Correcting the beam of the 300mW 638nm diodes

I'm sure there's research out there in this area already, but it's tough to find and I wanted to do it myself anyway. Experimenting is fun

This is how the prism expands the beam:



I've seen that one prism works great. But everyone uses two. Why two? Presumably to expand the beam more. This also means higher losses, though, and you can also expand the beam more on one prism by increasing the beam's incidence angle. I was curious on how all of these factors compare, so I spent an afternoon on it. (these surplus shed prisms were used) These are my results



And this is what is meant by angle of incidence:



So... If I wanted 1.5mRad, I'd need to have either one prism at ~66°, or a prism pair at ~59°. A single prism loses 25%, and two prisms lose almost 40% at these respective angles. The only advantage I can see to using two is that the resulting beam can have the same direction, while a single prism steers the beam off to an angle.

These numbers will be different for other prisms, but the general line shapes and positions with respect to each other should be similar.

thoughts?

That makes alot of sense.
I wish i could see someone post things like that with the knife edge idea.
[~~~~] then knife it and rejoin overlapping to get [~~] the use of a wave plate will keep the beam specs and pattern the same.
The idea to cut the beam in half then use a cube with a wave plate to bring them back
If you do that and use a single prism you could get an amazing beam with losses under 35%.

Thank You for posting this. It will be helpful when i do this on my own.

:gj:

Here is a comparison. I don't have much space in the house. 7 meters is about the best I can do. The difference will obviously be even more drastic at further distances.

They are both running at the exact same current (they're in series). The corrected one actually looks slightly brighter even though it has ~30% less power.







And here's an arctic and a 300mW 660 on the sides:




This didn't turn out well, but here it is anyway:



edit: forgot to add, running at 950mA.

What are you running them at.?

Impressive.

Cant wait to finish my build with Correction.:wave:

Woah, crazy wide beam, but it cleaned up nicely.

Are they doing 300mW? It looks waaay brighter than the 300mW 660.

Ghostchrome said:

Woah, crazy wide beam, but it cleaned up nicely.

Are they doing 300mW? It looks waaay brighter than the 300mW 660.

Click to expand...

He said that they run with 950mA so they should output alot more than 300mW
Look here: http://laserpointerforums.com/f52/rayfoss-mitsubishi-638nm-300mw-laser-diode-67381.html#post971328

@Cyparagon: Nice thread! +rep when i can again

Thanks for the info ! I am waitting 2 of those diodes from dtr,

Thanks for this.

Once corrected, would these be advisable to replace knife-edged setup for my diy scanner?
I have never owned either so I have no comparison, trying to work out which to go for..

That is still up for debate. What are you knife-edging?

Cyparagon said:

I've seen that one prism works great. But everyone uses two. Why two?.

Click to expand...

Youanswered in your post

advantage I can see to using two is that the resulting beam can have the same direction, while a single prism steers the beam off to an angle.

Click to expand...

Another advantage that you can design prisms with less reflection by exploiting so called a Brewster angle. Antireflection coatings help too ($xx$)

Len said something about the lens from a sled. Any clue as to what he means?

The focusing lens from optical drive sleds have very short focal lengths. Therefore, if you collimate a beam with one, you will have a much smaller initial beam (with increased divergence over a regular lens)

well from what we saw of his reading using one of those lenses gave decent results so im happy.