Artificial Sunlight (official topic)

My current setup is using a 150W 5600K Luminus LED, with 95 CRI. The LED outputs about 12,500 lumens.

I just picked up another Luminus LED that’s twice the power and lumens output (280W, 25,000 lumens) and I’m looking forward to trying that. I have a much larger dish sitting in my garage, and I’m thinking of getting a Luminus CVM-45 to test with it—that model goes up to 600W and 57,000 lumens, and is available for around $145 USD here in Canada.

One idea I mulled around but never pursued—sunlight makes your skin feel warm, which LEDs don’t do. I was considering whether I could simulate this with an additional infrared LED placed next to the primary LED, but haven’t pursued the idea. I’m not really sure if it would be safe? It’s been a while since I looked into it but I remember getting the impression that it might risk eye damage. I’m curious if anyone else has had this idea, or if anyone has any thoughts on it?

Where are we at on getting to a solution for an artificial sunlight that can be hang from a wall instead of requiring a large volume to set lights up? Is that even physically feasible?

Also, instead of using a 6000k light and a blue film to filter out and spread the blue light, would it be possible to use a 5000k light directly and simulate the bounced blue light with a secondary blue fill light? That’s more of a lighting solution than a physically accurate one, but it might simplify the final product.

Thoughts?

@moment  I doubt you'd get something hanging on a wall. Artificial sun beams need a lot of space in a DIY set up. I think a combination of two routes is best to nail down artificial daylight in a room. One being something like this: https://lightglasslighting.com which on further digging is very similar to the DIY Perks smashed tvs video. Use this as a fill, then combine it with this for soft beams:

I have this set up in my north facing office currently on its own, and it's ok, but not amazing. The shadows are verrrrry blue, and the light is a bit too yellow, which is quite a depressing experience imo. But I think balanced out with a good white daylight panel, it might feel nicer.

@moment Saying that, what you're looking for has been done, it's just really expensive to produce according to this guy after reaching out to him.

Posted by: @bigfield

↑

@moment  I doubt you'd get something hanging on a wall. Artificial sun beams need a lot of space in a DIY set up.

I agree that we’re unlikely to get anything small enough to fit on a wall. But I’m convinced that a large, shelf-top sun light is possible using similar techniques to rear-projection flatscreen TVs. I haven’t yet put the time and money into investigating in depth, but if you look at rear-projection TVs from right before the industry transitioned to LCD, they can be quite shallow for their size.This 42-inch TV for example has a diagonal width of 107cm, but the depth is only 30cm! At its most basic, this is accomplished with a fresnel lens at panel, an angled mirror in the back, and a projector lens at the bottom. It may be that the optics of the most “modern” rear-projection TVs are more complicated, but I haven’t seen anything to confirm that and haven’t yet gotten around to finding one and dismantling it.

Posted by: @timdmackey

↑

At its most basic, this is accomplished with a fresnel lens at panel, an angled mirror in the back, and a projector lens at the bottom.

Here are some different optical schematics for rear-projection TVs. As you can see, some of them use multiple mirrors, while the shallowest appears to use just one mirror.

Hello folks, I've been looking at the Coelux patent and it appears they use a compound parabolic concentrator (CPC) to generate the virtual sun. @DIY Perk, was your first project able to mimic the sun in such a way that it looked like the ones from Coelux etc., or was it actually just the linear light effect?

Posted by: @hans23

↑

Hello folks, I've been looking at the Coelux patent and it appears they use a compound parabolic concentrator (CPC) to generate the virtual sun. @DIY Perk, was your first project able to mimic the sun in such a way that it looked like the ones from Coelux etc., or was it actually just the linear light effect?

-- attachment is not available --

 

I thought about this concept before myself. Funny that that it is patented. Anyway i saw multiple problems with this setup which is why i did not further pursue it.

1. If the observer stands near to the sunlamp and has direct sight onto the emitting light source he will obviously see the light source but also the reflections of the parabolic mirrors on both sides thus giving the impression of multiple suns.

Now one could say alright. We only need to make the parabolic mirrors flat enough such that at an appropriate distance one will only see the emitting light source if one does not see the reflection from the parabolic mirrors and the reflection on one of the sides if one does not see the reflection on the other side or the light source directly. While this is true this leads to following impression then:

2. If the parabolic mirrors are rather flat and/or the observer stands far away from the sunlamp the observer will see that there are multiple disjoint light spots such that if he walks past the sunlamp he will see the first sun illusion, then nothing, then the second sun illusion which corresponds to direct line of sight to the light source, then again nothing and then the last sun illusion.

