Using a current output DAC to control laser diode brightness: which IC to use?

[phennessey]

I'm looking to control the brightness of a 50mW green laser diode (something like this: https://www.digikey.com/en/products/detail/ams-osram-usa-inc/PLT5-520FB-P/20534487) using an 8-bit current-output DAC, very fast (500ns between pulses).

My question is: how do I pick the right IC for this? I am struggling with how to properly read the data sheets (I also am a bit clueless about how to drive laser diodes. They have a feedback pin, but I'm not sure how I'm supposed to use it in practice.)

Here's one example I found on DigiKey: https://www.ti.com/lit/gpn/dac0808. The datasheet says "The DAC0808 is an 8-bit monolithic digital-to-analog converter (DAC) featuring a full scale output current settling time of 150 ns while dissipating only 33 mW with ±5V supplies." Does this mean it outputs 33mW when the data pins are set to 11111111? As you can see, I'm a bit clueless here...

If anyone is willing to hold my hand a bit and help me pick out a through-hole IC that would work, I would be most grateful. Bonus points if you are willing and able to help me plan the circuit for the laser diode.

[RoGeorge]

LASER diodes are not like LEDs.  They are not made to be powered by constant current.  To get nominal power you'll need to use the optical feedback photodiode, to control the right amount of power (in a closed feedback loop, usually there are dedicated LASER drivers ICs, each for specific applications).  That's why the high power laser diodes have 3 pins, and not just 2, the 3rd terminal is from the photodiode.  In the same capsule there are two diodes, one is the LASER diode, the other is a photodiode that reads back the optical power.

Also, many LASER diodes won't lase at all (might even get damaged) if the applied power is too small, but I didn't look careful at the model you plan to use.

Not to say, 50mW can get you (or others) permanent eye damage just from the reflected beam, without looking directly into the LASER diode.  https://www.lasersafetyfacts.com/laserclasses.html

Another thing, a 50mW LASER diode is not capped at 50mW.  If driven incorrectly, it can generate way more than only 50mW, so during prototyping/testing, think of it as if it were a much higher power device than those already dangerous 50mW (because of potential hardware or software bugs that might accidentally overdrive the LASER).

Even more dangerous if the green LASER is in fact an infra-red LASER diode with a frequency doubler crystal in front of it (to turn the incoming infra-red light into green light).  Those can emit considerable levels of invisible IR light, too, so not only dangerous green, but dangerous invisible IR, too.

[moffy]

This article might help fill in some detail: https://www.analog.com/en/resources/technical-articles/methods-of-controlling-laser-drivers-pots-and-dacs.html
But as RoGeorge suggests, the laser diode can use the integral photo diode to maintain a controlled optical power out.

[phennessey]

Thanks. I am well aware of how dangerous a 50mW green laser is.

My question was about picking out a current-output DAC IC. Holding out hope that someone will help me with that part of the question. I am trying to attenuate its brightness quickly using an 8 bit signal, and I am hoping for some help on that!

[RoGeorge]

What you are trying to do does not work as you expect, there is a lase threshold, which means you can not reduce the current too much.  Also you shouldn't make it too big either.  How big is too big depends with the die temperature, and varies quite a lot, that is why the need for a feedback photodiode.

Driving LASER diodes is not trivial, not for beginners.  My advice would be to find something else to build. 

The question about which DAC to choose does not make much sense, unless you first post which LASER driver you want to use.  The driver is another circuit, not the DAC.  It depends of the application.  Why the short pulses you mentioned?  What is the end use for the green LASER?

[ajb]

I'm looking to control the brightness of a 50mW green laser diode (something like this: https://www.digikey.com/en/products/detail/ams-osram-usa-inc/PLT5-520FB-P/20534487) using an 8-bit current-output DAC, very fast (500ns between pulses).

A 'current output' DAC doesn't mean that you can use it to directly operate a current-driven load like a laser diode.  The DAC0808 can only supply about 4mA max, so you would need to have some additional circuitry to amplify that control signal to the operating current range you want.  In other words, you still need a diode driver.  A basic linear constant current driver isn't terribly complicated, but your desired pulse rate of 2MHz makes it a lot less trivial.  If you want well defined pulses, then the driver needs to provide a loop bandwidth much greater than 2MHz, and cannot overshoot.  There are other driver types better suited to pulsed operation, but are more complicated to implement.  You may also need to provide a subthreshold DC current plus the pulsed current to get sharp edges on the optical pulses, at the expense of some continuous light output.  This is all highly dependent on the application requirements.  No offense, but if you are struggling to interpret the DAC datasheet then I think you will struggle a lot more before getting good results on this project.  What are you trying to do?

