RGB Dichroic Mirrors

[diachi]

I went with the flexmounts because of their relative abundance and that I get 3 of them. I will have to make a lens holder to hold the mirror in place. The plastic holder will mean that I will not have to stick the mirror into place, but stick the holder.

Thanks for the help Diachi :beer:

You shouldn't have any problems sticking the mirrors on, although that's more suited to square mirrors. Only problem then is removing the mirrors!

Post what you come up with for a mirror holder, I'd be interested in seeing how you go about it

 

[Benm]

It's a fair tradeoff: using simple cheap mounts can work out pretty well, you will just spend an insane amount of time aligning stuff - and possibly doing that again every time the system is moved.

You could spend 4 hours mucking around with them or shell out $100++ on 3 fully adjustable ones that you can set up in 10 minutes or so, choice is yours entirely. And if its a one-off build/install you may prefer putting in the time.

 

[Dreadnought]

Ok, I just bought a blue OSRAM 450nm module however I got a couple of questions on colour balance.

This thing is supposed to have an output of 50mW. Now, to get a good colour balance; would output matter. For instance, if the blue is at the front, red is in the middle and green is at the back. Would having 40mW
Red and 30mW green make the white too blue or would it have good colour output.

Fortunately, the module description does say that output is 40mW-50mW so it means that I could potentially put it down to 40mW but it might be hard.

I am asking this because there is some conflicting info. Multimode's RGB laser has blue as the biggest output but JLSE's has red as its largest output.

My setup isn't as compact as these guys but these 2 threads have good info and that I also plan to change colour intensities to get different colours.

My dichroics reflect blue last, so it might make it an issue.

 

[Bionic-Badger]

After searching around, I found these thick bronze 1" washers at my local hardware store and they fit the kinematic mounts for 1" optics pretty well. I like that I don't have to glue the optics straight to the mount itself.

The washer leaves about 0.5" diameter worth of space in the center which is usually enough. Unlike other 1" washers, this washer is thick enough that it is held in place by the set-screw without slipping. Pretty good for like $0.60 a washer. I'd like to find some actual optic ring mounts, but the real ones are often expensive.

 

[Bionic-Badger]

Re: color balance:

The ordering of the lasers won't matter much except for whatever losses that may be incurred by the optics. What's going to matter is the power and divergence. There are many different power ratios to pick from (e.g. RGB = 4:1:9 for 638/520/462nm), but often it all depends on what power lasers you can source. Having lasers with different divergences may lead to power ratios that don't match your original design.

What will probably happen is whatever mix you manage to produce will look "white" and be good enough. It's often very subjective, and absolute ratios based on color sensitivity (like the ratio above) are just ballpark guides.

 

[Dreadnought]

Re: color balance:

The ordering of the lasers won't matter much except for whatever losses that may be incurred by the optics. What's going to matter is the power and divergence. There are many different power ratios to pick from (e.g. RGB = 4:1:9 for 638/520/462nm), but often it all depends on what power lasers you can source. Having lasers with different divergences may lead to power ratios that don't match your original design.

What will probably happen is whatever mix you manage to produce will look "white" and be good enough. It's often very subjective, and absolute ratios based on color sensitivity (like the ratio above) are just ballpark guides.

Ok, thats good to know. That means that I am able to use that OSRAM module. If I have the wavelengths for my RGB system as Red 635nm | Green 532nm | Blue 450nm and I want to get a good white balance, what sort of mW's do I need for the reds and greens to get a good white balance. Considering that the design is like so:

What sort of outputs would I need for these freqs? Would the ratio be 4:1:2 or 1:2:2 or having power levels like JLSE's laser?

For instance, if the power outputs are like so: Red: 60mW | Green: 40mW | Blue: 50mW - would this be a good ratio of power outputs for white?

I don't really have access to some of the resources because I am on mobile currently.

After searchinRedround, I found these thick bronze 1" washers at my local hardware store and they fit the kinematic mounts for 1" optics pretty well. I like that I don't have to glue the optics straight to the mount itself.

The washer leaves about 0.5" diameter worth of space in the center which is usually enough. Unlike other 1" washers, this washer is thick enough that it is held in place by the set-screw without slipping. Pretty good for like $0.60 a washer. I'd like to find some actual optic ring mounts, but the real ones are often expensive.

That was something similar in what I was thinking to hold the mirror but my idea was to make a flat magnifying glass design to hold the mirror. I would stick the 'handle' onto the mount with epoxy.

 

[Bionic-Badger]

You can determine the "exact" ratios by using the luminous efficacy values for different wavelengths, which correspond to how bright your eye perceives certain wavelengths.