So we should not make the mirror flat to avoid 1. either. But one of those problems must occur as far as i understand it. Either these regions of reflections or whatever one wants to call them overlap, then we see multiple suns or they are disjoint. It will only be perfect at one fixed distance it seems to me. Furthermore one more note on 1. We also actually want theta_max to be rather small right? But as it get smaller the distance at which 1. does not happen will also grow.

This is why i discarded this setup but if i am missing something i would be glad to be corrected.

1. @purple did you end up buying the sheets from 1688?

@nolo any news on your projekt?

 Here's my crack at this project. I'm actually pretty happy with how it looks. It's a Nanlite LED + Fresnel shining up at a ceiling-mounted mirror covered with waterproof transparency sheets. The diffusion could be a bit bluer, but overall not bad! Thanks to all the folks in this thread for the inspiration!

Let me know if anyone wants the BOM.

Hello again,
I got some pictures of a system that works pretty well. As they haven´t made that system puplic yet, I can´t show it here. But I think some of you guys could help with finding the principle. I don´t know if it is possible to make a private Discord group or something like that to get a faster way to respond to each other! Because I thin if we could make a group call or something, we could fasten up the process! I also orderer a panel that produces a artifical skylight and would like to share it with youu as soon it is here! Just let me know.

And again the question, does the parabolic mirror and that one LED allready creates that Sun and Sky effect and we are just trying to make it smaller, or is it only for the parallel light?

@hans23 ooo what did you order? Would love to see what you've made too

Hey, I'm crazy interested in this Rayleigh thing and have scoured the internet trying to find something accessible. Anyone interested in checking this out? https://makerstock.com/products/acrylic-translucent-white?variant=39931672297569

Posted by: @devsw1

↑

Hey, I'm crazy interested in this Rayleigh thing and have scoured the internet trying to find something accessible. Anyone interested in checking this out? https://makerstock.com/products/acrylic-translucent-white?variant=39931672297569

I may be incorrect, but I don’t think translucent acrylic exhibits the Rayleigh/Tyndall effect. I think it just adds a haze/blurs objects behind it without affecting the color.

I bought some of the waterproof inkjet transparencies that were suggested by @Nolo, and they work really well! I think Nolo is right though about 5600K LEDs not being blue enough for the proper daylight effect. The blue color correcting film I ordered just arrived yesterday, so I’m going to try that out soon to see how it improved the color. 

New 600W LED just arrived! Now I just have to figure out how to cool it 😅

Has anyone had any success calculating the cooling needed for high powered LEDs like this? The LED I’ve been using until now is 150W, and keeps cool just fine using a low profile CPU heatsink with active cooling, but with 4x the wattage, this LED puts out waaay more heat. W

ith many CPU cooler manufacturers no longer publishing TDP ratings, it’s been really hard to find concrete information on what might be able to keep this thing cool. I can’t even find the information about water coolers either!

Luminus (the manufacturer of this LED) links to some information about heatsinks that will pair with their LEDs, but it’s all passive heatsinks which are monstrously large (think 10 pounds of aluminum).

I found this article about calculating heat dissipation needs for LEDs, but the math kind of went over my head. Maybe someone else could make more sense of this? I don't really understand all the concepts/properties needed for the equation, and how I could translate that to a CPU cooler.   https://ledgardener.com/find-heat-sink-cob

For now, I have a large Thermalright Peerless Assassin Air cooler that I bought a couple weeks ago. I’m going to put the 600W LED on it and hook up a temperature monitor to the LED temp junction. I’ll slowly ramp up the power from the minimum (<100W) and monitor the temperature until I see it getting close to the rates maximum. Wish me luck!

@timdmackey Wow, pretty powerful LED chip. With an aircooler it's almost impossible to cool the chip @ 600W Output. Matt did a Youtube video where he cooled a similar 500W LED chip with an aircooler:

But as you can see in the video, the cooler is very massive and loud, so I would definetely go with watercooling instead. With watercooling you can cool the LED almost noiseless with the right fans and big radiators. It will be a bit more expensive than an aircooler solution but IMHO worth it.

There are All in One watercooling solutions on the market with a 420 mm radiator which cost around 100 euros. You just need to build a custom bracket to mount the LED chip to the CPU block. I think you could run the LED chip at good temperatures and silent if you run it with 300 to 400 watts and 6 silent 140 mm fans (3 fans at each side of the radiator). To run the chip @ 600 watts all time is not healthy anyway in the long run, unless you really have a monster water cooling solution.

@robsi Unfortunately not, because I am currently moving to a new appartment.

But thanks to this thread, the knowledge about all materials needed is now available and I have a lot of ideas for the future, which I will post here as soon as I realized them. But first I have to renovate the new home. The sunlight imitation is like the icing on the cake at the end 🙂

@nolo thanks for your insights. It hadn’t occurred to me that I could mount fans on both sides of a water cooler radiator! I’ll definitely want to give that a try at some point.

