Lets Build a LED Light !!?
 

Let's build a light from high power led. I would like to do this as I want to learn more about this technology. I'm sure the end result will be superior with a group. There are pre-assembled components of many types available. Here's my first thoughts:


- color temp changeable
- dimable
- flood or spot depending on LED lens choice
- scalable

A high cost video lighting system I've seen uses warm and cold leds to produce a variable range of light temp. This can be done manually with dimmers, or through software control. Lights are typically controlled by the DMX protocol. There's nothing to invent here, but the proper user interface is the question.

An alternative is to use RGB leds. These likely have more color cast issues with white, but can produce all colors.

The light distribution pattern is mostly determined by the lens on the LCD. These come preassembled or can be added.

To meet the requirement of variable color temp a minimum number of cool and warm LCD will need to be used. This would be the smallest unit.

Thermal and case design are likely the most custom elements.

What's right with this plan? What's wrong? What's missing?

Venture capital, a website, business plan and reason why we should do it when we have guys around like Richard Andrewski who are already on it ;-)

Dan

Dan.... I just can't take you seriously with a post numbered 666.

I've tinkered a bit with diy led lights.... but as soon as I get trying to build a sturdy yet light casing that looks decent I run into high tooling or machining / casting costs. I've even considered the compucarve cnc machine ( $1800 at sears ) and cnc'n from polycarbonite. But.... it would take about 10 hours of milling per unit. + the bench time on the electronics.

I've found that {my} time is best spent shooting and then spending a percentage of the earnings on gear.

I would however love it if zylight would come out with a 1'X1' version of their z90.... with oh.... 17000 lumens

you only have to machine the first one, then you hire out the job to a CNC machine shop. 10 hours manually is probably 20 minutes on a CNC, and if you get a quantity, the rates are relatively cheap. I have noticed posts on this forum from at least one person in china that would assist in outsourcing the labor. if you can make one working prototype, you can most likely get the financing to produce the first run of lights.

to build a led light is darn easy (i builded already a dozen of them), the problem being that everybody will ask for a different model to fit his needs.
and if you build a one-size-fits-all, it will be too bulky, expensive for most of people.

Yikes, just noticed that. The sign of the beast. Making LED lights, hmm...are those somehow related? ;-)

Dan

I would love to see a high quality video specific LED light come to market at a great price.

Keep in mind what is out there for cheap:

All Products Lights LED Fixtures | | CHAUVET® Lighting

Mid range

Teleprompters - Prompter People

And on the higher end

Element Labs
http://litepanels.com/

Well, at least Giroud has the right spirit. :)

My only experience with LED lights is my flashlight...

But from your description my main concern would be the time needed to change from flood to spot if it requires a lens change. If you have to run back to the truck and look for a part to make the change, that could be a real time killer on the set.

Keep us posted on your progress.

OK. Let's see here. 4 years to the day since I harped on about building LED based lights for video use,on CPF, I'm actually starting to see some amazing progress in this direction. Even Lightpanels read my white paper and now have a bi-color panel for nuanced color output.

Word of warning. DIY lighting is so addictive and involved, I doubt you'll have time to do anything else once you're hooked. So if you like making movies, avoid this area of DIY.

Before you begin. Spend a year or so browsing CPF (what's left of it) and get to know the engineers. As Guy mentioned, the ball is already rolling and most main players have figured out what to do. The best thing that users can do is input useful feedback from the trenches.

Oh and yes, the parts do cost about 1/1000th of what the retail product sells for.