.

As a side note, always always always protect your LD from ESD by a Schottky and a Zener. Always always always.

[Pack34]

Right now I'm working with some PDLD and IPS lasers. These do not have a co-packed photodiode. The more I look into laser diodes the more it seems that anything to provide some feedback would make this much more reliable.

How are you measuring this? If you're able to make this measurement, surely you can just hook that measurement into an op-amp driving a FET driving the laser and control the laser accordingly (servo current as mentioned by T3sl4co1l.)

I'm using a laser power meter. It relys on the laser hitting a target some distance away.

What kind of laser are you using? FP, ECDL, Bragg-grating, ...? (I am now assuming a run of the mill red FP LD)

To get the power stability you want, take care of the following:
1) Limit the temperature, small heatsink (with or without a peltier) should suffice. (Absolute) temperature control of 0.1 deg C is irrelevant, unless you want to prevent mode-hopping etc. Since I assume a FP LD, you have many modes already.
2) Measure the output optical power, use the in-package photodiode (PD) or sample the beam with a beamsplitter and place a PD there. The latter has the advantage of being insensitive to the temperature of the LD, the former suffer from that effect.
3) Adjust the current through the LD with the signal coming from the PD.

As a side note, always always always protect your LD from ESD by a Schottky and a Zener. Always always always.

The diode is heat sinked with a TEC and there isn't an in-package PD.

Tomorrow I'll setup the test again with some thermocouple to see if the issue is temperature drift.

[chris_leyson]

Most certainly temerature drift due to self heating. If you drive the laser with constant current expect the output power to drop as the laser warms up.

[Pack34]

Just to chime back on this, it definitely seems directly correlated to the temperature of the laser diode. So I've been looking into how to stabilize temperature as much as possible.

I did notice something after I started re-working the heat sinking. The driver for the TEC was directly on the opposite side of the chassis from the hot side of the TEC. I'm wondering if that was the cause of some of the issues I've seen with it. The TEC driver failing when it got too hot.

I've reworked the thermals and TEC feedback, now to let it sit for 10 days to see how stable it is.

[Eng_hassan85]

Great Topic .. Good Luck

[jeremy]

Is the laser located in an air-conditioned environment? One thing I noticed once was that because the air-conditioning in the particular room I was using automatically switched off overnight is that there was a drift event right as the air conditioning turned on in the morning. Depending on exactly what power meter interface you are using (fiber coupled, etc), make sure that the hot air from your instruments is not passing over any of your lens holders/mirror stands, etc as the thermal expansion of the metal can misalign your system.

[collingtech]

any draft for the circuit ? and how the status?

[_Wim_]

Interested in an update also. Thinking about trying something comparable, so and tips would be more than welcome...

[grouchobyte]

@pack34

I have used ICHaus before. Good stuff. Unfortnately for my applications they never seem to be sufficient in one respect or another. If not restricted, can you provide the laser you are using and some details(schematic) of your drive circuit. It seems from your description that you may have a thermal issue. This is quite common. Below are some generic comments, not necessarily applicable to you.


Both APC and/or thermal control via a peltier and(TEC) controller are often  necessary to keep laser power and wavelength stable. But not always. It seems you have quite a variation in power and many things can contribute to that. Constant current drive without some sort of thermal control will lead to power drift

If you are using the internal back facet monitor diode in the laser package as the feedback element in your control loop, keep in mind that it is only monitoring the laser light in a spatially optimal location, with some possible  error in tracking light vs photodiode current as result of material drift or CTE

Also, look for error or hystersis in your control loop and possible circuit tempco errors(unlikely)

The laser will have a negative tempco power vs temperature. An increase in temperature decreases output power. The wavelength also shifts but in the opposite direction vs temp. If the change in output power is slow it is probably thermal. Use the time constant as a clue in a thermal inertia assumption.

Thermally, keep the self heating and ambient/case temperature stable.  A Peltier cooler and control circuit might make things worse if not implemented correctly. Beware.

External light monitoring is fraught with all kinds of problems as well since there are tons of places for drift and error to occur. Trust me on this one.

Validate your test gear and/or test setup.

Also, don't trust your laser vendor. Specs and the actual device can vary. Many lasers are hand fabricated. Lasers are not 2N3904s.

Good luck. I have spend 32 years mucking around with lasers, so this is an easy one for me.

[_Wim_]

@grouchobyte

Which laser vendor would you recommend? I would need a low power (< 50 mW) near infra-red laser source that is really stable in temperature. At the moment I am looking at Thorlabs products (https://www.thorlabs.de/newgrouppage9.cfm?objectgroup_id=4737), and would prefer a package with collimation optics already included (at least for the proof of concept) like a pigtail of butterfly style and of course internally mounted photo diode. But I find prices quite high for these kind of packages, especially so because this will be my first laser project, and this could cause some casualties during testing...

The power should be stable preferably within 0.5%. Wavelength stability is less important. The laser will be used in a turbidity (transmittance/absorption setup) measurement of a liquid. Detector will be a photodiode. Connection between the process and my laser/detector  I would prefer through fiber optic cables (hence the reason for the butterfly/pigtail style of package).
From the explanation of the OP so far, I am guessing he is working on something similar. If not, sorry for high jacking this thread and I will start my own thread if desired…

[grouchobyte]

@grouchobyte

Which laser vendor would you recommend?