TheLuminus CVM series is designed for film production use and long life, so I think I’d feel pretty comfortable running it at full power as long as my cooling system was adequate to keep the chip under the recommended case temperature. This chip has a maximum case temperature of 120°C, but I I think I’d probably feel more comfortable if it was running at the “typical” rated temperate of 85°C or lower.

As for noise, it’s not so much of an issue for this version I’m making, because it’s going to be in a large workshop. But quieter operation is always a plus, especially if the cooling performance of a water cooler is better!

@nolo Also, looking at that video you shared about Matt’s 500w LED, it’s incredible how much the prices on these chips have come down in 5 years. The CVM-45 600W LED that I bought is only $140 USD when buying them individually!

https://www.mouser.com/ProductDetail/Luminus-Devices/CVM-45-56-95-60-BA42-F5-2?qs=doiCPypUmgGURKFNcaBCGQ%3D%3D

 This is my artificial bathroom skylight. I'm using a 100w led with an 80cm satellite dish mirror.  I've found quite a good tyndall effect by mixing colloidal silica in epoxy. 

I think that's one of the best DIY artificial sunlight installations I've ever seen! Brilliant! I'll have to check out that collodial silica idea... can you share any specifics on how you went about it?

This is really beautiful, I'm also very interested about the details. Did you buy a ultrasonic probe device for mixing the nano particles with the epoxy ? And did you cut a hole through the roof to integrate the satellite dish?

@jordiemc please give us the details! Epoxy brand, type and amounts of silica, mixing method, and if you had to polish it...

Thanks for the positive feedback.  It's been a long project with a lot of experimentation. 

Due to the limited space in my loft above the ceiling, my mirror is quiet close to the window.  This is unavoidable, and the edges of the the mirrored vinyl sections cast shadow lines on to the window.  I tried very hard to create as perfect a mirror as I could but mirrored vinyl doesn't stretch very well. 

I even tried using a compressor to pull a vacumm to get the vinyl to stretch over the dish but I still got creases. https://photos.app.goo.gl/dVhX6RHgdnK9cCNh7

So i decided to create a JWT style mirror https://photos.app.goo.gl/GjygJoX9QjkX3bq38

With regards to how I created the window, I experimented with mixing nanoparticles such as titanium dioxide in resin hardener. However without access to an ultrasonic homogeniser I didn't really expect to get anywhere. 

I mixed in colloidal silica and it turned the hardener white, but I was surprised to find a day later that it was now translucent and had a blue hue when I held a light to it.  I suspect it has something to do with the refractive index of the resin being similar to that of the colloidal silica.

For my first window I mixed a roughly equal volume of colloidal silica to resin hardener.  Then I put the mixture in the vacuum chamber over night. I used a 2:1 resin kit.  For the size of window I made, at around 5mm thick, i required around 1.5kg of epoxy.

My bathroom ceiling was needing done anyway, so I figured it was the ideal place to install the skylight.  

https://photos.app.goo.gl/7jvFbYT4shvZU6CaA

If anybody has any questions on it, I'm happy to answer them.

@nolo I did cut a hole in the ceiling.  The effect is much nicer than the usual down lights.

@jordiemc do you think the vacuum was mandatory for the effect to work?

@fhr I think it could be done without a vacuum chamber, maybe if the hardener/colloidal silica mixture is left for long enough the air bubbles will rise and escape. Mixing introduces a lot of air. Probably why the solution looks white at first.

After pouring the resin I used a blow torch to try to get rid of bubbles. You need to be careful not to burn the epoxy though.

Posted by: @jordiemc

↑

@nolo I did cut a hole in the ceiling.  The effect is much nicer than the usual down lights.

Absolutely, your bathroom looks very pleasant with the skylight, it has a kind of "openness" to it now. No comparison to ordinary lighting, which gives rooms a somewhat depressing feeling. Your sklyight simulates the real sun and sky very realistic. The picture whit your bathtub is very nice, the "sunlight" looks very realistic! 

And thanks for the answers Jordiemc. The last picture illustrates good how you assembled the satellite dish at the roof of your house. Do you have any details how you mounted the LED, because it seems that you found a good way to hide it very well, as it is not visible in the pictures.

I also don't know if you missed this part, but it was discovered here, that we can use waterproofed inkjet foil instead of mixing nano particles with epoxy. Maybe this would be the easier solution and you could give this a try in combination with an acrylic glass sheet for stability and put the film on top.

But anyway it's good to know that you achieved a good result without a ultrasonic probe. Just for interest, do you have a link for the colloidal silica ? If mixing just by hand gives a rayleigh scattering effect, it's a good alternative to the inkjet film. The more options we have, the better 🙂