To get the ratios, you find the values of the wavelengths, for example (just using round numbers for wavelengths):

(R,G,B) = (640, 530, 450)nm --> (0.175, 0.862, 0.038)V_lambda

The normalized ratio values would be:

(0.175, 0.862, 0.038) / (0.175 + 0.862 + 0.038)
= (0.163, 0.802, 0.035)

Invert the values:

(6.143, 1.247, 28.29)

Normalize w.r.t. green:

(4.926, 1, 22.69) ~ 5:1:23

It's a pretty big ratio difference between green and the other colors.

Keep in mind that the ratio is not set in stone. The efficacy values are taken from photopic vision (daylight). When you're looking at laser beams/dots you're more in a mesopic vision regime, where you have high contrast dark areas and areas of brightness. The ratio will therefore be highly subjective.

A lot of the time your ratio will depend on what your limiting channel's power is. If you can't supply a whole lot of red, you'll have to tone down the green and blue. Likewise if you have tons of blue. Often you'll have enough green for your needs. Sometimes choosing a different wavelength can have dramatic differences in brightness (e.g. 462 vs 445); so you'll have to compensate.

For (638, 520, 455)nm I'd first try 4:1:8 and then adjust as needed. Some of the ratios you'll find on the net are using (635,532,473)nm wavelengths which have different brightness ratios than the wavelengths you're using. Again, it's pretty subjective and even "off" ratios can look pretty good.

 

[Dreadnought]

You can determine the "exact" ratios by using the luminous efficacy values for different wavelengths, which correspond to how bright your eye perceives certain wavelengths.

To get the ratios, you find the values of the wavelengths, for example (just using round numbers for wavelengths):

(R,G,B) = (640, 530, 450)nm --> (0.175, 0.862, 0.038)V_lambda

The normalized ratio values would be:

(0.175, 0.862, 0.038) / (0.175 + 0.862 + 0.038)
= (0.163, 0.802, 0.035)

Invert the values:

(6.143, 1.247, 28.29)

Normalize w.r.t. green:

(4.926, 1, 22.69) ~ 5:1:23

It's a pretty big ratio difference between green and the other colors.

Keep in mind that the ratio is not set in stone. The efficacy values are taken from photopic vision (daylight). When you're looking at laser beams/dots you're more in a mesopic vision regime, where you have high contrast dark areas and areas of brightness. The ratio will therefore be highly subjective.

A lot of the time your ratio will depend on what your limiting channel's power is. If you can't supply a whole lot of red, you'll have to tone down the green and blue. Likewise if you have tons of blue. Often you'll have enough green for your needs. Sometimes choosing a different wavelength can have dramatic differences in brightness (e.g. 462 vs 445); so you'll have to compensate.

For (638, 520, 455)nm I'd first try 4:1:8 and then adjust as needed. Some of the ratios you'll find on the net are using (635,532,473)nm wavelengths which have different brightness ratios than the wavelengths you're using. Again, it's pretty subjective and even "off" ratios can look pretty good.

Yep, I got a similar ratio after doing some calculations. I then started to look at more modules and then the thought struck me. I could use Tocket's Chroma program to see if I get a rough, different ratio. That way I can try to pick a white and then when it comes to it; adjust the beams perfectly to see a good white. I think I have the program somewhere in my downloads folder, but I cant check until I get access to my PC next week. Good thing you can calculate a good white even with loss from dicro. I even think Chroma even calculates a good white from scotopic instances which means a better white in mesopic conditions. Chroma does get a lot of slack but its good for rough estimates like our calculations that we did earlier.

--------------------------------------------------------------------------------------------------------------------------------------

I have finally found Chroma in my downloads folder, and have figured out that 50mW, 60mW and 50mW (532nm,635nm,450nm) is a good output pick. I don't mind if its a little inaccurate but I have ran this calculation to make it less ideal: [number of watts] * [1- [loss of dichroics]]

So the calculation is like so:

50*0.75
60*0.75
50*0.85

Here is some pictures of the output of Chroma:

HTML code: #D9DFFF
RGB code: R: 217 G: 223 B: 255
HSV: 230.53° 14.9% 100%

Its more like a cold white which is perfect, further calculations seem to make this the ideal colour outputs. I could even have this by having a 1:1:1 ratio:

HTML code: #BFE0FF
RGB code: R: 191 G: 224 B: 255
HSV: 209.06° 25.1% 100%

Because of how powerful its going to be (around 100mW), the brightness will make the beam more white because of factors like human perception and even White's illusion.

 

[Dreadnought]

Sorry for the double post, but I recieved my laser module for this project. I will put some beam pics someshere but for now, I need to buy the rest of the parts. I also need to make a wooden box to house everything because plastic is too flimsy and I dont have my hands on any metal. I will also need to buy a fan to remove the excess heat in the box

 

[Apocalypse]

DTR's shop used to sell them, but it's out of stock now.

I think he still has some of them dichroics left. Maybe you can PM him.