Maker Pro
Maker Pro

Need help building a laser diode driver

Hi all,

I have a project to create a 2 Axis CNC laser engraver using a 2W 445nm blue laser diode. This is all fine and dandy, except I'm not well-versed in laser circuits, or any other power circuits for that matter. I've done a fair share of logic circuits, so I know the basics, but I need some mighty help with this.

What I know so far is that a laser diode demands all the current it can get, so it needs a driver. As the diode heats up, its efficiency decreases calling for an increase in voltage in order to maintain power at the same current level. I know from a few tutorials that this can be achieved using an LM350.

My biggest quandry is how to drive the diode close to its specified power. All of the tuorials I've seen so far call for a 1W diode and biuld for an output of 1W. Seeing as my diode is 2W, how would i get it to perform at this level?

Any help and clarification would be appreciated. Thanks! :)
 
You need a constant current driver.

Power will be Vf * I

Thanks for your quick reply :)

So which factor would power depend on the most in this scenario? Should I be pumping more current constantly, or should I look to maintaining higher voltage to achieve something like 1.8W consistently?

Sorry if i'm ignorant, I'm very new to power electronics :S
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Vf comes from the datasheet for the device. You may also be able to measure it if the device is operated at a low current.

If is probably also specified in the datasheet.

In general though, you set the current higher to get higher power output.

Note that the laser you're talking about is very dangerous. What are you doing to prevent reflections going where you don't want them?
 
Vf comes from the datasheet for the device. You may also be able to measure it if the device is operated at a low current.

If is probably also specified in the datasheet.

In general though, you set the current higher to get higher power output.

Note that the laser you're talking about is very dangerous. What are you doing to prevent reflections going where you don't want them?

Seeing this application, would you recommend an lm350, or something else? Unfortunately, datasheets don't exist for this laser diode since it's proprietary and comes only from projectors.

As far as the reflection considerations, my idea was to place a black box over the entire unit while operational. I'm open to better ideas, however.

Edit: fixed autocorrect
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Presumably you know the power of the device.

Next connect it to a 9v battery via a 1k resistor. It should lase weakly. Measure the voltage across the device. This is your initial estimate of Vf.

Calculate the current using P = Vf.If

Set up a constant current driver for half of If and (after verifying the current) connect it to the laser and measure Vf again. Note that the laser will be dangerous. Without a heat sink it may also quickly overheat.

Using the corrected estimate of Vf, recalculate If and adjust the current source.

You would probably be safe making a yellow, orange, or red box. But be sure to check the transparency at the laser wavelength. You would want it to be well under 1%.
 
Thanks for your quick reply :)

So which factor would power depend on the most in this scenario? Should I be pumping more current constantly, or should I look to maintaining higher voltage to achieve something like 1.8W consistently?

Sorry if i'm ignorant, I'm very new to power electronics :S
You seem to be under the impression that you can control voltage and current separately, you cannot. A constant current driver adjusts the voltage to whatever causes the desired current. That is all there is to it. You choose the current, and the driver chooses the voltage.

Bob
 
Presumably you know the power of the device.

Next connect it to a 9v battery via a 1k resistor. It should lase weakly. Measure the voltage across the device. This is your initial estimate of Vf.

Calculate the current using P = Vf.If

Set up a constant current driver for half of If and (after verifying the current) connect it to the laser and measure Vf again. Note that the laser will be dangerous. Without a heat sink it may also quickly overheat.

Using the corrected estimate of Vf, recalculate If and adjust the current source.

You would probably be safe making a yellow, orange, or red box. But be sure to check the transparency at the laser wavelength. You would want it to be well under 1%.

So far, I've found a post on a forum that outlines the average current limits for M140's. The safe operation cutoff seems to be around 1.8 amps. So I've figured a resistor with a value of 0.75 ohms and 3W of dissipation across the Vout and adjust pins of the LM350 should produce a constant current of 1.667 amps (for the sake of safety of the diode before I know its capabilities). My next question is whether I can use a potentiometer to slowly increase current instead of just dumping current on the diode at once. If possible, what value and dissipation should it be?

Also, I've seen other schematics include a diode and a capacitor. Can you explain why this is?

Looking forward to hearing from you. :)
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
There is no problem increasing the current from zero to the constant value in a single step.

0.75 ohms and 1.7A sounds correct for the LM350. You'll definitely need a heatsink for it. 1.7A is sailing close to the wind if 1.8A is the absolute max. I would ensure the laser module is very well heatsinked, since temperature is a major consideration when you want a long diode life.

With an LM350 is it impractical to use a potentiometer to vary the current as it all flows through the sense resistor. More complex designs use a sense resistor, a comparator and a voltage reference. In these, a potentiometer can be used on the voltage reference side to adjust the reference voltage and therefore the current. (these other bits are inside the LM350)

I haven't seen those other schematics. Perhaps you could post a link to one of them so we can see what you're talking about.
 
There is no problem increasing the current from zero to the constant value in a single step.

0.75 ohms and 1.7A sounds correct for the LM350. You'll definitely need a heatsink for it. 1.7A is sailing close to the wind if 1.8A is the absolute max. I would ensure the laser module is very well heatsinked, since temperature is a major consideration when you want a long diode life.

With an LM350 is it impractical to use a potentiometer to vary the current as it all flows through the sense resistor. More complex designs use a sense resistor, a comparator and a voltage reference. In these, a potentiometer can be used on the voltage reference side to adjust the reference voltage and therefore the current. (these other bits are inside the LM350)

I haven't seen those other schematics. Perhaps you could post a link to one of them so we can see what you're talking about.

