Current limiting resistor for laser diodes?

[Winston]

I've seen nothing but constant current drivers for laser diodes online, but I have never seen an explanation of why this is required. Could an appropriate current limiting resistor be used instead for short duration (1-3 second) bursts with long (30 minutes or more) cool-down intervals in between?

I want to possibly use a near-IR laser diode as a thin fishing line cutter, that line being used to reef a small hobby rocket parachute until the rocket is closer to the ground and since everything will be homemade including the PCB and must be as small and light as possible, I need an absolute minimum parts count.

[wolf32d]

The output light intensity of a laser diode is proportional to the forward current. Constant current drivers are therefore the most useful power supplies for lasers.
Nevertheless a current limiting resistor will do the job as it does for LEDs. If the resistor limits the current to a low enough value you can operate the laser continuously.

[Winston]

The output light intensity of a laser diode is proportional to the forward current. Constant current drivers are therefore the most useful power supplies for lasers.
Nevertheless a current limiting resistor will do the job as it does for LEDs.

Then I still don't understand the need for a constant current supply. If the required operating voltage and max allowable current draw of the laser diode is known, why can't just a simple resistor be used just as with an LED? What benefit does a constant current supply provide?

[bktemp]

If the required operating voltage and max allowable current draw of the laser diode is known, why can't just a simple resistor be used just as with an LED? What benefit does a constant current supply provide?

The max current is often not known, but only the maximum output power. Each laser diodes needs a different current to produce a the required output power.
The benefit of a constant current source is, it produces a more stable output power than a resistor and you can adjust the current much easier using a control voltage than adjusting a low value resistor.

[wolf32d]

What benefit does a constant current supply provide?

The output light / injected current ratio doesn't depend on temperature. On the other hand the I-V characteristic of the laser diode does depend on temperature. During operation the laser diode gets hot and the forward voltage tends to drop a bit. If you need a stable light source (e.g in photonic circuits) you need a constant current source that compensates any changes in the forward voltage drop.

In your case the intensity stability shouldn't be a big deal though.


EDIT: actually the output light / injected current ratio does depend a little on temperature. That's why high end laser drivers have a monitor photodiode in the feedback loop that provides a measurement of the actual output light intensity.

[Winston]

Thanks to all for the explanations.

I guess that besides achieving optimal, steady output at varying voltages with die temperatures than can reach high values judging from laser diode and high power LED packaging, it's a matter of protecting a relatively expensive component, with laser diodes being a VERY expensive component if not buying Chinese stuff on eBay. For very short duration use not requiring optimal, steady output, while also considering the low cost of the driven device in this case, nothing more than a conservative value of current limiting resistor is needed.

Now that I know that laser diodes have similar requirements to LEDs thanks to the explanations found above, I searched for and found this:

What Type of LED Driver Do I Need? Constant Current vs. Constant Voltage

http://www.ledsupply.com/blog/constant-current-led-drivers-vs-constant-voltage-led-drivers/

[LaserSteve]

Laser diodes are amazingly sensitive to ESD and surges.  The diode's central  structure  itself often does not fail, but the combination of high current density and high internal optical flux from a turn on surge or overdrive blows one or the other cleaved  end face off the diode. These cleaved "facets"   serve as the optical cavity resonator mirrors of the laser. If you damage the laser, it usually becomes a dim LED. This is a very common failure mode.   Keep in mind the density  of photons inside the diode is much, much greater then what emerges.   

This has a name in the business, COD, for Catastrophic Optical Damage, which usually occurs in the nanosecond to microsecond time scale. 

  The fact is, most DC power supplies have a very undefined startup, often with spikes and surges, that cause the simple limiting resistor to be a BAD idea.  In fact most current limited bench power supplies under 300$ in price readily spike LDs in constant current mode.

So the diode driver deals with the exponential portion of the  LD current curve, prevent spikes, provides a soft start, and limits overcurrent.  Most cheap diode drivers depend on the fact that the current source op-amp has a long startup time for its  internal current sources, no matter what.  This provides an inherent soft start, but only if using a single supply rail.

Yes, you can run a LD like a LED, but unless the DC power supply is PERFECT without startup surges, they quickly die.

    If your going to try this, and its a mission critical phase of your rocket, you should consider some sort of active current limiting, even if its a two transistor, three resistor  active current source.   Then include something like a Lasorb(tm) across the diode to protect it from ESD and overvoltage spikes.

