An experiment I want to do requires me light up a box with rainbow light which has frequencies of 380nm(Upper violet) or less to over 9000nm(Room temperature infrared).
It'd be most useful if it was radial, lighting up the soil with a junction between all colours at the centre and around that arcs of different colours.
Preferrably, the light energy will be independent of wavelength, i.e. for every watt of violet there's a watt of red.
It's a biology experiment, I want to put a simple plant (moss or algae) in a mutagen and see if it will create a form of chlorophyll that will favour unusual and possibly useful spectrums of light.
>>907558
that's gonna be tough...
from 380 to 980 or so you could do it piecewise with LEDs and phosphors, but specific wavelength lightsources are going to add up. Above that, you may need to go incandescent and filters... and I have no idea what you will do for filtered output longer than 2000nm. You're already talking about crazy expensive optics.
Stuff like a prism is not going to be linear or usable outside of the spectrum right around visible...
>>907565
Does it become any easier if I forsake a smooth gradient, and instead go with a couple dozen sources of increasing wavelength?
>>907569
Sort of, but I can't think of a way that's not hella expensive, so some other clever anon might have better ideas.
Like, you want to do what a spectrometer does internally. Except with really high power output.
>>907558
Do the wavelengths have to be separated like in a rainbow, or can you just go with a 'white' light source?
If they have to be separated, my first thought would be to get a couple of high-ish power white light sources, then use a monochrometor to filter out a few specific wavelengths. Of course, the light sources I have in mind aren't really that cheap, and monochrometors aren't either, and at best you would only be able to set up a few colors at a time.
Another idea for getting exact, yet arbitrary, wavelengths at a high power would be to get a dye laser setup. Then you can change the color by changing the concentration of dye in the gain medium. Spread the light out using a 10x microscope or telescope eye piece or whatever. Again though, that would be very expensive. You might try searching the universities and businesses in your area to see if anyone already has this equipment and wants their name on this science project.
>>907570
There's probably materials that would allow me to shine white light through and filter for the colours I want, but that'd be super inefficient.
One thing that could possibly work is if the materials were mirrors to anything they didn't allow to pass through, and so one could make a mirror chamber to get quite a lot of the light to successfully go in through the correct ports.
The two problems though are that it's doubtful that the full range of special mirrors exist (if any exist, for that matter), and also the lights themselves may absorb a lot of light hitting them.
>>907621
The colors have to be separated, though still be distributed over large areas of soil.
So, do monochromators end up losing most of the light going through them? Bear in mind that I'm going to have to use lots of powerful lights over a long time (months at least), and filtering out most of the light is going to mean I'll need even more powerful lights, could mean some crazy power bills unless it was solar powered (meaning added initial cost)
>>907623
The only truly efficient way to do it is to have 100 different lasers emitting 100 different wavelengths, then turn down the optical power output so that they are all at the same power as the weakest one.
That said, using a monochrometer may be wasteful, but it's doable. You'll definitely have to find a way to keep the heat from the white light source out of your experiment. Decent white light sources go for a few hundred bucks each, and a decent monochrometer goes for a few hundred bucks.
The cheapest way that I can think of to do it with lasers is to find the cheapest, high power laser diodes that you can, and rig a bunch of them up together. You can definitely find some high power cheap diodes in the 650nm range, you can definitely find some high power cheap 808nm diodes. You can probably find some cheapish 405nm ones. The reason those 3 colors are available for cheap is because red is a simple semiconductor to make, the 808nm ones were mass produced for burning cds, and the 405nm are used in Blu Ray players. If those 3 colors are enough for your experiment then you can probably rig that up for a grand total of ~$100 or so.
>>907631
Also, you can probably find some fairly cheap and high power 1024nm diodes. Moreover, you can put a frequency doubler in front of it and get a nice green wavelength too. So that's 5 colors.
Running laser diodes would be cheaper than running a white light source for several months, although if they are constantly on, you will approach the expected life of the diodes, so you should have some backups.
>>907631
Looking far too expensive for me, I'm now much more strongly considering a passive option for getting the light in the box.
My OP image indicates that water can definitely lens infrared light, and all the required frequencies will be present in sunlight.
A prism gives a non-circular but still radial distribution of light, and even if glass filters out some of the required frequencies, a prism-shaped container of water will still work as a prism.
One then merely has to use some large mirrors (even foil could possibly work) to put lots of light into the prism, and then use a filter to bring down the frequencies with the highest energies.
>>907634
Agreed OP. Sunlight is most likely your best bet here.
Again, off the top of my head, but you might consider getting some optical fiber and placing a few strands in each of the colors from your prisim. Then you can direct those colored lights however you want by arranging the other ends of the fibers. This may help you evenly distribute the different colors, and you can do a course adjustment of optical power for each wavelength by having more or less fibers in that area of the rainbow. Btw, if I recall, yellow and green are going to be the two colors that will have to be toned down.
