PWM drivers for lasers. Why/why not?

No, there always needs to be at least one capacitor. An inductor provides voltage for half the time, and the capacitor provides voltage for the other half (as far as I know). Without a capacitor it would be extremely choppy with lots of bad spikes.

Here's a good resource on how boost and flyback converters work: Flyback Converters for Dummies

Thanks, I'll read that once it loads on this dial-up tier satellite.

rev0 said:

Do you mean regulation via adjustable PWM duty cycle? It's been discussed before, while this does vary apparent power, it doesn't do a darn thing for safety since the peak power would still be the maximum

Click to expand...

That depends on how much power is in a single pulse.

Obviously, a 1W laser that has a 0.5% duty cycle averages out to 5mW actual power. However, if it's running at 5 millihertz, the "on" pulses are one second long, at 1W more than enough to damage the eye.

So at lower frequencies, this is correct, and if it's all you guys are saying, than I concur. However, at sufficiently high frequencies, PWM beams become physically indistinguishable from the equivalent-power CW beam. The question I originally had, was whether sufficiently high frequencies to get the energy content of each pulse below the danger threshold were practical.

So, let me work this out for a 1W beam. Basic assumptions are as follows:

The blink reflex, say is 1/10 of a second, 100ms. 5mW is the maximum power that will fit a 100ms pulse without damage, per the CDRH standard. So, a 50mW pulse maxes out at 10ms, 500mW at 1ms, and 1W at 500 microseconds for the same amount of energy (and accordingly, the amount of "work" -- including that which we regard as "damage" -- that it does.)

If the longest "safe" pulse of a 1W beam is 500 microseconds with the 0.5% duty cycle required to get it down to a 5mW average, the resulting lowest safe PWM frequency is 10Hz -- one pulse per eyeblink, essentially.

That's flickery as hell, so the next step is to increase the frequency. At 1KHz (to pick an arbitrary frequency where ordinary hand motion won't produce an easily appreciable string of dots) the minimum frequency, the "off" time is ~5ms long and the "on" pulse is now 5 microseconds long -- that's the maximum CDRH energy dose of 500 microseconds sliced up into 100 more pieces.

So the question then, is whether doing 5 microsecond pulses for a 1W emitter is practical at the hobbyist level. If it isn't -- and from reading posts here, I think that may indeed be the case -- then the safety claim is true for all likely frequency parameters we have available to us.



EDIT: it appears that the answer to my last supposition is indeed "yes". While researching for drivers for 1W+ builds, I ran across this thread by Xer0 with the same question, and even the same calculation I used (down to the same numbers) and some good info from LSRFAQ: http://laserpointerforums.com/f65/safe-pulse-energy-53164.html

Canuke, a few more
things to consider.

One, thank you for taking the time to do the math. +1 is the best I can do.

It is dangerous to assume a single pulse will trip the bio mechanisms, the eye may try to close the iris down the bio version of a few F stops before it trips closed. So you may have 3-4-5 pulses before you get a full close.

Next up, it is energy density per unit area, not just total power. I keep forgetting that part, myself.

Active collecting area, dark adapted eye, is a 7 mm diameter circle.

Worse case is "aided viewing", a person with a pair of glasses on can be considered a 50 mm diameter collection area.

Last of all, The eye has a optical gain of roughly 100,000.

The math is here, along with the tables, see pages 22,32, 118, and 207
This is the only complete version I have found on the internet, many of the other safety standards and documents are "pay to play" or incomplete.

Army laser safety, download the pdf, click here:

Laser Main

Steve

I would +rep you if the android browser would let me canuke

One part whe even dont thought about, but a guy at a german board mentioned: The fact that a µC, which is need for a usable (small enough) PWM circuit, could hang up / get some kind of error, and go full blow suddenly. Maybe its a like-chance-of-getting-hit-by-an-thunderbolt thing but it should not be just ignored...

Xer0 said:

One part whe even dont thought about, but a guy at a german board mentioned: The fact that a µC, which is need for a usable (small enough) PWM circuit, could hang up / get some kind of error, and go full blow suddenly. Maybe its a like-chance-of-getting-hit-by-an-thunderbolt thing but it should not be just ignored...

Click to expand...

If you code it properly, the chances are lower or the same as the driver itself failing.

LSRFAQ said:

Canuke, a few more
things to consider.

One, thank you for taking the time to do the math. +1 is the best I can do.

Click to expand...

Ah, that's where those rep points came from. :thanks: guys!

LSRFAQ said:

Next up, it is energy density per unit area, not just total power. I keep forgetting that part, myself.

Click to expand...

This is based on the assumption of a visually perfect lens/cornea, it is worth noting; the "dot" on the retina isn't going to be that tiny for most of us. Ironically, astigmatism acts as a mild sort of built-in protection, as it means that the light is always out of focus in at least one axis.

That's not to argue that the safety rules are to be disregarded or watered down, though, anymore than having multiple safeties on a pistol means we can point it at whatever

Thanks for the link to the safety guide, I knew that there was more to the story than the math I ran through.