I'm a laser professional. I have a large pile of LDs on my desk at home that have been killed by cheap drivers that disregard startup conditions, and so called "constant current" Ebay bench power supplies that have a 350 uF or so electrolytic cap hanging off their output stages.   In the rare cases when we test large LDs on constant current bench supplies, we place a manually switched shunt across them at startup.

Look, every 5$ to 35$ laser pointer even  has a current source in them, and for good reason. If the Chinese could depend on the internal resistance of the battery and a cheap resistor to   limit the current, they would.   One of the worse things you can do is have moderate to large inductances anywhere near the laser VCC.

You'll see COD sooner or later, even with the most carefully chosen low inductance film resistor.  Most LDs have a current risetime measured in nanoseconds, and it takes them a while to stabilize internally.  Until close to stable, they are very vulnerable to COD.

I'm hesitant to send you to laserpointerforums.com, the signal to noise ratio there is very bad, but you'll find a variety of low cost, light weight, laser diode drivers advertised there. Some of which are of quite dubious quality.   

Three other things:

No cheap, lightweight  circuit is failsafe. If the LD is hot enough to burn fishing line, you need to consider eye safety  and eye protection for those recovering the rocket.
It only takes about 15-20  mW of laser light to get to the low end of the recorded eye damage threshold in a perfect case.  You'll have a lot more then that to get a clean burn of fishing line. For something like this, an experienced laser user, pro or hobbyist, would design a beam containment box or some form of beam stop.

Back reflections of the  laser beam into the laser diode are the second leading cause of diode failures. Make sure your design features adsorptive, not reflective, material around the lens focal point.

Laser safety glasses are far, far, cheaper then blindness.


Steve








 

[jwm_]

Another reason is that LEDs are often run with a lot of headroom compared to laser diodes. Even dirt cheap LEDs can handle 20-30ma yet are generally run at 1-3ma. You can have an extremely ill behaved voltage source and still not come close to hitting the danger zone for leds when they are derated that much.  In general, laser diodes are used when you want them as bright as possible so you run your diode capable of 500ma at 500ma, much less room for voltage variation. Applications with high power LEDs like good flashlights will use the same constant current drivers laser diodes use.

That said, for your application, where the voltage source is presuably a known battery and you are not going for the absolute maximum power, a resistor is probably just fine.

However, there is a much easier solution to cutting a fishing line using a resistor. just wrap the fishing line around a low value resistor and burn it out. Dave even has a video on this:

[Winston]

Laser diodes are amazingly sensitive to ESD and surges.

Bad to hear that confirmed as other mentions of that sensitivity weren't so definitive. The brilliant Chinese shipper pressed the diode leads into white closed cell foam and enclosed them in a a short plastic cylinder with lid that one might use to store contact lenses. Oh, but he did enclose all of that in a proper ESD bag. I'll never know if these units were fried by that stupidity or from their first use.

The fact is, most DC power supplies have a very undefined startup, often with spikes and surges, that cause the simple limiting resistor to be a BAD idea.

I'll be driving these through a SOT23 MOSFET from a single cell lipo.

If your going to try this, and its a mission critical phase of your rocket, you should consider some sort of active current limiting, even if its a two transistor, three resistor  active current source.   Then include something like a Lasorb(tm) across the diode to protect it from ESD and overvoltage spikes.

I'll be testing the circuit over a very large number of cycles using a lipo cell identical to the flight hardware to test reliability even though there will be no serious hazard to anyone if the chute never opens because of the low rocket mass and a certified and approved launch site in the middle of nowhere.

I'm a laser professional.

Yes. I've seen your other posts here and was hoping that you'd see this thread and provide these excellent tips.

No cheap, lightweight  circuit is failsafe. If the LD is hot enough to burn fishing line, you need to consider eye safety  and eye protection for those recovering the rocket.

I'll be the one recovering them and the LD will be enclosed in an opaque case.

It only takes about 15-20  mW of laser light to get to the low end of the recorded eye damage threshold in a perfect case.  You'll have a lot more then that to get a clean burn of fishing line. For something like this, an experienced laser user, pro or hobbyist, would design a beam containment box or some form of beam stop.

