Hacker News Reader

Comment by smlacy | Hacker News Reader

Every time I see this link, which is about every year or so, I spend an inordinate amount of time trying to build a "laser". It really feels possible, using a combination of diffuse/reflective to "corral" all the light into going mostly in the same direction. Would love to see if anyone has gotten close.

View on HN

A parabolic reflector -- all (specular) reflective, no "diffuse" or "transmissive" -- will get all the light from a point source going in the same direction. If you really want all the light doing that, you need an infinitely wide parabolic reflector so you'll get an infinitely wide beam. But if you're happy for, say, 99% of the light to end up in the beam, you can make a finite portion of a parabola that does that. Shrink it in towards the source and (aside from discretization issues) you will get as narrow a beam as you like with 99% of the light in it.

A real physical light source can never have zero size as the source here (aside from discretization issues) does. As a result, you can't get an arbitrarily narrow beam from it. There's a quantity called "etendue" that measures a sort of combination of spatial and angular spread-out-ness, and no combination of optical elements can decrease it except by absorbing some of the light.

As others said you can use a parabolic reflector to direct the light in one direction.

I think this can work:

On one side place a parabolic reflector to direct the light in one direction.

On the other side place a circular reflector to direct the light back to the source. And place a tiny hole in this half circle to let the light out.

Quick drawing: http://zenphoton.com/#AAQAAkACAAEgWwAzALEBEwC0AO4A/wAAtADxAL...

It should be possible, what you want to do is try and draw a parabola with fully reflective surfaces where the focal point is the light source.

https://en.wikipedia.org/wiki/Parabolic_reflector

Reply on news.ycombinator.com