LASER diodes are not like LEDs.  They are not made to be powered by constant current.  To get nominal power you'll need to use the optical feedback photodiode, to control the right amount of power (in a closed feedback loop, usually there are dedicated LASER drivers ICs, each for specific applications).  That's why the high power laser diodes have 3 pins, and not just 2, the 3rd terminal is from the photodiode.  In the same capsule there are two diodes, one is the LASER diode, the other is a photodiode that reads back the optical power.

This is highly dependent on the application and the specific laser diodes used.  Many modern diodes, including quite high power ones, do not copackage a photodiode and can be controlled perfectly well (for many applications) with constant current.  Optical power does vary noticeably with temperature, so thermal management is important if the exact output power is critical, or if running near maximum thermal power. 

Even more dangerous if the green LASER is in fact an infra-red LASER diode with a frequency doubler crystal in front of it (to turn the incoming infra-red light into green light).  Those can emit considerable levels of invisible IR light, too, so not only dangerous green, but dangerous invisible IR, too.

 

This is a good thing to be aware of, but the OP has specified a 520nm diode, which will be a direct injection type.  Doubled YAG would be 532nm (plus 1064nm and 808nm depending on how well aligned/filtered it is).  The latter may be extra challenging to operate in pulsed mode due to nonlinear and temperature sensitive effects in the doubling system. 

[phennessey]

1. I need to EITHER modulate the BRIGHTNESS of a laser diode from zero (no light) to an arbitrary brightness (not sure what that is yet)
OR I need to modulate the FREQUENCY of a laser diode (OR modulate the emitted light) between either 525–550nm or between 375–400nm (whichever is easier/cheaper)
2. The diode wavelength must be in the range of 375—550nm (525nm is preferred if the brightness method is used)
3. I need to modulate it accurately and reliably at a rate of around 2MHz, with adjustment time between pulses no more than 100ns.
4. I need to modulate it with 256 levels of brightness/wavelength (8-bits)
5. I need to be able to try a different power diode in case the max brightness isn't bright enough

These are the non-negotiable design specs. I'm hesitant to say more than that because this is a potential product idea that could become commercialized. If someone wants to sign an NDA and contact me privately, I would be more than happy to discuss!

I realize I'm in over my head. I've decided that will not stop me, but I know full well that I will have to rely on the expertise of others. If you're not able or willing to help with these restrictions, there is no need to reply. Keep in mind that I have every intention of learning as much as I need to in order to accomplish my goal. I'm not giving up.

[Buriedcode]

I'm unsure why you would need to module the wavelength of the light emitted - but that, AFAIK is impossible with diodes, they are fixed wavelength.

The on you linked appears to be a single diode - not DPSS IR 1064nm driving a frequency doubling crystal to 532nm - so you shouldn't require any IR filtering afterwards.
You would however require optics if you want to collimate the beam since all bare laser diodes output a divergent beam with an oval shape.

Modulating the brightness in 256 steps is cetainly possible, but you also mention modulating the amplitude, which just brings more questions: is this on/off 100% AM (amplitude modulation) or is it switched between two different levels? Is the modulation square, sine or triangle wave? AFAIK, LD's can be modulated at > 100MHz, but I don't know if dedicated IC's allow for this.

As has been stated by the other replies, what seems like quite a straight forward job isn't so straightforward.  Unlike LED's, LDs are pretty delicate, and require careful feedback design as their output can vary significantly with temperature (which increases with on time).  I second (or third?) the recommendation for an IC thats dedicated to driving LDs - that accepts external control allowing you to set the brightness and modulation with whatever means you have (DAC, PWM etc..).

Again, I cannot think of an application where one would need to modulate the amplitude OR the frequency, the two are very different and have different applications.

Make life easier for yourself and seek a laser "module" from a reputable manufacturer, that includes all the optics, and driving electronics, allowing you to control the amplitude and/or modulation externally without having to handle the specifics of the LD yourself.

[phennessey]

I thought modulating the amplitude is the same as modulating the brightness. If that's not the case, then forgive my ignorance here. I only care about modulating either the BRIGHTNESS or WAVELENGTH the light.

I cannot think of an application where one would need to modulate the amplitude OR the frequency

Some specific materials are reactive to changes in brightness OR frequency. Modulating the wavelength would also work for my application. Is there a way to modulate the light frequency AFTER it is emitted from the diode? I'm not sure, but that's why I asked. That might be easier than trying to modulate the diode itself.

I don't know how else to explain this brightness modulation — it needs to be set to 1 of 256 levels of brightness. I can't use PWM — it must be constant brightness at each instant, with a very short fall/rise time (ideally 50ns) between brightness levels.

Now, knowing all this, does that change your opinion on what direction I should take? I am trying NOT to have to buy a $4000 piece of equipment here. My goal is to do this with relatively cheap off-the-shelf parts.

I second (or third?) the recommendation for an IC thats dedicated to driving LDs - that accepts external control allowing you to set the brightness and modulation with whatever means you have (DAC, PWM etc..).