I plan to include exactly that even though the LD is enclosed in an opaque box.

Back reflections of the  laser beam into the laser diode are the second leading cause of diode failures. Make sure your design features adsorptive, not reflective, material around the lens focal point.

These particular 808nm near-IR diodes have no optical window nor any collimating optics. I believe they may be intended as green laser pumps(?) from what I've read elsewhere. After a lot of searching online, I found a graphic that shows that the uncollimated output is an ovoid with the long axis in line with the long axis of the die.

Laser safety glasses are far, far, cheaper then blindness.

I only plan to power the LD during testing behind an opaque shield with no reflective paths to my eyes which will be closed prior to each short test with the results being recorded by a video camera with typical sensitivity to IR light (it can clearly see and record IR remote control pulses.) Automated longevity test cycles will be done in an unoccupied room under an opaque cover. I've read elsewhere that 808nm is the worst for eye safety.

[Winston]

However, there is a much easier solution to cutting a fishing line using a resistor. just wrap the fishing line around a low value resistor and burn it out. Dave even has a video on this:

Thanks, I'll check that out. I'd considered resistors driven beyond their power rating, but thought they'd eventually fail because of that at the worst time, like as the rocket is falling from the sky. Also, I'd wanted to pass the fishing line through the reefing line cutter case, so wrapping around the resistor is not possible. However, if the resistor method proves to be mechanically and electrically durable enough, there's no reason the resistor couldn't be mounted external to the case, even being made replaceable by placing it between screw or solder terminals.

Below is a link to my previous posts here related to this effort. The thermal method didn't work out because the resistive heater I found to be otherwise ideal required a two cell lipo to heat adequately and that battery was thicker than I wanted. I need this unit to be as small as possible:

https://www.eevblog.com/forum/projects/heating-element-suggestions-for-line-cutter/msg715168/#msg715168

[salbayeng]

Ditto for the hot wire cutter

[Siwastaja]

This is not black magic, you can calculate everything.

Your simple DIY constant voltage supply with a series resistor IS a constant current supply like any other. The laser doesn't know.

Your supply has accuracy spec just like any other constant current supply on market.

You can easily calculate the maximum worst case current, and if it's within specs, your constant current supply made of CV supply + resistor is just as good as any other.

If you do the math, you'll see that the bigger the proportion between the voltage dropped over the resistor vs. over the load, the better the current regulation. That's why limiting the current with resistor is either inaccurate, or inefficient. But if the efficiency is of no issue, then you can have rather good CC accuracy by dropping enough voltage over the resistor so that Vf(diode) changes over operating conditions become insignificant. This is exactly the same with LEDs; as you may know, you never want to drive a 20mA, 2V LED with a 2.1V supply and 5 ohm resistor; you do it with a 3V supply and 50 ohm resistor instead.

CV supply + series resistor is one of the easiest types of CC drives to analyze, which makes it rather safe to design. It's extremely easy to calculate the worst case surge current from I = U/R. This is much more difficult to guarantee when your CC circuitry has active parts. CV supply + resistor is also utterly trivial to drive on/off with a single MOSFET, without the risk of repetitive surges.

This is why I would advice to simply use the resistor, if the alternative is to build an active supply that is too difficult for the DIYer to analyze. Even a total beginner can analyze the resistor solution and guarantee the maximum current!

It's naturally very inefficient, because you need to drop quite a bit over the resistor; laser diodes require more attention to the current than LEDs. An active linear circuit could make this drop-out smaller, and a switch mode converter would almost eliminate it.

Surge / ESD / reverse voltage protections are important with laser diodes, but it's completely different subject from how the constant current operation is implemented. These protections are required regardless.

The magic in a commercial, expensive "laser driver" is not in constant current; it's in protections and robustness, and possibly efficiency in some cases.

A simple 7812 or 7815 regulator with a big output cap, with the resistor AFTER the cap, is the simplest and most robust constant current source for a beginner. With any other CC regulator, you must know the dynamics. As LaserSteve mentioned, CC lab supplies are not usually good; they are simply not meant for fast current limiting, but easily produce current surges.

[RoGeorge]

AFAIK LASER diodes usually have 3 pins (or more), and in the same case a LASER diode and a feedback photodiode are packed together.

