Laser driver

[Raj]

I wanted to drive a numb08 laser diode.
I'm thinking of using lm338 to build a supply that'll limit voltage to 4.5v max, and 2.77 A current.
I'll use multiple Tvs diode and Zena diode, right next to laser diode for protection.
I want to ask, is it good enough?
Do I need to change anything?

[Zero999]

A laser diode only needs the current to be limited. The forward voltage will sort itself out. Only one LM338 is required to regulate the current.

The supply voltage is 5V and the LM338 needs 3.75V across it ore more to regulate the current to the desired level. The laser diode could drop as much as 4.8V, so you need an 8.55V supply, in order for it to work.

The overvoltage protection diodes will not do anything and aren't needed.

A simple series resistor would stand more of a chance of working, than the complex circuit you've posted.

[danadak]

If you decide to modulate diode the regulator transient response
may be of concern.


Regards, Dana.

[janoc]

A laser diode only needs the current to be limited. The forward voltage will sort itself out. Only one LM338 is required to regulate the current.

The supply voltage is 5V and the LM338 needs 3.75V across it ore more to regulate the current to the desired level. The laser diode could drop as much as 4.8V, so you need an 8.55V supply, in order for it to work.

The overvoltage protection diodes will not do anything and aren't needed.

A simple series resistor would stand more of a chance of working, than the complex circuit you've posted.

Eeeep! Do not drive a laser diode with a simple resistor! That will work for a short term test but it is a very poor solution. It is the best way how to turn your laser diode into a crappy LED - exceeding the diode current or optical output power even for a very short time (e.g. because of temperature fluctuation) will kill the optical surfaces and it won't lase anymore - and a simple resistor or a constant current driver has no way to prevent this. These are bare laser diodes, you will absolutely need a driver with an optic feedback to limit the output power because that changes both with temperature and with the diodes aging.
 
Moreover, that NUBM08 is a pretty beefy 38W bank, LM338 will be very going very hot if you try to use it as a constant current driver. Furthermore, you will likely need one for each of the diodes (NUBM08 is a bank of 8 lasers), if you put the diodes in series you will need a fairly high voltage supply. 

Have a look here to see how to correctly (and safely) drive semiconductor lasers:
http://www.repairfaq.org/sam/laserdps.htm#dpstoc

[Raj]

 
Moreover, that NUBM08 is a pretty beefy 38W bank, LM338 will be very going very hot if you try to use it as a constant current driver. Furthermore, you will likely need one for each of the diodes (NUBM08 is a bank of 8 lasers), if you put the diodes in series you will need a fairly high voltage supply. 

http://www.repairfaq.org/sam/laserdps.htm#dpstoc

nice point but, I have a single diode. Lol

[janoc]

nice point but, I have a single diode. Lol

Well, that wasn't obvious from what you have posted - the part number and  the datasheet are for the 8 diode bank.

[Zero999]

 
Moreover, that NUBM08 is a pretty beefy 38W bank, LM338 will be very going very hot if you try to use it as a constant current driver. Furthermore, you will likely need one for each of the diodes (NUBM08 is a bank of 8 lasers), if you put the diodes in series you will need a fairly high voltage supply. 

http://www.repairfaq.org/sam/laserdps.htm#dpstoc

nice point but, I have a single diode. Lol

How do you know it's not 8 diodes inside one package, all connected in series?

[Raj]

 
Moreover, that NUBM08 is a pretty beefy 38W bank, LM338 will be very going very hot if you try to use it as a constant current driver. Furthermore, you will likely need one for each of the diodes (NUBM08 is a bank of 8 lasers), if you put the diodes in series you will need a fairly high voltage supply. 

http://www.repairfaq.org/sam/laserdps.htm#dpstoc

nice point but, I have a single diode. Lol

How do you know it's not 8 diodes inside one package, all connected in series?

It's trolling, but nice point...I think, Because that'll lead to multiple focus points instead of one. They can simply increase the size of laser diode junction to make it more powerful Android it'll be more efficient that way.

