Monday, February 16, 2009

Glass Onion - Glueless Model

Friction joints shouldn't be cut using the lasercutter... It lacks consistent precision. Even when cutting different increments, not all the notches came out the same and I had to sand them all down to make it fit. Surprisingly when all put together, the gear mechanism self stabilized with the cumulative friction. In another words, they stayed together fine but when one leaf fell out, the whole assembly did as well.

Inspired by HeeJo Shi's Marble assembly.


  1. Hayes,

    Lasercutters can have several issues: Power,speed and Z height of the beam can vary. Material thickness is not always consistent.

    You can gain more control in a couple ways: try an array of tests where you try out your critical detail connection but you change the dimensions by, for example, 1/100" each time. One of the tests will probably work the best. Then you input that information in a 2nd test. You should also ask for the power and speed information when you did the test and specifically request it again.

    Iterate and correlate machine, material and geometry.

  2. hi josh,

    thanks for the response. i did several lasercut passes, each varying 1/128" at a time. what I discovered was the problem was with both the material and the limitations of the consistency of the laser itself. The big problem was that even with a drawing with consistently sized notches, the laser cuts wider still in some areas of the bed than others. I ended up with a cap piece in which half of them fit and half didn't. When i moved to the next thickness in, a completely different set of notches fit and others didn't. Tito says it has to do with distance of the laser head from the source, the topography of the bed, the depth of material, and the newness of the lens. I also work there and have the luxury of doing these multiple iterations instantly, and still there were too many parameters for me to consider.

    i think the lesson that i'd have to impart is that don't design critical joints such as these expecting it to work perfectly. IE: slotted joints etc have more tolerance than the mortice and tenons that i used. i just wanted the cleanliness of this particular detail and paid for it with several hours of sanding.

  3. i see. there are too many variables and the results just aren't consistent. however, if you were lasercutting steel or plywood these errors would be relatively small. jumping in scale would probably produce a greater degree of accuracy. another source of (small) error is how the machine leaves a kerf. There is nothing to be done about that no matter what. The waterjet does this too ofcourse and so a truly perpendicular cut isn't possible. Tablesaws will do you better. And so will the mill. The mill is by far the most accurate of any machine the school has, but it doesn't do 1/64" wide cuts like the laser :) So, we're caught between accuracy and fineness.

    We'd really like to see you take these investigations further and understand how you can deal with them with a different machine and/or material. If you can achieve basic consistency and precision, what then? How much of a good fit it too good? For example, you might prototype one joint that fits very well on the mill but then you remember that you need to do 6 at once for your detail and the cumulative pressure required to make the detail assemble is very high. Do you 1) make your detail fit more loosely so that assembly is easier or 2) develop a tool or assembly method that allows you to exert force in a smarter/stronger way?

    Both might be valid.

  4. nice break down... i guess that #2 was what I was suggesting. it may have more significant implications ( or at least broader ones) for architecture since we can't all do projects where we sit onsite with a computer and babysit the GC.
    The material matters a great deal... wood would have had far more flexible properties and the friction joints would've been far easier to craft. plexi- for which there is no full scale equivalent, is brittle and has much less surface tension, making precision more important, but understanding that disparity is a lesson within itself.