In order to drive the LASER diode safe at it's maximum power, a constant current is not enough. They are very prone to thermal failure. Optical feedback from the phtodiode is a must, otherwise you risk to damage the LASER diode. The constant current value must be continuously adjusted according to the optical feedback from the photodiode.

[LaserSteve]

      Um, surprisingly the best simple cheap, driver with the least failure rate is a LM317 configured as a simple constant current source with an external metal film or wire wound resistor.   They have an inherent soft start/slow startup  characteristic, but not by design.  Do not fall into the trap of using an adjustment  potentiometer across the sensing resistor, as the wiper lifts during rotation and the LM317 goes to full current. There are ways around that by designing a resistor network, but that is beyond the scope of this post.

    You need between a 3 and 6 mm focal length lens, to relay the image of the diode die onto the fishing line.  You could go as long as 10 mm and be fine, but the working distance will get impractical.  You'll find a good source of diodes/lenses  from a guy named DTR...  He at least uses astat packing.  Most diodes ship with a shunt across the leads or high grade Astat packing for a reason.  I'll slide a CMOS chip or a PIC across a table bare in static season here in Ohio. But for LDs I get out a static strap and control the humidity.  Once mounted in circuit, they get fairly robust with only minor shielding.  Depending on the diode, they often have anode or cathode tied to case, newer high power diodes have a floating case. That can influence your design, as well.

High power diodes rarely have the built in photodiode, because the power off the back side of the laser would FRY most tiny photodiodes, and the market calls for regulating the power downstream in the process, with current sources controlled by external light feedback if needed.

808 nm diodes are for DPSS laser pumps, using Nd:YAG and ND:Vandate, which both have a series of adsorption bands at 805-809 nm. Quite a few get used with fiber coupling for medical devices.

Beware the naysayers who say just use a resistor. They  may not have much experience with LD ESD and COD, or the fact that some LDs (blue mainly)  have some rather unique multilayer  hetero-junction structures that probably oscillate during startup.      A SMD "317" or LT3080 do not have much mass, and for a short series of one to three "burns" may not need much heatsinking.

I'll disagree with the statement that "Optical Feedback" is always needed.  I have many LDs in Current Only service, but with carefully designed drivers intended for the purpose.  I will agree that systems with optical feedback tend to last longer, but that may only be due to careful design.

There are some US Vendors who will sell you a diode at only twice or so compared to  the Chinese price, I'll have to check my notes when I get home.  I'll PM you.

I cant speak highly enough of having a LASORB across the diode on a flight device in a non conducting rocket body. I know the engineer behind it, and the product is really good at what it does...

http://www.lasorb.com/



This is not "Magic", like RF, it is a subject that requires a higher level of care and understanding.   Yes, in a perfect world, a mere resistor would suffice. However the actual world is a really complex place.  I jokingly say "Life is a Transmission Line". There is a caveat to that, I mean the complex RF variety, although the power delivery version is equally fascinating.


Steve

[Fungus]

I still don't understand the need for a constant current supply. If the required operating voltage and max allowable current draw of the laser diode is known, why can't just a simple resistor be used just as with an LED? What benefit does a constant current supply provide?

a) It lets you run very close to the ideal power. This is hard to achieve with just a resistor because of variance between individual resistors/diodes. Each diode would either have to be matched to a particular resistor or you'd need some sort of trim control. This would add expense.

b) A constant current driver won't vary much with temperature. Both the diode and the resistor could have large temperature coefficients and the output could change as they heat up.

So constant-current has a few important advantages.

OTOH a resistor will work if you're not too fussy about exact power output.

[LaserSteve]

Link for small quantity Lasorb sales site:

https://pangolin.com/orders/lasorb.php

Of course pricing goes down if your a OEM...

I have no interest in the company, other then knowing the owner, and installing the things on every laser I own.  Which has greatly reduced the dead diode pile.
I do use/install enough diode lasers in a given year to see most of the failure modes.

Steve

[mikerj]

This is not black magic, you can calculate everything.

Your simple DIY constant voltage supply with a series resistor IS a constant current supply like any other. The laser doesn't know.

Only if the Vf of the laser diode remained constant, but in real life it changes over temperature.  If the supply voltage and resistor value are both much higher than the diode's Vf then you will approximate a constant current which will be close enough for most purposes.  If your supply voltage is only a little higher than the laser diode's Vf (which infers a low value series resistor), then the current will not be stable over temperature.

