Red laser driver PWM

Hello all,

I'm looking to drive my single mode 638nm diode HL63603TG from DTR. I want to drive it at 300mA and be able to PWM it to dim the output from 0-100% duty cycle or somewhere near that at ~10kHz. I'll be using a single cell Li-ion battery which should output 4.2-3.5V.

I thought of using the LM317 as a constant current source but I'm not sure if the LM317's voltage drop is too great since the red diode needs around 3.2-3.3V to operate.

I also thought about using the X-drive buck which has an enable pin for modulation. However, I don't know the maximum PWM frequency that it allows. I'm looking for around 10kHz. It says here that the input voltage of the x-Drive is

X-Drive V6 input voltage range is 3.4 to 16.0 volts

SXB-2 all versions input voltage range is 2.8 to 6.0 volts

Click to expand...

so I think using a single cell Li-ion battery would be fine.

If anyone has better ideas or if my ideas are flawed please let me know!

Thanks!

I doubt a switching regulator would be overly happy keeping up with 10 kHz.

The LM317 will definitely not work. If set up as a constant current source, it already requires 1.25 volts over the sense resistor, plus whatever it needs to operate itself (could be up to 3 volts depening on current).

Your worst case scenario is 3.5 volts suplly with 3.3 volts output, leaving only 0.2 volts. This is not enough for any constant current source i know off to operate reliably.

One thing you could do is build a constant voltage boost converter, and run a linear solution off its output. This way you can buffer the supply voltage (say 4 or 5 volts) to cope with the switching load, and do the pwm on the linear controller as fast as you like.

Not the most compact or efficient solution, but i think it would be the most reliable.

Thanks for the suggestion!

One thing I'm worried about are start up voltage spikes. Do you know any boost drivers that are ok for laser diodes? Also, do you know of any transistors similar to the LM317 but with an enable pin? I'll keep searching for options.

Apocalypse said:

Thanks for the suggestion!

One thing I'm worried about are start up voltage spikes. Do you know any boost drivers that are ok for laser diodes? Also, do you know of any transistors similar to the LM317 but with an enable pin? I'll keep searching for options.

Click to expand...

IIRC you can add an enable to a LM317 by using a small transistor to short the adj pin to gnd.

A LM317 is not a transistor, it is a voltage regulator IC that comes in a case that is also used by many transistors, but there is a lot more in there.

If you run it als a constant current source you typically connect a resistors between out and adjust that drops 1.25 volts (the lm317 internal reference).

This leaves no other option that putting a transistor across the laser diode (i.e. between adjust and ground), which shorts out the laser diode when you pull its base high. Note that this shuts off the laser diode by creating a short circuit across it, and the LM317 will dissipate -more- power when the laser is off.

Such a setup would be fine for a failsafe that detects a scanning error in a projector, but it's not a good way to go about dimming your laser operated from a battery since it will draw the same amount of current no matter how bright or dim.

Sorry, missed the part about the battery Voltage in the OP. 317 cant be used. The 317 can be modulated with a transistor on the adj pin. As Benm said the power dissipation would be too high for battery operation anyway. I have to get my mind out of projector mode here at LPF.

Switching current regs are hard to design with low noise in mind. Especially with 0-100% modulation. I would also love to see some circuit ideas if anyone has any.

Wonder if I could do it the other way around. Have a constant voltage buck converter before a boost converter here. Pin 3 is the enable pin. LM3410 specs from TI

Input voltage range of 2.7V-5.5V
Output voltage range of 3V-24V

Click to expand...

I found a buck converter LM2830-Q1 which I could set to a voltage of 2.7-2.8V using equation R1=(Vout/Vref-1)*R2 on the data sheet page 15 equation 12 using Vref=0.6V, R1=36k, R2=10k resulting in Vout=2.76V. I can then run it through the boost converter which has an enable pin to boost it to 3.3V for the diode and also modulate it using PWM.

One thing I'm worried about is the start up spike though. Does anyone know the acceptable spike for a laser diode? There are start up wave forms on page 17 of the LM2830-Q1 data sheet. If someone could check it and make sure it is ok for me. It looks ok to me.

You could, but it's not desirable to do it in that order.

Losses are relatively higher when working with the same amount of power at a low voltage (hence high current). This includes losses in switching transistors, inductors, diodes and such.

You can build a circuit that can output a voltage that is higher, equal OR lower than it's input voltage. Common topology for this is a 'sepic' design which combines boost and buck into one circuit (it has 2 inductors and an extra capacitor). Another option would be the 'Cuk' converter, which doesnt have the extra capacitor but can only output negative voltages. The latter would not be a problem when diving a case-isolated laser diode.

Benm said:

You could, but it's not desirable to do it in that order.

Losses are relatively higher when working with the same amount of power at a low voltage (hence high current). This includes losses in switching transistors, inductors, diodes and such.

You can build a circuit that can output a voltage that is higher, equal OR lower than it's input voltage. Common topology for this is a 'sepic' design which combines boost and buck into one circuit (it has 2 inductors and an extra capacitor). Another option would be the 'Cuk' converter, which doesnt have the extra capacitor but can only output negative voltages. The latter would not be a problem when diving a case-isolated laser diode.

Click to expand...

Do you think I can use the LM3410 which is used in the benboost driver and configure it in SEPIC mode (Page 25 on the data sheet) rather than Boost mode?

I don't see any reason why not, it is a textbook sepic topology in that datasheet. It will not fit on the board from that design, but if you design your own circuit board it should be no problem. If you get the part in the SOT-23 package you should be able to hand solder it with some experience.