A couple of tips if you want to go this way, regular old multi-mode fiber (the cheap stuff that they use for audio connections) is what you're going to want. Also, putting a collector on the tip of each fiber will help focus light in to the fiber. Typically a high precision lens system would work best, but you can go with hemisphere lenses that sit flat against the flat fiber, or you can even try your hand at using epoxy or glue or whatever to create your own hemispheres on the fiber tips. UV cured epoxy may work the best. Good luck OP!
Hold on, if I'm going to do a sunlight powered prism, amplified by mirrors, how will I compensate for the changing angle of the sun?
>>907739
Either manually adjust the incident angle, or make some mechanism that tracks it. There are tracking algorithms and software available for arduino, you just need to find them(Github).
>>907764
That's necessary for telescopes, but I'm just needing to do light collection here, focusing light from whatever angle into a single beam.
It is actually possible to do without any moving parts I think, but the only solution I've figured out so far is a transparent cylinder that has a low index of refraction around the outside and a high index of refraction along the core, and I haven't a clue how such a thing could be created.
>>907777
On this note, is there a way I can get a loooong sheet of material that has a high IOR at one end, and a low IOR at the other?
One way I can make such a cylinder is to wrap up such a sheet.
>>907777
You know, a paraboloid (like a satalite dish) could do it.
>>907777
Not sure if trolling, but you realize that's what a optic fibre is, right?
OP you may want to look into monochromator and monochromator gratings.
Having a constant power will yield EXTREMELY different results depending on what you're asking for.
Similarly to audio noise, you either want same power per frequency (for each 1 nm band) or a log power distribution ( same power from 200-400nm than 2000-4000nm ).
Either way, you also need a very precise measurement device if you want this to be controlled.
I really cannot help you without a more precise methodology explanation or a drawing.
Well, you could do the prism thing suggested by others, but instead of sunlight you can use a sulphur-plasma lamp. That will give off a broad spectrum white
>>908225
I need to take running costs into account though, and if it's possible with glass or mirrors to get a constant beam going into the prism with sunlight, then I don't need an electical light.
>>907621
>If they have to be separated, my first thought would be to get a couple of high-ish power white light sources, then use a monochrometor to filter out a few specific wavelengths
Well, you wouldn't use a slit to throw things out, just separate them. Use a grating or prism... problem is, one of that will result in even light, even spacing, much power, or provide anything much over 1000nm without special glass.
>>908225
Why would he use a multi-line discharge light source since his application is probably one of the few where blackbody is closest to ideal? He *wants* all the infrared an incandescent puts out.
He's just going to have a hell of a time using optics with it.
Thread recap for those who tl;dr:
Electical lighting was deemed too expensive, both in terms of equipment and operation.
Hence, sunlight is now the best bet as mirrors can easily collect a large amount of the light, while a prism and filter can create a rainbow with the right energy distribution.
The problem now though is that one must make it so that the sunlight always hits the prism at the same angle, so that the red part of the garden stays red and the blue part stays blue.
Four possible solutions so far, each with an upside but also a problem:
>Manual: Cheap, but imprecise and maintenance intensive
>Tracking system: Simple and flawless, but expensive
>Transparent cylinder with a low IOR exterior and high IOR core: Simple and may work according to simulations, but not simple to manufacture
>Optical Fibre: I really don't know what setup would be used, so I can't make judgements
>>908365
One good source of decent parabolic mirrors are the kind used in decent telescopes. Not the long skinny tube kind but the short and fat kind. With a decent parabolic mirror, a wide variety of angles of sun light will bounce off the parabola and travel through the vertex of the paraboloid. Just put your prisim at the vertex and get a stable rainbow. Or, put some columnating optics at the vertex so you now have a columnated beam of light, then put the prisim in front of this column of parallel rays of light.
>>908466
Can you explain or show me the mirror setup?
>>908657
Well, if no one can answer this, anyone able to advise on how to make IOR gradients?
>>908927
Fuck off OP. No wonder no one wants to help you with this. If it's really so hard to understand then I'll draw you a picture of the setup in paint *After work*. Until then, why don't you try to understand what a paraboloid is, why you can use it as a solar collector, how a negative lense can columnate a spherical wavefront, and how your going to get enough power through your water lense without melting the plastic and boiling the water.
Sage.
>>909011
I've run simulations of a couple types of parabolic mirrors, and the simulations indicate that there is no sole focal point for sunlight coming from multiple directions.
Which is a real shame, I really wanted them to work.
>>909106
well we only have one sun
and at the distance it is the rays can be considered parallel
>>909107
But for a fixed mount collector, one must consider the sun as being the sum of all possible locations, which gives you a scattering of rays.
>>909116
Only over the whole day. At any given moment the sun is sufficiently parallel.
>>909119
But then I need to track it, and people were talking like it was an alternative to tracking.
Found this giant satellite dish for $1, haven't bought it yet but I intend to.
How can I motorise it to track the sun without spending a fortune on motors?
