Mechanics: (almost?) final chapter

The final design of the mechanics of SpiROSE have finally been chosen! A week full of design decisions, finding a design, testing the mechanical
design, simulating, talking with the mechanics to see if it’s doable, etc.

There have been lots of designs, the most importants are the following.

Blade-based designs

The design pattern is based on a stack of RoseACE blades. Every blade contains 32, 64 or 128 LEDs, and the stack is made of 32 to 42 blades. This
leads to a design looking like the following.

Blade-based design pattern without top arm.

The red squares are the RGB LEDs. A big central axis is used for turning the LEDs and transmitting the electrical ground. A special PCB at the bottom
of the axis is used for the Single Board Computer module, an FPGA and a brush for transmitting the Vcc from the fixed base. Wires connect the blades
to the main rotating PCB.

This design is similar to the design made by Volumen Display. It have the advantage to be modular: you can add or remove blades without changing the
mechanical design of you system, and another advantage of being more aerodynamic (it acts more or less as a big 30cm diameter fan :).)

The issues with this design are mostly mechanical issues. In fact, it is extremely hard to make a stable design this way because of the approx.
1300rpm speed of the blades. The mechanics said it would wobble, and the solution making the wobble to disappear (adding another bearing on top of the
axis) introduces other even harder problems (how to fix the PCBs). The other issue is the occlusion. The central axis makes a big occlusion and the
helicoidal shape of the blades overshadow the lower LEDs, giving the feeling of a big blurred central axis.

Another blade-based alternative design was to stack custom plexiglass(r) laser cut discs, containing the PCB. It looks like the following. The Disks
could be screwed to each other up to the base circular PCB, making the axis. The issue with that design is that the plexiglass would obstruct so much
the light coming from the LEDs that it makes the feeling of an even bigger central axis, instead of solving it…

Plexiglass(r) based axis.

Due to the issues related to this design pattern, it was finally removed from our design space.

Blade-based pattern with top supporting arm.


Plate-based designs

This design pattern is a HARP-like pattern. A big central plate gives a matrix of LEDs. This plate is turning on itself and makes the same shape as
the blade-based design makes. This pattern needs a supporting arm on top, leading to the following design.

The difference with HARP would be the number of LEDs, and the fact that our goal is to remove the occlusion issues they got. The occlusion they got
was due to multiple effects:
– The fact that LEDs are on one face of the PCBs makes the system not able to render anything when the LEDs are not visible to the user’s eyes. The
solution we got was to use two PCBs with LEDs, back to back. It makes possible to always have LEDs facing the user’s eyes.
– The fact that the LEDs have an emission cone (approx. 120°), giving this feeling that there is kind of a central occlusion because the LEDs cannot
illuminate when the user’s eyes are not in the emission cone. To solve this issue, light guides/diffusers will be experimented on LEDs next week to
see what we can do (this subject will be discussed in another post :).)


The winner

The winner of these designs is the Plate-based design, having less mechanical issues. The mechanical design approved by the mechanics is the following. It will be changed a little bit due to motor changes, Budget adjustments, etc. but it will be looking like this one.

Next week is not anymore on the mechanical design, but on the embedded systems and budget choices!

See you next week for more fun!


Leave a Reply

You can use these HTML tags

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>




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