"safe" pulse energy

Im not into that complicated calculations that define when an eye exposed to a laser can get damaged. so as i am plannig to build a PWM for throttling down an 445 for pointer-reasonable uses, i search for an easy formula to calcuate an "eye-safe" setting.

eye-safe in the case, that the eye is able to dissipate the energy of the single pulses safely until the fast wink reflex is closing the eye.

i need a relation of pulse lenght and output power

so for example, when pulsing at 5Khz, its basically 200µs pulse lenght.

with the relation pulse energy / dutycycle, 5mW could be achieved with:

50% duty at 10mW
5% duty at 100mW
0,5% duty at 1W

for example. but at everything under 400mW is not in the default working range of our Casio 445nm Diodes, they behave too instable at every low setting with rising/falling temperature, maybe too instable to pwm without squezzing an photodiode + regulation circutit into the host.

And at stable 1W, it's still 1µS pulses of Class 4 Laser radiation! still too dangerous for the eye?


So the final question is... what current setting we could use, assuming an 200µs pulse interval, and 5mW desired average output, so an accidental hit in the eye could not cause more damage than a real 5mW?

no Pro a clue? *push*

My guess if there afraid to tell you, since its a little dangerous.
What if you make a mistake and the pulse is too long or there data is wrong,
What if the microcontroller crashed and it stays on and someone goes blind?
I know you want adjustable, but its hard to do safe,
except with a ND filter. Hmm.

I hope you succeed, but even if the beam is dim-able, still treat is as if
it can blind someone just in case. Better safe then sorry.

A quick bit of rambling, as you at least deserve a partial answer, instead of just DON'T TRY THAT!.

The reason no one replied to you post is that maybe 5 people in all of LPF have the gear/knowledge to attempt this. The other 4 of them are probably at their day job or graduate school right now.


I can run the numbers for pulses, but what I can tell you is they are already well below the diode's threshold. You would be better with CW. Using a black glass neutral density filter and spreading out the divergence. It gets you more "effective" brightness, as the eye is not set up for detecting short pulses.

I've reviewed the books recently to help a friend get a operators permit in a state where it is required. I can tell you that PWM is not the way to go for a stationary beam, and your handheld is a stationary beam, you can't move your wrist fast enough or in a repeatable fashion, so you assume its stationary.

Its power density per unit area at the eyeball, not total power, that needs to be measured and controlled. There is a big difference between the two.

Making a 1 watt laser "eye safe" is a tricky proposition. Do you own a " Nist Tracable" power meter with a silicon cell detector? Thermal power meters are not accurate enough for this task, nor would I trust my eyes to a home made or low cost power meter.



If I just assumed the beam was TEM00 for my calculations:
You'd have to hold the power way above threshold, yet below the point where the diode goes to a complex transverse mode for the numbers to be valid anyways. Tests by the holography crowd have indicated this is possible with selected diodes and is above 10 mW yet below ~100 mW, but is not repeatable from diode to diode or across a wide temperature range. Many of the Nichia Blues never get a good transverse or longitudinal mode, even at low power. Low frequency pulsing shifts the mode and the divergence in real time, so now the measurement gets complex, fast.

That is tricky to do...

So, If I were to attempt this,which I will not, I would:

Get the beam profile down to near Tem00 by reducing power and running CW for stability. Next I'd put in a ND filter, expand the beam, and put in a safety circuit, looking at sampled optical output power, past the ND filter, that trips the laser off within microseconds.

After I'm sure that is the case, using a 1/2" CCD detector in my Spiricon Beam Analysis Unit, I would get the beam structure and uniformity over time verified.

You then need to choose a minimum viewing distance. You would adjust a upcollimator to a known divergence and once again get out the Spiricon to measure that diameter and divergence. Not 100% trusting the Spiricon with my eyeballs, I would then do some tests with scanning a pinhole detector across the beam at various distances. Note that the quality of your upcollimation interacts with the minimum viewing distance.

Next up, for final adjustment and test, You then need a sensitive, calibrated,sensor at a known 7 mm diameter, the size of a average pupil. No part of your beam can exceed the MPE when measured with that 7 mm diameter sensor. Then every time I used the unit I would check for power and mode variations. Ie Quality control done preshow.



PWM does not buy you anything, and the published safety table for the narrow pulses you would need is more restrictive then the continuous wave table anyways.

Eye safe with a CW, stationary, beam is a tricky thing, and I would be doing about 5 pages of math, and then sending those 5 pages out for review, before I would even dare to publish them. I'd still need the actual pointer, optics train, expander, and diode for measurement before I could remotely hint that its safe, there is so much variation from device to device and mode shifts with these diodes, that what is a good range of adjustment for one might not work with the next off the line, even in the same optics train with the same power supply.

Sure, reducing the output power and beam duration for CW or gated CW lasers, is always a good thing in mitigating potential damage. But delivered power is NOT the only variable in laser safety.
You have to look at diameter, divergence,wavelength, mode, distance, rep rate, pulse width, aided or unaided viewing, and quite a few other variables to make a eye safe laser.



I'm not saying it cannot be done, the Chauvet Fat Beam is a approved and tested laser with world wide approvals that uses the beam expansion technique. But I am saying its tricky and requires a lot of prep work, equipment, and follow up.



Steve

I would run it at the lowest possible (200 - 300mw ?), then use a filter to drop the output.

Maybe if you can create a heatsink which covers the diode. Then have a small hole so only part of the light from the diode can escape.

You would want the material to be non-reflective though, so the light doesn't go back into the diode.

Thanks a lot to Steve! so lets forget the PWM thing...