DIY clear epoxy optics?

[snarfer]

I wonder if anyone has tried casting their own LED optics out of epoxy, such as this optical epoxy (pdf), which is advertised as 98% optically clear.

 

[G1K]

Assuming you could get all the bubbles out of the epoxy (easy to do in a vacuum), your cast will only be as good as the mold used to make it. The casting is easy compared to getting a good mold.

Just something to think about.

Ryan

 

[saabgoblin]

Hello Snarfer, is the contact info listed on the PDF a retail or wholesale contact? let me know if you find a good supplier because I want to start playing around with this idea as well. Pm me if you wish.

 

[RocketTomato]

You might have problems with homogeneity which would affect how well the optic behaves.

Some of the Norland Optical UV curing adhesives/epoxies might work better though they are expensive.

It is probably easiest to get a glass optic near the shape/size that you want and grind it down and polish it. Amateur astronomers have been building custom optics for years for their telescopes and would be a good resource to check into.

 

[saabgoblin]

For my purposes, I am looking at filling in gaps in the Peak Led 5mm lights with Glow in the dark powders or other colorants to reduce the side spill of these lights. I prefer the beam pattern of their older recessed leds as opposed to their newer offerings so I want to fill in these gaps to see what effects I can produce with the more efficient snow leds.

 

[Oznog]

The external optics seem to be all acrylic.

Most general purpose adhesive/composite epoxies are either a yellowish tint, or will soon turn yellow when exposed to sunlight. Obviously since epoxy-cased LEDs were the standard for decades that other products are available- just not sure if they're easy to get.

Scratch resistance is often an issue. Actually I wondered about refractive index, and was surprised to see that epoxy and acrylic were quite similar.

Anyhow, my note there is that if you need to machine a mold for an epoxy optic, why not machine a block of acrylic for your optic?

I've done casting with a vac pump. No prob but you can add bubbles by trapping air at the mold-resin interface if you're not careful. There is also pressure casting, where they put the filled molds under high pressure to shrink the bubbles. You can't vac pump a filled mold in general, the expansion forces a bunch of your resin out of the mold then all the bubbles break and it falls back but your mold may be missing half the resin. So unless you've got a huge conical overflow reservoir on top that will happen.

The clarity of the result depends on the mold having very smooth, glass-like casting surface. Unfortunately 3D printing doesn't leave that in the model. You might spray it with some sort of surface coating that will be smoother before using it to make your molds. The great thing about acrylic is that the right chemicals or flame can polish it, whereas epoxy only responds to mechanical polishing which can be tedious and very difficult to get into the nooks and crannies.

 

[snarfer]

Just to put this into a little better perspective, what I'm trying to do, and I'm sure I can't be the only one, is to find some cost-effective way of producing prototype LED optics according to my own designs in small batches, up to say 50 or 100 pieces at a time. The project also includes creating a software tool that will optimize the shape of lenses and/or reflectors with the aid of the Differential Evolution algorithm.

I have talked to some companies that produce polycarbonate optics, and they charge something like 5,000 to 15,000 dollars to create a custom mold. That's completely out of my price range at this point. Of course once you pay for something like that, the cost per optic is tiny.

On the other hand, there's no way I want to be machining lenses one at a time. That would end up taking a really long time and cost a lot on a per optic basis.

So I'm trying to find some happy medium between the two alternatives, something I could do for few hundred dollars worth of tools and materials and get halfway decent results.

Most general purpose adhesive/composite epoxies are either a yellowish tint, or will soon turn yellow when exposed to sunlight. Obviously since epoxy-cased LEDs were the standard for decades that other products are available- just not sure if they're easy to get.

An engineer friend told me that the urethane based compounds will turn yellow no matter what the manufacturer says, but that the epoxy products should be okay.

Some of the Norland Optical UV curing adhesives/epoxies might work better though they are expensive.

Norland has some very interesting products, but it looks like they are adhesives, not for casting.

Anyhow, my note there is that if you need to machine a mold for an epoxy optic, why not machine a block of acrylic for your optic?

This is absolutely a fair question. It might indeed be more cost-effective to machine plastic parts than a steel mold.

One possibility might be to create an original by hand, and then use that positive to create molds, thus avoiding machining altogether. I am really hoping that I won't have to have any machining done, because it is just too expensive and time-consuming.

Unfortunately 3D printing doesn't leave that in the model.

It's too bad about 3d printing not leaving the proper surface. It seems like that would be the process that really makes the most sense in terms of designing the part in software. I am going to do some more research on that. It could really solve a lot of problems, and keep me away from the dreaded machine shop expenses.

 

[gadget_lover]

Just to put this into a little better perspective, what I'm trying to do, and I'm sure I can't be the only one, is to find some cost-effective way of producing prototype LED optics according to my own designs in small batches, up to say 50 or 100 pieces at a time. The project also includes creating a software tool that will optimize the shape of lenses and/or reflectors with the aid of the Differential Evolution algorithm.

I think you are taking on a task that is more complex than you think it is.

If the objective is to create batch of optics with known qualities, then you need to very accurately reproduce the shapes produced by your calculations. This will be nearly impossible to accomplish by hand, though you will get lucky at times and come close.

The beauty of ACCURATE computer modeling is that you should not need to make a dozen tries to get it right.

A few hundred dollars will not get you the necessary tools to create either the molds or the machined prototypes. You can, however, get a CNC mill for under $3,000 and a CNC machining center for under $10,000. Then there is a lot of learning to do to learn how to use the CNC hardware.

Part of what you are paying for when the shop charges $5000 for a mold is their expertise. They will be able to advise you on things like compensating for shrinkage/expansion, getting the part out of the mold, polishing it, etc.

You can turn some plastics by hand on wood or metal lathes, but your results will not be consistent nor repeatable.

Daniel

 

[snarfer]

If the objective is to create batch of optics with known qualities, then you need to very accurately reproduce the shapes produced by your calculations. This will be nearly impossible to accomplish by hand, though you will get lucky at times and come close.

I agree. The two most viable possibilities that have emerged in this thread, so far, are (1) CNC machining, which is, I believe, the standard way to do it, and (2) rapid prototyping, i.e. 3d printing.

DIY 3d printers (Fab@home) may not have sufficient accuracy, but there appear to be quite a few companies offering rapid prototyping of plastic parts at cost low enough (something like $50 a cubic inch). It would add some intermediate steps to the mold-making process though.

Part of what you are paying for when the shop charges $5000 for a mold is their expertise. They will be able to advise you on things like compensating for shrinkage/expansion, getting the part out of the mold, polishing it, etc.

Absolutely. This is the expertise I would like to develop in-house, by trial and error if necessary. Because when you have a specialized light that you are lucky to sell 50 of, it doesn't make sense to spend 5-15,000 on a mold that's only good for that light. Even if the lights can be sold for 1,000-1,500 dollars each, there are lots of other expenses to deal with.

Because we are investing in making the housings for the lights out of composite materials, carbon fiber and lightweight bent aluminum, we are going to be doing a lot of casting anyway. So it would make sense to cast the optics as well.

Of course I realize there's a tradeoff too. There are companies doing amazing things with optics that we would never be able to reproduce without serious cash, or perhaps at all. For example, growing reflective crystalline structures on the exterior surfaces of a polycarbonate optic, refractive indexing, and so forth.

 

[OddOne]

If you have the ability to calculate out how to shape an optic and have access to machining equipment, you could locate some cast acrylic rod and machine/polish your own. Cast acrylic rod in optically-clear form is pretty easy to locate and relatively inexpensive.