analog modulating 523nm "Pointer" modules

Hi folks.
i have an bare 100mW green module from Susie here, the ones for "own builds", without heatsink and only a click-on driver. I want to use this for a scanner. now my experience with pointer greenies is rather bad. had a 5mW one tested with current modulation while measuring photo diode response on a oszi... cruel. modehopping 2 times, non linear curve.
this will be even more worse with 100mW i think.

Which possibilities i have to linearise ist besides of using a TEC?
µC with LUT? PID-regulation over photocell behind thin glas plate?
or toss it and buy laserwave one with analog driver? (bad: they are huge)

Best start by trying, so we can at least see what we are up against. Do you have the ability to plot input current vs output power with a few waveforms? 1kz triangle would be a good start.

The combo crystals used in laser pointers just keep getting better, and are becoming less and less sensitive to thermal variations. I personally believe that with a tec keeping the module at a reasonable temperature it should be possible to get something decent out of it. The mode hops that happen due to thermal lensing are always going to be there, but even many commercial lasers have this O_o

I have toyed with the idea of incorporating some feedback, but the turning points in the graph and hysteresis could really wreak havoc with the control loops. A low power DSP solution would probably be the best bet, you could incorporate some stateful PID loops to do something about the mode hops.

Witnessed another laserist running some tests the other day with O-Like pointer cores just hooked to a new analog driver. A few of them ran just fine as is, with good linearity.

One however, made excellent power while being ran CW or with bright frames. If you left it off for a few seconds, it took a while to warm back up and had no to almost no output at low brightness levels. In other words, it needed to heat back up or be heated to gain linearity when blanked for any length of time.

I have a 180 mW pointer core that is doing the same thing.

Steve

Quote:

---

Originally Posted by krazer

I have toyed with the idea of incorporating some feedback, but the turning points in the graph and hysteresis could really wreak havoc with the control loops. A low power DSP solution would probably be the best bet, you could incorporate some stateful PID loops to do something about the mode hops.

---

Yes this one i got told too by a laser engineer in a corporation i made my college practice. these hops would confuse a PID regulation completely. He told me a fast enough µC with LookUpTable could do the math. 4kB of space would do even 16bit. But however the Table must be first generated from fully graphing the laser. i have no clue how to do that...

If you think a simple LUT is enough, it would be much easier to implement it on your computer, do any of the lasershow programs out there support a linearity correction already? It would suprise me if none do.

The problem I suspect you will encounter is that the correction table will be different depending on the frequency components of the modulation signal. It may work well for a first order approximation though.

I am curious as well, don't most projectors suffer from these nonlinearities? Most laser manufactures do not test let alone specify their modulation performance, and I hear very little about it, but it seems like for graphics it would make a big difference...

n though.

I am curious as well, don't most projectors suffer from these nonlinearities? Most laser manufactures do not test let alone specify their modulation performance, and I hear very little about it, but it seems like for graphics it would make a big difference...[/QUOTE]

Pango has a 4 point correction during setup based on adjusting 4 lines to equal brightness.

Steve

So what now?

For a LUT, would a µC be fast enough, or have it to be a DSP?

Quote:

---

Originally Posted by Xer0

So what now?

For a LUT, would a µC be fast enough, or have it to be a DSP?

---

16 or 32 Mhz micro would be probably fast enough, but I think there will need to be some "predictive coding" besides a LUT, and some delays for the green and red lasers to line up with 473 blues.

Steve

If you are willing to use a adc/dac that are on a parallel interface you could get away with even a 1MHz micro, running at 32 or 48MHz will give you plenty of time for predictive power correction algorithms.

If you want to save space and go for a serial one, you could do it with a SPI interface running at 10-20MHz.


Megapoints to the first one to write the correction algorithm as a sound processor plugin for windows that you can apply to a soundcard dac *retardo

I know Pangolin Quickshow has linearity correction features. though I'm not sure I've seen many other softwares with it.


As for pointer modules, CNI has started selling their pointer modules and calling them "OEM laser modules", I think Dave has tried them out on modulation, but they aren't really as cheap as the o-like or generic pointer modules.

Quote:

---

Originally Posted by krazer

If you are willing to use a adc/dac that are on a parallel interface you could get away with even a 1MHz micro, running at 32 or 48MHz will give you plenty of time for predictive power correction algorithms.

---

do µCs not already have inbuilt ADCs and DACs ?

Quote:

---

Originally Posted by Xer0

do µCs not already have inbuilt ADCs and DACs ?

---

They are not even close to fast enough.