I actually don't fully agree with EH about the penetration thing. The beam angle is sort of irrelevant. The number of total photons emitted is what matters, regardless of the angle. Angle is just a trick.
You get better "penetration" by spreading the light out and making it more diffuse, not by concentrating it into a central beam. The term penetration is sort of a bogus term to begin with when using it in the context irradiating weed.
Penetration is a fuzzy subject. It comes from the depth of field of a given light source, between the point where any closer and it burns to the point where any further and it's inadequate. The higher the wattage of the light source, the larger the distance between these two values will be- and since HID lamps tend to run high wattage- up to 1kW in wide use- it stands to reason that people would think this gives greater penetration of the canopy.
Yet there are several ways to game this outcome; one way is to put the light source on a mover, and then as the light moves, so does shading. Another is multiple light sources, so the cross lighting does the same thing.
One last consideration is that at least for 600W and 1kW HPS lamps, the light source itself is fairly long, allowing for substantial shading reduction.
I don't think these make HID lighting better, but it does account for the perception of 'penetration'.
I'm still designing a fixture for my own application, but I think LED can easily create similar penetration levels to HID lamps, and with enough height in the space above the canopy, perhaps even more.
But why would this be necessary or desirable? ScrOG techniques already flatten the canopy so much that any advantage gained wouldn't be well utilized. I see precious few grows where top quality bud is produced at a level far below the canopy, so the point would seem to be null. Better to have plenty of light sources all optimized to work within a foot of the ScrOG canopy to reduce shading and get maximum umol values where they belong.
Thoughts?