[Zero999]

 
Moreover, that NUBM08 is a pretty beefy 38W bank, LM338 will be very going very hot if you try to use it as a constant current driver. Furthermore, you will likely need one for each of the diodes (NUBM08 is a bank of 8 lasers), if you put the diodes in series you will need a fairly high voltage supply. 

http://www.repairfaq.org/sam/laserdps.htm#dpstoc

nice point but, I have a single diode. Lol

How do you know it's not 8 diodes inside one package, all connected in series?

It's trolling, but nice point...I think, Because that'll lead to multiple focus points instead of one. They can simply increase the size of laser diode junction to make it more powerful Android it'll be more efficient that way.

Yes, you're right, I forgot this isn't an LED.

[mikerj]

Eeeep! Do not drive a laser diode with a simple resistor! That will work for a short term test but it is a very poor solution. It is the best way how to turn your laser diode into a crappy LED - exceeding the diode current or optical output power even for a very short time (e.g. because of temperature fluctuation) will kill the optical surfaces and it won't lase anymore - and a simple resistor or a constant current driver has no way to prevent this. These are bare laser diodes, you will absolutely need a driver with an optic feedback to limit the output power because that changes both with temperature and with the diodes aging.

This sounds like scaremongering; you can run laser diodes from constant currents perfectly safely with a small amount of care, provided you aren't trying to exact the maximum possible power from them.  Feedback is required if you need to compensate for temperature and ageing effects, but in many applications some variation in output is not a problem.  Thermal protection would make sense for higher power diodes like the one the OP is using however.

What you do need to be careful about is ensuring minimal transients from the driver output during power on/off.  I can remember a technician at work killing a good number of expensive 980nm pump lasers by trying to drive them from an HP bench supply.  It killed them when the output was enabled due to the time taken to switch from CV to CC operation.

I'm still a little confused about the OP's intent, does he just want to drive one of the eight diodes within the NUBM08 array?

[Raj]

Oh, yes I'm trying to drive a diode extracted from separated from the array (the diode is a used one) Right now I'm in the testing phase. If it can cut through materials I want it to, I'll turn it into an engraving machine later.

[janoc]

Oh, yes I'm trying to drive a diode extracted from separated from the array (the diode is a used one) Right now I'm in the testing phase. If it can cut through materials I want it to, I'll turn it into an engraving machine later.

If you have extracted the diode from the block, do not forget to put it in some other heatsinking material or it will cook itself.

This sounds like scaremongering; you can run laser diodes from constant currents perfectly safely with a small amount of care, provided you aren't trying to exact the maximum possible power from them.  Feedback is required if you need to compensate for temperature and ageing effects, but in many applications some variation in output is not a problem.  Thermal protection would make sense for higher power diodes like the one the OP is using however.

Well, the datasheet explicitly mentions the requirement for optical feedback on this one. Considering how powerful the laser is, I would say it is a very good idea to include it. The problem is that the optical power goes up the colder the diode is with the same current (the lasers and LEDs are more efficient when cold). So if you are unlucky, on a cold day the diode will destroy itself because the output power got exceeded for a short moment and ruined the mirrors.

Constant current driving with no optical feedback is only really usable for things like cheap laser pointers.

And yes, you are right about the supply transients - the maximum current must not be exceeded even for very short time (milliseconds!) otherwise the diode will be destroyed.

[janoc]

Oh, yes I'm trying to drive a diode extracted from separated from the array (the diode is a used one) Right now I'm in the testing phase. If it can cut through materials I want it to, I'll turn it into an engraving machine later.

If you have extracted the diode from the block, do not forget to put it in some other heatsinking material or it will cook itself.

This sounds like scaremongering; you can run laser diodes from constant currents perfectly safely with a small amount of care, provided you aren't trying to exact the maximum possible power from them.  Feedback is required if you need to compensate for temperature and ageing effects, but in many applications some variation in output is not a problem.  Thermal protection would make sense for higher power diodes like the one the OP is using however.

Well, the datasheet explicitly mentions the requirement for optical feedback on this one. Considering how powerful the laser is, I would say it is a very good idea to include it. The problem is that the optical power goes up the colder the diode is with the same current (the lasers and LEDs are more efficient when cold). So if you are unlucky, on a cold day the diode will destroy itself because the output power got exceeded for a short moment and ruined the mirrors.