[tggzzz]

No cheap, lightweight  circuit is failsafe. If the LD is hot enough to burn fishing line, you need to consider eye safety  and eye protection for those recovering the rocket.

I'll be the one recovering them and the LD will be enclosed in an opaque case.

... if all goes according to plan. Plausible worst case: rocket goes off course, box damaged on impact, picked up by a child and laser fires.

Quite frankly a laser seems a bit of an overcomplicated way to break a small cord. OK for a James Bond film, but inelegant engineering.

If the simple resistor concept turns out to be inadequate, I'd consider making a helix out of a resistance wire and passing the cord through the centre. Apply enough current and it (and the cord) will rapidly get hot!

Alternatively, realise that your requirement is not novel, and people have already developed solutions. Don't reinvent the wheel; you'll only create a square wheel! Here's one homebrew mechanism https://www.apogeerockets.com/education/downloads/Newsletter266.pdf for pyrotechnic bolts in a small rocket.

[lukier]

I cant speak highly enough of having a LASORB across the diode on a flight device in a non conducting rocket body. I know the engineer behind it, and the product is really good at what it does...

Why is this LASORB so magical? What is wrong with a simple TVS ESD suppression diode across the inputs? I used this method on my 445nm 1W & 2W diodes and it seems fine:

http://laserpointerforums.com/f51/inexpensive-esd-static-protection-55979.html

Just curious, because PESD3V3U1UT are super cheap and readily available (e.g. from Farnell), while LASORB is complete opposite.

[ajb]

I cant speak highly enough of having a LASORB across the diode on a flight device in a non conducting rocket body. I know the engineer behind it, and the product is really good at what it does...

Why is this LASORB so magical? What is wrong with a simple TVS ESD suppression diode across the inputs? I used this method on my 445nm 1W & 2W diodes and it seems fine:

Speed and clamping voltage.  How do you get a TVS that will reliably clamp a transient to a level that is safe for the LD, but won't conduct during normal LD operation?  Let alone over a wide range of operating regimes and temperatures, let alone with any decent speed?  The main selling point of the Lasorb is that it's triggered by a fast rising edge rather than at a specific voltage, and it actively clamps the LD voltage during a transient.

It's essentially a MOSFET with a high-pass RC filter at the gate.  Here's the patent for the basic idea, although there are some details that the patent doesn't cover:  https://patents.google.com/patent/US20110110005A1 

[Winston]

Just use a small hobby blade or such like  fixed to a small solenoid / relay arm / servo or such like. Much cheaper, just as small, and comparably light weight.

After a failure to perform as desired of the ceramic enclosed vaping element shown in my previous thread on this linked to above, I went to that concept using a 3.3V micro linear servo (Spektrum 2.9g LINEAR LONG THROW SERVO SPMSH2040L) with an amazing amount of torque (329g+). It works, but the entire unit ended up being larger than I'd wanted. Below is a link to an image of my custom 3D printed case. The hole in the bottom of the left half of the case is for the PCB mounted SMT visible light optimized phototransistor which detects the ejection of the 'chute with the attached reefing device:

https://c2.staticflickr.com/8/7453/26769762884_6630bbb835_o.jpg

Or for the resistor method you can get something like nichrome or whatever 40AWG or 44AWG wire strands that should get plenty hot enough with an ampere or two through them for a half of a second.  You could probably even use lamp filiament from a 3V miniature bulb or something.   If getting higher current at lower voltages helps then just use a small 10:1 or whatever hand wound transformer driven by a push pull or similar arrangement to produce a square wave at 1x or 2x the battery voltage peak-to-peak in the primary at whatever current is suitable for your needs -- 500mA, 1A, 2A, etc.  That would be more than enough current to heat a small wire strand or such within a second.  And along those lines they also have those electro thermal things they use for starting the motors, right, one of those should work.

I've already bought a large number of different small gauges of nichrome and Kanthal wire to come up with an optimal coil design for the very small single cell lipo I'm using, the coil suspended between thin brass hobby tubing at both ends for G-load support with the fishing line passing through the center of the coil.

I've also investigated in Digikey and Mouser parts selectors but not experimented with using over-powered thick metal film SMT resistors mounted above and below the fishing line passing between them.