Great! Can you point me to such an IC, or help give me some specific search terms that would send me in the right direction? Perhaps this? Or this?

[ajb]

I thought modulating the amplitude is the same as modulating the brightness.

Since we're discussing modulation, it's important to be clear whether we're talking about minimum, maximum, average, or difference between min and max.  'Amplitude' of a modulated signal in this case could refer to the difference between min and max, which may not be the same as maximum or average.

So you want the output to go between zero light and 'some' light every 500ns, and you want the 'some' level to be controlled by the DAC, and you want the rise and fall times to be as fast as possible, is that right?

The need for zero light output between pulses will somewhat limit your rise/fall times because it takes some amount of time to build up enough energy in the laser cavity to go from spontaneous emission (below the lasing threshold) to stimulated emission (actual lasing).  This is why the lasers used in high speed applications, like telecom, aren't turned completely off between pulses -- keeping the diode just below the lasing threshold makes it much easier to get sharp rising edges because most of the energy is already there.  Offhand I don't know if you'll really be limited by those factors, my experience with laser diode modulation is at lower frequencies.  Typical constant current drivers may also be a bit challenged here, but getting stable rise and fall times with no overshoot at these timescales will require some careful design anyway. 

Is there a way to modulate the light frequency AFTER it is emitted from the diode?

Technically yes, tunable lasers are available, or you could use a wide band source and a tunable filter.  But those are more complicated/expensive than just modulating the intensity.

Great! Can you point me to such an IC, or help give me some specific search terms that would send me in the right direction? Perhaps this? Or this?

That link is kind of basic, but does at least show a simple DAC-controlled programmable current source.  You might look at laser driver ICs or appnotes designed for telecom usage, as those will be the closest to your operating requirements.  But even a 'ready made' IC is going to require some careful design and selection of the components and circuitry around it to achieve stable and consistent operation.  If you really want to pursue this project, be prepared to start with something lower performance and work your way up to the performance you're aiming for.  You will blow up fewer diodes that way.

[phennessey]

'Amplitude' of a modulated signal in this case could refer to the difference between min and max, which may not be the same as maximum or average.

Thanks, that's my bad. I mean modulating the brightness over time.

I don't need the light to be zero between pulses. I just need to modulate the brightness quickly over time. Ideally the brightness changes occur every 500ns, and the rise/fall time is about 100ns.

tunable lasers are available, or you could use a wide band source and a tunable filter.  But those are more complicated/expensive than just modulating the intensity.

Yeah, this is not likely to be the direction worth pursuing. I just wanted to be sure there isn't some super obvious and easy way to modulate light frequency quickly over time in a way that is precisely controllable, preferably not by filtering a wide-band source, but by modifying the emitted light's frequency somehow (diffraction grating?).

This IC looks promising: https://www.laserdiodecontrol.com/shop/100mA-14V-oem-laser-driver-ati

[bson]

1. I need to EITHER modulate the BRIGHTNESS of a laser diode from zero (no light) to an arbitrary brightness (not sure what that is yet)
OR I need to modulate the FREQUENCY of a laser diode (OR modulate the emitted light) between either 525–550nm or between 375–400nm (whichever is easier/cheaper)

You can also modulate the duty cycle rather than the period frequency, which tends to be simpler.

But still, why a period of 1MHz?  You can more easily pulse it at a few kHz and vary the duty. Or even lower frequency - usually when dealing with LEDs a few kHz is more than sufficient to avoid visual flickering and strobe effects, but it's not like this should matter with a laser.

[phennessey]

But still, why a period of 1MHz?  You can more easily pulse it at a few kHz and vary the duty.

This laser is being used to expose a light-sensitive material very quickly, so I need the brightness modulation to be real-time, at no less than a 1MHz rate. PWM won't work.

[eliocor]

maybe using an electro-optic modulator?
eg: https://www.thorlabs.com/navigation.cfm?guide_id=2090

[moffy]

But still, why a period of 1MHz?  You can more easily pulse it at a few kHz and vary the duty.

This laser is being used to expose a light-sensitive material very quickly, so I need the brightness modulation to be real-time, at no less than a 1MHz rate. PWM won't work.

Shouldn't be an issue as the diode chosen has a quoted 100MHz BW. The question is how quickly do you need to change the modulation value and do you want to do it analogue or digitally, I guess the presumption is digitally since you are asking about a DAC. Is the modulation just switching between two different levels or is the level constantly varying?

[LaserSteve]

iC-Haus makes off the shelf diode driver chips with the required bandwidth. So does Analog Devices.

The diode in question requires a current limited constant current source. It will faithfully follow a waveform, no need for light feedback.

 Do not "spike" the diode by trying to drive it with a current limited bench power supply.  Most constant current bench PSUs in labs overshoot like crazy on a microsecond scale.