That would depend, but I wouldn't make that a priority in the selection process unless some other paramaters are nailed down first

I would need a low power (< 50 mW) near infra-red laser source that is really stable in temperature.

Near IR covers a pretty wide range....specifically what wavelength?

At the moment I am looking at Thorlabs products (https://www.thorlabs.de/newgrouppage9.cfm?objectgroup_id=4737), and would prefer a package with collimation optics already included (at least for the proof of concept) like a pigtail of butterfly style and of course internally mounted photo diode.

Multimode or single mode fiber pigtailed? How do you plan on terminating the far end?


 But I find prices quite high for these kind of packages, especially so because this will be my first laser project, and this could cause some casualties during testing...

True, hopefully laser safety precautions are being taken. You are in Class III territory.

The power should be stable preferably within 0.5%. Wavelength stability is less important. The laser will be used in a turbidity (transmittance/absorption setup) measurement of a liquid. Detector will be a photodiode.

Ge, InGaAs or Si? Active area?

 Connection between the process and my laser/detector  I would prefer through fiber optic cables (hence the reason for the butterfly/pigtail style of package).


Agreed. Thermally a butterfly pkg is way more stable but it does drive cost up.

From the explanation of the OP so far, I am guessing he is working on something similar. If not, sorry for high jacking this thread and I will start my own thread if desired…

No problem. Glad to help. FYI: I am laid up  recovering after breaking my shoulder so my writing on an ipad may be cryptic and somewhat truncated or incoherent (excuse the pun)

[_Wim_]

Hi, thanks for your reply.

- Near IR covers a pretty wide range....specifically what wavelength? => 830nm
- Multimode or single mode fiber pigtailed? => I was under the impression that for my application this would not matter much (short distance, no datatransfer), but was going to buy single mode
- True, hopefully laser safety precautions are being taken => that is the goal, but as a beginner is this area, something can always go wrong. I was thing zener diode/series of schotkey diode and an RC network.
- Ge, InGaAs or Si? Active area? => Si (based on cost), and the larger the active area the better I would think (again limited by cost). I do not need high speed (at all). Are there any benefits in noise/dynamic range/stability to opt for InGaAs or Ge?
- Recovering after breaking my shoulder => wow, sounds painful, hope you recover soon. Explanation was not cryptic at all, very clear.

[_Wim_]

Forgot to answer one question:
Termination of far end: I was planning to use something line this (grin lens): http://www.thorlabs.de/newgrouppage9.cfm?objectgroup_ID=1340

[grouchobyte]

Okay, great.

1. 830 nm ....lots of laser vendors  but not the cheaper telecom ones.....check out Roithner. You are in the EU, I presume. Your flag suggests that.

http://www.roithner-laser.com

2. Multimode would be better simply because launching 830nm light into a single mode fiber is challenging from a optics standpoint and so the cost may be higher

3. If you are talking about damage to the laser diode itself, a fast schottky diode across the pn junction of the LD will suffice in preventing reverse spikes. Zeners? i haven't used a zener since 1979. Giggling makes my shoulder hurt.

4 Ge covers a wide wavelenth range but is noisy. InGaAs wont cover you at 830nm (well barely with extended InGaAs). Si seems like the best material for the detector. I agree.

Best

Grouchobyte

[_Wim_]

Zeners? i haven't used a zener since 1979

I have, and I was only born that year... 

Thanks for the excellent feedback.

[grouchobyte]

Zeners? i haven't used a zener since 1979

I have, and I was only born that year... 

Thanks for the excellent feedback.

When you mentioned "casualities" I had imagined real people getting hurt. Sorry, it must be the drugs I am on

Yes, lasers are very easily zapped by circuits and the stupid people who use them and design them. As a previous poster had earlier warned.....YES, use every means possible to prevent laser damage including the use of zeners, snubbers, limiters and fuses (just joking on this one). Always, Always Always...... Better safe than sorry with lasers at potentially KiloEuros each, ouch

[LaserSteve]

For the spikes, Google "Lasorb" by Pangolin.  I'm a big fan of this device and I've been working with diode lasers for 12 years. The red model  will work just fine with IR lasers. It's a ultrafast shunt device based on a mosfet and transistor on the same die.

STEVE

[grouchobyte]

For the spikes, Google "Lasorb" by Pangolin.  I'm a big fan of this device and I've been working with diode lasers for 12 years. The red model  will work just fine with IR lasers. It's a ultrafast shunt device based on a mosfet and transistor on the same die.

STEVE

Thanks Steve! I learned something today from an expert.  Finally after all these years of rolling my own(so to speak)

@grouchobyte

[LaserSteve]

Grouchobyte..  I know the inventor personally.   I've seen the hand-held Tesla/Oudin coil demo, that you probably saw in the video. That thing switches in less then 800 picoseconds, limited only by lead inductance..  Worth every penny, and shipping inside the package of many telecom lasers...
Glad you like it. Bill solved a serious problem with that one...

Steve