In both cases, melted plastic fishing line residue buildup caused clogs might be a problem and, ideally, I want this cutting mechanism to be inside the case to physically protect it and to shield it from the cooling windstream from the rapid descent of the rocket with its ejected but not deployed 'chute and reefing device. I don't want to need to be opening the case on a regular basis to clear clogs.

Optical is almost never the answer if you're looking for something that can be done by mechanical means.

From my design considerations and experience so far, I hope it is the answer in this case. We'll see.

HOWEVER, I'm definitely open to any and all ideas for anything that fits ALL of these criteria:

1. Can cut thin fishing line using minimum current which is limited by the use of a very small single cell lipo and cell voltage drop.
2. Highly resistant to failure from G-loads.
3. Minimum chance of clogging due to melted fishing line residue.
4. TINY.

[Winston]

AFAIK LASER diodes usually have 3 pins (or more), and in the same case a LASER diode and a feedback photodiode are packed together.

In order to drive the LASER diode safe at it's maximum power, a constant current is not enough. They are very prone to thermal failure. Optical feedback from the phtodiode is a must, otherwise you risk to damage the LASER diode. The constant current value must be continuously adjusted according to the optical feedback from the photodiode.

The pinout diagram I have for my LDs shows two pins insulated from the case and IDed as + and - with a third pin attached directly to the case and marked "not used." I'll be operating the LD for probably no more than 3 seconds (we'll see), so heating shouldn't be an issue. If possible, I'll also be operating it well below max rated power.

[Winston]

No cheap, lightweight  circuit is failsafe. If the LD is hot enough to burn fishing line, you need to consider eye safety  and eye protection for those recovering the rocket.

I'll be the one recovering them and the LD will be enclosed in an opaque case.

... if all goes according to plan. Plausible worst case: rocket goes off course, box damaged on impact, picked up by a child and laser fires.

Quite frankly a laser seems a bit of an overcomplicated way to break a small cord. OK for a James Bond film, but inelegant engineering.

If the simple resistor concept turns out to be inadequate, I'd consider making a helix out of a resistance wire and passing the cord through the centre. Apply enough current and it (and the cord) will rapidly get hot!

Alternatively, realise that your requirement is not novel, and people have already developed solutions. Don't reinvent the wheel; you'll only create a square wheel! Here's one homebrew mechanism https://www.apogeerockets.com/education/downloads/Newsletter266.pdf for pyrotechnic bolts in a small rocket.

Addressing all of those points, please see my reply to evb149 found below.

[Winston]

      Um, surprisingly the best simple cheap, driver with the least failure rate is a LM317 configured as a simple constant current source with an external metal film or wire wound resistor.   They have an inherent soft start/slow startup  characteristic, but not by design.

An SMT equivalent of that is exactly what I plan to use if this current limiting resistor approach doesn't prove "good enough."

Do not fall into the trap of using an adjustment  potentiometer across the sensing resistor, as the wiper lifts during rotation and the LM317 goes to full current. There are ways around that by designing a resistor network, but that is beyond the scope of this post.

Pots will get nowhere near these tests. Fixed value resistors only will be used.

You need between a 3 and 6 mm focal length lens, to relay the image of the diode die onto the fishing line.  You could go as long as 10 mm and be fine, but the working distance will get impractical.

I hope to avoid optics by placing the line close enough to the LD to be melted without optics, that proximity needed due to the uncollimated beam spread ("7 deg parallel, 30 deg perpendicular"). I'm concerned about plastic vapors fowling or otherwise impairing the LD die within its windowless package. The fishing line will pass in front of the LD suspended between small diameter brass tubing. Natural airflow within the case should be away from the diode due to the location of the charging port but that's just a WAG since the orientation of the reefing device case against the airstream will vary.

808 nm diodes are for DPSS laser pumps, using Nd:YAG and ND:Vandate, which both have a series of adsorption bands at 805-809 nm. Quite a few get used with fiber coupling for medical devices.

Thanks, cool to know!

Beware the naysayers who say just use a resistor. They  may not have much experience with LD ESD and COD, or the fact that some LDs (blue mainly)  have some rather unique multilayer  hetero-junction structures that probably oscillate during startup.  A SMD "317" or LT3080 do not have much mass, and for a short series of one to three "burns" may not need much heatsinking.