The preferred hobby level test driver is an LM317 current source made with a metal film resistor. Lm317 has a slow enough start-up as to be very safe as a LD test driver for CW work.
Do not use a potentiometer with the LM317, the wiper "lifts" as you turn it.


The LD is sensitive to static discharge, it will need heatsinking,  and collimating optics.

Laser hobbyists buy collimators from DTRs Laser Shop, Meredith Instruments in US  or Roithener Laserteknic in Europe. Get at least a glass lens one.

If the module your building will be handled by humans,  install a Lasorb across the diode within 4 inches of its body. www.lasorb.com and a reverse protection diode in parallel. ESD kills little green diodes.

Laser Diodes drift in wavelength around 0.2 nm per degree C.  The green ones are tunable over +/- 3.5 nm with grating feedback. Which OP does NOT need.

EOMs are expensive, the external modulator of choice would be an AOM, but OP should not need it.

Avoid bright  back reflections into the laser diode.

Buy many extra LDs until you learn how to get the design right. Trust me on that.

50 mW is at the point where you need at least OD4 laser safety glasses,. At least until you learn how to contain the beam.  Thorlabs.com has those, probably in OD6 or OD3 for "Alignment" glasses that allow you to see the beam.  Watch out for "specular" reflections as you learn.  OSRAM has a good laser safety app note, read it!

Optical bench plates from Base-Labs and optical specific parts from Thorlabs or Newport  are a very good idea if precision exposing film.
 
You probably want to learn about first surface mirrors, half inch optical posts and mm1 or km1 mirror mounts.

Single mode green LDs have elliptical beams, down the road you may need correcting prism sets or cylindrical lens sets  from Chinese Laser Show projector parts vendors [30$ a set]

For really precise work you may need to learn about Spatial Filtering, but get the current source working first.

Find a copy of Phil Hobbs book, Electro-optics, Making it All Work.

Your idea of using a DAC for current control will probably rapidly evolve to PWM in most applications, with switching the diode from just below threshold current to some desired level.

I used to use daisy chained 74hc85 magnitude comparators and synchronous counters to generate PWM for brightness control with on/off style AOMs. That way all you had to do was send a Byte to the modulator circuit, no DAC required. it's a neat trick for delivering controlled amounts of light, without struggling to write timer/counter code in a processor.  That will also work with LD driver chips. You can build the same concept in an FPGA, but the discrete parts are far cheaper.

Analog current control is very possible.

Thorlabs Det-10A or their other models  are great fast photodetectors for use with an oscilloscope. Don't shine the laser directly into cameras or photodides. Scattered light or a glass wedge, or AR coated plate as a beam splitter is preferred as a sampler.

LaserSteve

[duak]

If you want a variable wavelength light source, you pretty much have to start with white light, say from a xenon arc lamp.  Dan Gelbart did this in the early 80s with an A-O modulator set up as a variable pitch diffraction grating for a color film image recorder.  See US patent 4,456,338 for the general concept.  By chirping the drive signal to the AOM, a swept spectrum of varying wavelength light can be generated.  Of course, if the drive signal is not varied, the output is monochromatic.

About 20 years ago, I saw a demo of a similar unit where a mems device was used to replace the A-O modulator with pretty much off the shelf parts.  It was interesting to see any number of different wavelegths dialled up as needed.  This was before efficient high brightness LEDs were available so I imagine nowadays the arc lamp wouldn't be needed.

[LaserSteve]

Variable wavelength these days is AOTF, or PCAOM.  You can get 8-10 perfectly overlapping color channels at 80% or greater efficiency on the Polychromtic AOM with a multi-line laser source.  I prefer models by AA E or Gouch and Housego.   Gelbert, common sense genius that he is, was beat by Ted Maiman on the color diffraction with AOM.

For. Gilbert's version you need polarized light.

As you can get a metal halide short arc car headlight and ballast for next to nothing, that sure beats the cost of a lab grade Xenon Arc.

I mention this because the small arc still beats led in radiance..

The PCAOM on my bench tunes in 4 nm steps, used with a mixed-gas ion laser for color display.  I'm told AOTF can get close to 1 nm.

It's a great idea, thanks for mentioning it.

Steve
 

[LaserSteve]

I'm betting the OP is trying to make a holographic optic or a mask, in which case Op probably should stick to the laser for best spot size

Steve

[tggzzz]

As the other posters have indicated, I think using a DAC to control a laser's brightness is unlikely to work.

I suggest the OP might like to read a rather good book, "Building Electro-Optical Systems: Making it all Work" by Phil Hobbs, https://www.electrooptical.net/building-electrooptical-systems/

Phil Hobbs has a good track record. He has built many "weird" electro-optical systems that push the limits, and is referred to many times in The Art of Electronics.