Lux: A 100% Open Source Camera


[Kevin Kadooka] recently finished his open source camera. The Lux Camera is 100% open source. Lux uses no parts from other cameras – not even a lens! To date we’ve only seen this with achieved with pinhole cameras. [Kevin] isn’t new to camera hacking. He was the man behind the Duo camera, which had a successful Kickstarter campaign in February of 2013. Duo is a DIY camera, but it still required lenses from Mamiya-Sekor, and a shutter from Seiko. Lux is a different animal. It has a manual focus 65mm f/5.6 Single Element lens. The shutter is [Kevin’s] own solenoid based leaf shutter design. Just as in the original shutter, an Arduino controls shutter operation and timing.

The main camera body and many of its parts are 3D printed. [Kevin] got some very nice quality parts from Shapeways 3D printing service. We have to say that some of the assemblies look a bit complex for desktop printers. However since everything is open source, anyone willing to put the time in could adapt them for the average RepRap or Ultimaker. [Kevin] has posted detailed build photos, as well as some photos taken with the Lux on his flickr stream. The pictures have  a decidedly holga-esque look to them, due in part to the single element lens. Even with this limitation, we love the idea of having a brownie style camera built completely from scratch.

29 thoughts on “Lux: A 100% Open Source Camera

  1. I love the look and execution of this project, but I fear this is where it ends as far as optical quality goes, which is a real bummer considering the quality that typical 120mm film can produce.
    The odds of maker technology in the near future being able to even equal late 19th century triplet based optics is very very small indeed.

      1. Even then, the range of options is limited and the net gain seems negligible. Given the expense it’s not really worth it.
        If you could get a primitive double gauss type formula working from stock glass, I’d be impressed. I just do not think it can be done in a way that gives a result that matches a properly designed lens from a century ago.

        1. Not entirely true: You can design around stock lenses, but you’re often limited in materials (Silica, BK7, possibly Flint/Crown glass if you’re lucky) using software. It’s not exactly trivial, but definitely possible (and not terribly complicated). I believe there are reference designs available for camera systems, so you’ve also got some place to start.

          1. Yes, that was the definiton of “limited” I was looking for. As I said, using off the shelf lenses even is a problem, especially for 100% openness. With a lot of effort at some expense, you could potentially build a primitive cooke triplet, but late 19th century forumulas like the planar or even tessar are way out of reach.

          2. Ah ok. I was thinking along different lines. Trying to produce something outside of stock radii would be prohibitively expensive, unless you’re a large company, or just happen to have all of the necessary grinding and coating equipment lying about.

    1. No doubt they’d have to be hand ground, but that’s not beyond the grasp of any number of makers that frequent HAD. It’s a matter of doing the maths. Same with printing the enclosure/mount-points for a triplet. Not all that difficult beyond not being able to print glass out of a nozzle… sheesh.

      1. Doing the math isn’t the issue really. It’s just that we won’t have a method, even for hacker spaces, to easily create the glass that matches the result of said done math, even for simple ancient formulas. Even without cementing and coating, it will not be done.
        Granted, 3d printing will make producing housings and perhaps even helicoils a remote possibility in a practical sense, but still you will not see anything beating cheap 1920th optical tech, I’m sure.

    2. You are right; finding suitable optics is a big challenge. I chose a positive meniscus lens mostly because it was inexpensive (<$10 and AR coated? A steal if you ask me), and I knew it would work.

      Using two achromats in a rapid-rectilinear arrangement may have provided better quality and speed. The problem then is placing the shutter between the two doublets, which complicates things somewhat – I didn't want to deal with that. Besides that there are other "simple" optical layouts that could be replicated, the Cooke Triplet and Petzval also come to mind. Already, those are several magnitudes more difficult to implement, even using "off the shelf" lenses from Edmund, Anchor

      I agree that doing the math is not difficult, there are ray-tracing programs like ZEMAX that can be used to design complex lenses. Establishing a glass foundry to make the various types of flint and crown glasses, with the correct curvatures – that is another matter entirely.

      A step in the right direction would be to find (or make) a positive meniscus achromatic doublet. The shape factor of the pMNS lens reduces spherical abberation, while the achromatic nature reduces chromatic abberations, to a degree.

      1. Well, either way, this is outstanding work! If I can find a suitable time and place (i.e. someone with darkroom access), then I will be building one of these.

        I wonder if you could get away with molded aspheres of various plastic compounds? You can get polymers with a wide variety of refractive index, and making the molds is a little easer than grinding your own lens.

        There was a post here a few years back about a guy who was molding parts out of resin, but I can’t seem to find the link at the moment. His process would be suitable for this kind of precision optics work.

  2. love it, great project.

    as far as the arduino instead of the 555, my guess is the ease of modifying code or adding any future features that he may not have specified in the last part of the project. This is one project that i can see benefiting from using the arduino over the 555

  3. I’m super proud of Kevin!
    I’ve been mucking around with a camera design for ages, and he’s cranked out two in half the time.

    Really cool. Keep up the great work.

  4. I was excepting some kind of digital camera. About a year ago I was looking for a digital camera sensor with suitable resolution and characteristics; however, as I have understand, 99% of image sensors available for hobbyists don’t have any datasheets.

    1. Even in industry/academia image sensors are hell. If you sign NDA’s you can get documentation, but it is often so poor that you really need a FAE (or lots of time) to make it work.

  5. His STL files look like they’re going to be impossible to print on an FDM machine. I’m going to try, but I don’t think this is going to be reproducible without an SLS or DLP machine.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.