As far as 1.667 amps, I picked that as a safety value since most people on the laser forum seem to say that 1.8 amps is the safe cutoff. I've seen people push the diodes past 2 amps and some change, but because its not rated, I'd rather not push it :p

I will be using an arduino to control the xy coordinate system driven by g-code. Do you think it's possible to toggle the power source to the LM350 through the arduino effectively? Also, what input voltage would you reccomend? I've seen anything from 5 volts to 7.3 volts to 12 volts. Does it matter?

Also, I have uploaded a picture of the type of schematic i was talking about. Here, they have a cap and a diode bridging the leads to the LD. I'm not entirely sure what their purpose is.

Thanks as always for your awesome responses :)
 

Attachments

  • F7ELPDOGMFIQ75I.MEDIUM.jpg
    F7ELPDOGMFIQ75I.MEDIUM.jpg
    32.7 KB · Views: 406

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
There is a simple constant current driver in the LED resource which allows the load to be switched by a microcontroller. The best way to adjust power is to vary the current but that will require a more complex circuit. With the one in the resource, you can use PWM to control the average current.

The circuit you have shown will average the current over time. This can be useful if driven by a PWM output as it will result in a more constant (but variable) current. The disadvantage is that the laser cannot be turned on or off quickly.
 
There is a simple constant current driver in the LED resource which allows the load to be switched by a microcontroller. The best way to adjust power is to vary the current but that will require a more complex circuit. With the one in the resource, you can use PWM to control the average current.

The circuit you have shown will average the current over time. This can be useful if driven by a PWM output as it will result in a more constant (but variable) current. The disadvantage is that the laser cannot be turned on or off quickly.

This is basically what i have gathered so far. I'm not sure if it's correct, but hey, thats why I'm here haha

Also, I'm not sure if I was clear enough, but I meant that I was thinking to put the 10 ohm pot in series with R1 and use it to do test scenarios at low power/gradually heat the diode before full operation. My rationale was that 10.75 ohms on Vout would give somewhere close to 150mA, which would probably be under the If of the diode.
 

Attachments

  • 1430265092381-1907794785.jpg
    1430265092381-1907794785.jpg
    123.6 KB · Views: 189
There is a simple constant current driver in the LED resource which allows the load to be switched by a microcontroller. The best way to adjust power is to vary the current but that will require a more complex circuit. With the one in the resource, you can use PWM to control the average current.

The circuit you have shown will average the current over time. This can be useful if driven by a PWM output as it will result in a more constant (but variable) current. The disadvantage is that the laser cannot be turned on or off quickly.

Could you please look over the schematics I drew and uploaded, and tell me if I'm on the right path? Is this a proper way to power such a diode, or am I missing something? Thanks
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
The circuit you have drawn will work however you need to switch power to it. This will require almost as many components as the circuit I recommended.
 
The circuit you have drawn will work however you need to switch power to it. This will require almost as many components as the circuit I recommended.

So, I took another look at the LED resource that you recommended. Is it as simple as popping a MOSFET in between the diode and LM350 driver circuit, or am I being totally ignorant of some important consideration again? From what I've read, MOSFETs can be driven at relatively any voltage, so a microcontroller should be able to handle it?
I feel like such a newbie :S but hey, I guess you have to start somewhere
 
Check out figure 3.5.

Yes, I saw that diagram, but I'm still none the wiser on why I can't do the same with an LM350 circuit... I haven't used MOSFETs before, and I'd like to learn why this isn't possible. Everywhere I've checked defines a mosfet as a high-current-bearing, low voltage switch. Can you please clarify this for me?
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
That circuit uses the mosfet to both switch the load as well as regulate the current.

Sure you could add a mosfet to switch the load and use an LM350 to regulate the current, but it seems a bit superfluous. I think we have a resource describing how to switch a load with a mosfet, so if you just want to add a mosfet as a switch, go for it.

A mosfet is simply a type of transistor where the conductance across two terminals is defined as a function of the voltage across another two terminals. Sure, you can use them as a switch, but they also operate in the "in-between" region where they are not completely on or off. Mosfets need to be chosen for voltage, current, and power dissipation (among other things) and 60V devices allowing tens of amps are pretty easily and cheaply available.
 
That circuit uses the mosfet to both switch the load as well as regulate the current.

Sure you could add a mosfet to switch the load and use an LM350 to regulate the current, but it seems a bit superfluous. I think we have a resource describing how to switch a load with a mosfet, so if you just want to add a mosfet as a switch, go for it.

A mosfet is simply a type of transistor where the conductance across two terminals is defined as a function of the voltage across another two terminals. Sure, you can use them as a switch, but they also operate in the "in-between" region where they are not completely on or off. Mosfets need to be chosen for voltage, current, and power dissipation (among other things) and 60V devices allowing tens of amps are pretty easily and cheaply available.

Right. I did a little bit more research and I think I understand now. However, to modulate the current, does the gate voltage need to change? I'm uder the impression that after the threshold gate voltage, resistance drops off incrementally. If so, I think I'm still going to have to go with the LM/MOSFET combo since I'm using an Arduino's fixed 5V supply unless that can be modulated in some way too. The real challenge now is finding a MOSFET that will have a significantly low Rds(ON) for 5V... I think they're called logic-level MOSFETS... any suggestions? I took a snoop around on Mouser since that's where I'm ordering from, but I couldnt pick anything. Looking forward to hearing from you :)
 
Top