PWM LD Controller Design

This is a help request, not a finished design.

I'm looking for a 405 nm @ 2-5 W, PWM LD driver.

Since I'm totally new to the LPF community, I've been looking at loads of
various designs both off-the-shelf and DIY circuit designs. However, now I
feel even more confused and undecided how to go about this. I would very
much appreciate what the experienced LPF member would recommend.


Background:

I'm planning to build a CNC-like laser cutter/etcher, using a
RaspberryPi-3. What will make this cutter non-usual is that it will
attempt to cut some materials only normally possible with professional
machines, using much higher powers. This will require using the highest
available power laser diodes. (I assume as of today, that these are
rated around 3-5W of optical output power.)

The Rpi3 already have a HW driven PWM built-in on GPIO#1 (header
pin-12), while all other GPIO can also be configured as software PWMs.
Thanks to the PiFM project, we know we can drive the HW driven PWM up to
at least 100 MHz, so I'm considering the possibility of using this. However,
this option may also overload the processor, that also need to control the
CNC steppers...

The project will ultimately be Open Source and published with full DIY
instructions.

So the LD driver design specifications I'm looking at is something along the lines of the following.

Other useful criteria:

All close-to-LD parts are kept close to LD, while other parts of the
driver are kept off the moving CNC board. Connected by moving a light
and flexible 3 lead cable.


Questions:

1) What is better, using the Rpi to generate PWM control signals
(over GPIO) or to use off-the-shelf components to offload MC for
CNC operation?

(The Rpi3 already have a HW driven PWM built-in on GPIO#1 (header
pin-12), while all other GPIO can also be configured as software PWMs.
Thanks to the PiFM project, we know we can drive the HW driven PWM up to
about 100 MHz.)

2) What are the best and cheapest design options using PWM components?
(LM555, 74AC14, MAX3667, HV9925 etc)

3) What design options would keep the number of components to a minimum?

4) How can you best control the PWM driver? (Through a DAC or something else?)

5) What is better for the LD, that the PWM pulse shape is a square-wave, sine-wave or something else?

6) Will the LD always be lasing at lower PWM duty cycles (at high frequencies)?
(Assuming f=1MHz, what is the minimum duty cycle for lasing?)

7) Can you improve LD start-up performance by using a zero/baseline current/voltage offset?
(For example, by letting the LD current idle at some constant but very low value?)


I'll surely have many more questions as the fog clears... But these are already overwhelming for most people.
I would appreciate anything or any suggestions that could help me get started...

The reason I've chosen to use 405 nm LD is simply that:

1) They (and accompanying optics) are readily available (and thus cheaper).
2) They are still visible, so more secure to use.
3) Their light is highly absorbed by the materials I'm considering.

But if the design/price/performance would be improved by using an even shorter wavelength, that would be fine too.

The problem with using a 405nm laser diode is that the single mode ones are limited to 1 watt optical power at best. Multimode diodes have problematic beam profiles as they use a larger emitter that the single mode do. Any laser you get in a lower wavelength would be very expensive and would be cost prohibitive. There are several drivers prefab that would supply the constant current needed to drive a diode to the power you are looking for and are able to be modulated either by TTL or analog. All the diodes that will give you an output of 5 watts have terrible beam profiles. They could only be used at a very close distance unless you are willing to do some major beam shaping. This is probably not what you want to hear, but it is the state of the art.

Aha, that is interesting. Thanks, this is clearly important to know!

So I was reading the following threads:

(1) "Laser Diode Technology": (Nice overview!)
(2) Single- vs multi-mode beam shapes (some pictures)
(3) Single- vs multi-mode laser power divergence

In particular the comparative power calculation in this post from (3)
was interesting. And the online calculator here.

Looking at the following figure in (1):



Not sure how accurate these are, but it seem that as you increase power,
the modes get less spread out. Where two adjacent longitudinal-modes are
calculated as:

Δλ = (λo)^2 / (2nL)

Where λo is the primary lasing mode wavelength, n is
the mode number and L the length of the lasing substrate.

However, for the project I consider, I don't think the multi-mode operation is
a problem, because of two reasons.

1) I'll be focusing to a very narrow spot on close range of not more than 4 cm,
so divergence should not be too bad at this range.
2) I'm looking to transfer energy (burn) so the wavelength spread should not
matter unless it causes lens/focusing problems.

Finally, it's even possible that the multi-mode operation may be of advantage,
because of differences of "substrate" absorption spectra.

Operating in multi-mode does give some problems with focussing, but it depends on your application if that this a problem or not. I'm not sure how small you want the spot to be in yours. You can forget about reading the pits and islands on an optical disc using a multimode laser, but at 4 cm distance i think you could still do pretty fine (say 0.1 mm) wide work with a multimode and proper optics.

The graphs for wavelength distribution you show are, afaik, more or less demonstrating the lasing threshold. At low power, below the treshold, lasers act more or less like leds with a rather broad output spectrum. With single mode diodes this turns into an exact wavelength when well beyond the treshold, while multimodes never really get to that point.

I thought I said pretty much the same thing. Within a short distance and with good optics a MM diode might be okay depending on what you need it to do specifically.

paul1598419 said:

There are several drivers prefab that would supply the constant current needed to drive a diode to the power you are looking for and are able to be modulated either by TTL or analog.

Click to expand...

I know the forum here is full of drivers, and I've looked at several, only to be more indecisive. Could you recommend one to get me started?
I've been tearing out the LD from a bunch of laptop DVD players and like to get started. (I think I've zapped most of them though, since I never took any ESD precautions and I didn't know at the time they were so sensitive.)

Benm said:

You can forget about reading the pits and islands on an optical disc using a multimode laser, but at 4 cm distance i think you could still do pretty fine (say 0.1 mm) wide work with a multimode and proper optics.

Click to expand...

Yeah, the 4 cm was just an arbitrary maximum. I really don't care how close the beam have to be, as long as it can move above the board. I was even thinking to keep all the DVD optics and use it, but that is too close I think, and regardless pointless if I have to use a custom diode.

I'd recommend checking with Lareerer's posts as he has built quite a few of these drivers that everyone here uses. I know he has some high power ones with modulation. I don't have a link at the moment.

SIEVA said:

I was even thinking to keep all the DVD optics and use it, but that is too close I think, and regardless pointless if I have to use a custom diode.

Click to expand...

You don't need all the dvd optics - specifically you don't need all the stuff that is required for reading the discs. In a typical dvd sled the diode is positioned 90 degrees from the disc and goes through a 45 degree mirror. This serves to keep the design compact and is probably of no use to you.

That said a module like the typical 'aixiz' one can have the focus set fairly close from the lens (a few cm depending on actual lens and diode) which could work out well - keep in mind that the 'dot' will be more of an ellipse than a circle with multimode diodes though. This may or may not matter in your application depending on how small you need it to be.