>>909753
use step motors bro
>>909753
use 2 large buckets with water and weighted floats in them.... tie rope to the floats and then to the dish... use a pump to move water from one bucket to the other... as water level changes floats move and the rope pulls the dish along.... small pump takes longer to move the water, but gets the job done...
>>909758
if you want to get stupidly cheap with this use two light sensitive circuits to wire up a control for the pump... put one at the bottom of a tube and mount it to the dish (so it has to be aimed at the sun) set it up so that it runs the pump until it's receiving a decent amount of light, and then stops running the pump... set the other one on top and set it up so that the system only runs when thats receiving light (so it doesn't run the pump all night trying to find the sun thats not there... ) then set the buckets up with a bell siphon so that when one gets too full it dumps the water back into the first resetting the system....
https://www.youtube.com/watch?v=lyrvcCqv5V0
>>909760
I have absolutely no clue how that would work for tracking the sun.
>>909757
This. Stepper motors and gearing. People use that kind of shit on telescopes all the time so there should be plenty of info on the web about it.
>>909757
>>909830
If I use a stepper motor, how easily can it end up to a different state to what's expected (e.g. should have turned 10,000 steps but is actually a few steps below or ahead), and how easily can it be wrenched into an unexpected position?
If it's realistic that the motor will end up in a different state to what the controlling computer thinks it's in (Can still happen from the computer being turned off when the dish is in anything but a default position) then the computer will need some sort of error correction if I'm not to perform frequent checkups.
Strong wind is common where I live.
So, would it be typical for a step motor to be able to move to a stated position rather than just performing a specified rotation?
Things to google:
Spacial coherence basically how directional the light is
Temporal coherence how monochromatic the light is or light frequency range.
Intensity and irradiance will describe the wattage your getting out. The spatial coherence and the beam focus can wildy change intensities on the surface this light is projected onto.
Light guides and lenses for everything outside visible light are super expensive.
IMO go for LEDS, highly directional, pretty monochromatic look at their output charts.
>>909824
you put the dish on a swivel (small circle) that lets it swing to track the suns movement east to west.
each side of the dish is attached to a weighted float in a bucket (although a drum would work better) the weight of those floats needs to be enough that they will drag the dish on the swivel, but also have enough boyancy that they float at the level of the water... that way you can use a small pump to move water from one bucket to the other moving the floats and changing the angle of the dish via the blue hose, the red hose is just a clever way of getting the water out of the second bucket back to the first without using another pump or making you go out to reset it... it works because as the water level in the bucket rises it also rises in the red hose, once the water level gets high enough to get over the bend at the top it goes down the tube and since it can't let air in it sucks water out of the bucket until the water level is below the inlet of the hose.... you'd need a control to turn the pump on and off, and it would be a good idea to give it a sensor so that it won't keep trying to find the sun at nighttime or when their is tons of cloud cover..
I'm sure there are flaws in this.. but I was bored and this seems like it'd be a working simple solution that minimizes motors ... if you used 2 drums for the water you could use 5 gallon buckets with rocks in them as weighted floats... you could probably pick up a fishtank pump from harbor freight for under 30$ and you'd need some tubing... and rope... as to the sensor it'd be easy to set up one using 2 photosensitive cells... if you want to be memey about it you could use an arduino
>>910141
as to those photocells, they convert light to electricity, stronger light means more power, you'd put one on the dish with a tube over it so that only light coming straight at the dish would hit the photocell, set it up so that when the output of that photocell drops below that it runs the pump.. use the second cell without a shroud that turns the entire system off when the level of outside light drops below a set level (ie clouds, nightime, apocalypse)
>>910141
This would work well if the sun did move perfectly east to west, but unless I was on the equator and seasons didn't exist, this wouldn't work to enough accuracy for my purposes.
>>910151
you can angle the dish north/south and adjust for seasons.. what are you doing that requires enough accuracy to account for day to day differences in this? you could just adjust it once a week.. or is this going someplace that is not easily accessible?
>>910152
I live fairly far south, so the day-to-day difference could be big enough to be a problem.
>>910154
ah well damn.
seriously though I want to go try and build this system now... it was fun to think through, and probably a crapload cheaper than buying a bunch of stepper motors...
maybe try googling "how to track the sun" .. I tried and it pulled up a bunch of videos on DIY solar panel setups... you could probably adapt something like that...
>>910156
I'm thinking that I might only need two stepper motors.
One for X rotation, one for Y rotation.
I'll have the centre of rotation slightly below the dish, so the beam of light it creates always points towards the focal point of a convex lens.
Problem is though that unless I can use mirrors in a clever way, to collect light from dawn to dusk, the lens has to be very large.
>>910154
Are you trolling? The sun moves less north and south the closer to the equator you get. Or do you live in Antarctica?
>>910173
NZ.
https://www.youtube.com/watch?v=IKyNYCVbeAc
OP here, just found this.
I noticed that a lot of the light is absorbed though.
Is it possible to aquire lens-grade gelatin?