Constant current driving with no optical feedback is only really usable for things like cheap laser pointers.

And yes, you are right about the supply transients - the maximum current must not be exceeded even for very short time (milliseconds!) otherwise the diode will be destroyed.

[jmelson]

Eeeep! Do not drive a laser diode with a simple resistor! That will work for a short term test but it is a very poor solution. It is the best way how to turn your laser diode into a crappy LED

Well, I've been doing exactly that with an old 670 nm laser diode in my photoplotter for the last ** 20 ** years, without damage.  I don't need to push it to the maximum output level (5 mW) so it has some margin.

Jon

[Raj]

Oh well.... Underdriving is a different matter.I ain't gonna do that. What use it will be too me after buying such expensive diode

[IanMacdonald]

Also since you posted in the Beginners section, I think it ought to be added that experimenting with this kind of unit is best left to those with some experience of high power optics. You need highly effective eye protection which precludes the beam or a reflection entering your eye from any angle, for a start. 

[Raj]

I do have a Google for the specified wavelength of od 4+

[janoc]

Eeeep! Do not drive a laser diode with a simple resistor! That will work for a short term test but it is a very poor solution. It is the best way how to turn your laser diode into a crappy LED

Well, I've been doing exactly that with an old 670 nm laser diode in my photoplotter for the last ** 20 ** years, without damage.  I don't need to push it to the maximum output level (5 mW) so it has some margin.

Jon

You do realize we were talking about a laser with 4W+ of optical output power, right? That is going to be cutting stuff. There you don't have a lot of margin between the current at which it starts to lase and current where it blows up already. Your 5mW diode from a laser pen kinda doesn't compare here.

[Raj]

Don't worry guys, that's why I came here, to get my design checked by experts.
And I'll be wearing my glasses while testing it, fitting it into the air.
And when the time comes to install it on CNC, I'll have it in a room, in which I am absent, controlling via camera and motors. And still I'll eat the glasses

[janoc]

Don't worry guys, that's why I came here, to get my design checked by experts.
And I'll be wearing my glasses while testing it, fitting it into the air.
And when the time comes to install it on CNC, I'll have it in a room, in which I am absent, controlling via camera and motors. And still I'll eat the glasses

A much better idea is to build a proper enclosure for the machine, so that any stray reflections cannot escape and with an interlock/safety switch on the doors that cuts the power to the laser when opened.

Those "naked" gantry-style CNC engravers/cutters so popular on AliExpress and elsewhere are an accident waiting to happen.

It also saves you from indigestion from those glasses.

[Raj]

Lol, thanks for the suggestion. I'll add a cut of switch on the door of that room. That way, the whole room will be the enclosure

[janoc]

Lol, thanks for the suggestion. I'll add a cut of switch on the door of that room. That way, the whole room will be the enclosure

  I find your attitude to safety profoundly disturbing.

You may have a pair of glasses (of who knows what quality) and know not to look into the beam but can that same thing be said about your family members, random onlookers, family pets, etc. that may be curious about what you are doing and walk in the room while the device is on?

Not to mention the possibility of a mechanical failure (e.g. the laser mount giving way or that door switch not working) or a stray specular reflection from whatever you are trying to cut catching you out.

[Raj]

Thanks, I already have warning lights at the entrance of the room through which you gain access to the room in which, there's gonna be a laser

[quadzillatech]

why do these threads get so far off track without any proper answers??,makes it a waste of time to read!.

[janoc]

why do these threads get so far off track without any proper answers??,makes it a waste of time to read!.

You do realize this is discussion forum not a textbook, right? And also you are about 2 years late to the game ... (*points at the date of the last post ...*)

[Raj]

why do these threads get so far off track without any proper answers??,makes it a waste of time to read!.

You do realize this is discussion forum not a textbook, right? And also you are about 2 years late to the game ... (*points at the date of the last post ...*)

Wow! two years already!.
Man, I need to speed things up...I still use the breadboard prototype as if it were permanent fixture