I hope I can get away with it this time. The reason I'm so damned component adverse here is because while I can handle tiny, single sided, homemade SMT PCB production, such tiny boards if necessarily double sided due to trace routing issues will be a real PITA.

I'll disagree with the statement that "Optical Feedback" is always needed.  I have many LDs in Current Only service, but with carefully designed drivers intended for the purpose.  I will agree that systems with optical feedback tend to last longer, but that may only be due to careful design.

If this doesn't prove to be a total dead end anyway, I hope to drive the LD significantly under its rated power for a very short duration with a very long cycle time.

There are some US Vendors who will sell you a diode at only twice or so compared to  the Chinese price, I'll have to check my notes when I get home.  I'll PM you.

Thanks, those should be sellers who actually know what ESD protection is all about. I'll be testing these Chinese LDs tomorrow.

I cant speak highly enough of having a LASORB across the diode on a flight device in a non conducting rocket body. I know the engineer behind it, and the product is really good at what it does...

http://www.lasorb.com/

Something I had not thought of. Plus, the 'chutes are nylon and the reefing device case will be PLA. I'll bet conductive paints for the case exterior are expensive. Anyone know? The LASORB devices would be nice but the through hole package is kinda large for my purposes and the SMD kit has a minimum order of 100 and consists of four (more) SMT components.

This is not "Magic", like RF, it is a subject that requires a higher level of care and understanding.   Yes, in a perfect world, a mere resistor would suffice. However the actual world is a really complex place.  I jokingly say "Life is a Transmission Line". There is a caveat to that, I mean the complex RF variety, although the power delivery version is equally fascinating.

Steve

And it is a lot of fun and very educational. I greatly appreciate your comments and those of others here. Collective brainstorming is very helpful. I don't know how many times I've thought, "Now, why in the hell didn't I think of that?"

[Winston]

This is not black magic, you can calculate everything.

Your simple DIY constant voltage supply with a series resistor IS a constant current supply like any other. The laser doesn't know.

Your supply has accuracy spec just like any other constant current supply on market.

You can easily calculate the maximum worst case current, and if it's within specs, your constant current supply made of CV supply + resistor is just as good as any other.

If you do the math, you'll see that the bigger the proportion between the voltage dropped over the resistor vs. over the load, the better the current regulation. That's why limiting the current with resistor is either inaccurate, or inefficient. But if the efficiency is of no issue, then you can have rather good CC accuracy by dropping enough voltage over the resistor so that Vf(diode) changes over operating conditions become insignificant. This is exactly the same with LEDs; as you may know, you never want to drive a 20mA, 2V LED with a 2.1V supply and 5 ohm resistor; you do it with a 3V supply and 50 ohm resistor instead.

CV supply + series resistor is one of the easiest types of CC drives to analyze, which makes it rather safe to design. It's extremely easy to calculate the worst case surge current from I = U/R. This is much more difficult to guarantee when your CC circuitry has active parts. CV supply + resistor is also utterly trivial to drive on/off with a single MOSFET, without the risk of repetitive surges.

This is why I would advice to simply use the resistor, if the alternative is to build an active supply that is too difficult for the DIYer to analyze. Even a total beginner can analyze the resistor solution and guarantee the maximum current!

It's naturally very inefficient, because you need to drop quite a bit over the resistor; laser diodes require more attention to the current than LEDs. An active linear circuit could make this drop-out smaller, and a switch mode converter would almost eliminate it.

Surge / ESD / reverse voltage protections are important with laser diodes, but it's completely different subject from how the constant current operation is implemented. These protections are required regardless.

The magic in a commercial, expensive "laser driver" is not in constant current; it's in protections and robustness, and possibly efficiency in some cases.

A simple 7812 or 7815 regulator with a big output cap, with the resistor AFTER the cap, is the simplest and most robust constant current source for a beginner. With any other CC regulator, you must know the dynamics. As LaserSteve mentioned, CC lab supplies are not usually good; they are simply not meant for fast current limiting, but easily produce current surges.

Thanks. I plan to design for a worst case fully charged cell voltage of 4.1V which will only head very slowly downward from there due to discharge and the rapid, but small, voltage drop due to internal resistance when the LD is activated. If I can get away with it, I'll also be operating the LD significantly under its max rated current. If I can't, I'll just buy a higher power LD and operate it at the full power level of my current LD.