About my blog

New depth sensors such as the one embedded in the MS Kinect system work in real-time, enabling to create an instant 3D scan of objects. In this project, we are exploring a large-scale installation that demonstrates this by a 400+ array of motors and LEDS. This work includes interaction design, hard-coding and lots of soldering etc. The result will be put on exhibit on various international shows.(Expert: Jouke Verlinden)




De meningen ge-uit door medewerkers en studenten van de TU Delft en de commentaren die zijn gegeven reflecteren niet perse de mening(en) van de TU Delft. De TU Delft is dan ook niet verantwoordelijk voor de inhoud van hetgeen op de TU Delft weblogs zichtbaar is. Wel vindt de TU Delft het belangrijk - en ook waarde toevoegend - dat medewerkers en studenten op deze, door de TU Delft gefaciliteerde, omgeving hun mening kunnen geven.

Kinetic Mirror

Project Evaluation

We all chose this minor because we like to build
things or want to learn more about how to build things and that’s exactly what
we did at this assignment.  We were lucky
we had a good team, everyone had different skills and we could fill each other
gaps and necessities.  From the start we
could easily understand each other and make a good planning for the project. After
the first week we have analysed the problem we divided the work, but in the end
we can say that everyone  did a bit of
everything.  From the start we knew we
will get busy creating the frame for a big scale kinetic mirror of 400 pixels
and also make a better version of light diffusers. That was it. It would have
been nice if we also learned some more about the programming as well. We thought
that we could do some research with the software and mechanics of the actual
prototype of 16 pixels but we were just too busy building frame prototypes. The
process was very simple, we start with making a program of requirements
(forces, weights, design) than conceptualisation and finally making a
prototype. We could not get to building the real one but were not ready to make
it either. We learned while making the frame mock-ups that always something
goes wrong and that we have to do a lot of trial and error mistakes to get
nearly to something that looks like your design.  In this project we learned to use new
machines like the thermoforming machine. 

this was a fun and instructive project to do. We were able to actually face
real-life problems and bring our imagination to the real world making something
that will be used international.  We are
very pleased with our learning’s and achievements during this project and hoop
to have the chance to build the real frame in a near future.


Project Evaluation

We all chose this minor because we like to build
things and that’s exactly what we did at this assignment. But it would have
been nice if we also learned some more about the programming as well. We were
just too busy building prototypes.

Nevertheless this was a fun and instructive project to
do. We were able to actually face real-life problems for example producing the
Led-diffusers. Something in theory can seem quite easy,
but when you scale it up to 400 Led-diffusers the costs suddenly either
sky-rocketed or the production time would become way to long! In the end we
believe we found a good solution to the problem and were happy to show of our
final project.

This project isn’t over yet! We knew from the start
that it would be highly unlikely that we would have been able to deliver a
finished product in just four weeks time. Besides the delay in the delivery of
the electronics there are still some minor issues we need to work out. For
example the assembly of al the pixels and the construction of the actual frame.
This project may technically be over, but we hope to keep contributing to the
project and be able to assemble the actual kinetic mirror early next year!


Science Fair

(unfortunately the picturs do not yet work, will be updated soon!)

Despite that the 400p kinetic mirror was not
completely finished, the science fare was a great success! We did have a lot to
show, for instance a scaled version of the final frame. We made this small
frame out of steel beams with bolts as connection points, so we could show
those interested how our final frame could be diss-assembled. On the bottom
plate there is room for the powersupply, mac mini and some watertanks for

Besides this small protoype, we made a big mock-up of the real frame, to show
the actual dimensions of the 400p version. Inside this pixelframe we placed a simple
white screen so we could project virtual pixels, in this way we could show the
guests a virtual model of the 400p mirror. The beamer was connected to the
Kinect and the Mac mini, the virtual pixels were created by the same software
that will eventually drive the motor fader pixel, so our virtual display was
allready modeling the persons standing in front of it. If the user interacted
with the Kinect-sensor this would be realtime projected on the screen. Bellow
is a photo taken when someone placed his arm in front of the Kinect sensor.

We also showed the original prototype with 12 pixels and our new led diffuser
caps. By switching power on and off, this prototype kept initiating the
start-up check over and over again to show the moving and flashing pixels. Next
to this we played a movie of the 12p prototype reacting on the movement of the
yellow towel.

All these elements together gave the guests a good picture how the 400p version
would be in reality.

It was interesting to see what our fellow students worked
on this period. All the stands looked great!

Renders of our frame designs

Here is a render of our final Frame concept




 Before we came with this design we had a pipe like frame, but becouse our scharp angles was very difficult to make.

As team we had to make hard choices while designing, trying to acchive a light weight, strong and easy to build frame design.




Preparation Science Fair


The first 1:1 mockup frame of the final concept. 😀
A quickly made tool for perfect allignment and glueing the diffusers with its brace. 
LED-diffusers in progress. 
The first twenty LED led-diffusers are done! Together with the moulds en self made cutting mould to obtain same height distances between all the led-diffusers.
And a final overview of the twenty diffusers with the brace glued. 


The past week we’ve been working on a lot of aspects of the frame, firstly we made a more detailed mock-up of the 400 pixels:



Did some impact tests to see what the worst-case scenario would be:



Worst case scenario: if everything goes wrong and all the 400 pixels receive power without the control of microchips we are looking at a maximum force of 2 kN. Next step was to take the previously posted frame mockup and run it through some simulations:


Results above showed some major deflections if the frame was subjected to this loading. This prompted another quick brainstorm session at IO:




Idea came up to do something with the frame of an adult swingset, sturdy cheap and actually quite aesthetically pleasing. Below a quick sketch of what we are currently thinking off:



Power supply

Calculations made in conjunction with Rob Luxen has shown us that we’ll be needing at least three computer power supplies. Below are some images of the ones we’ve chosen:




Prototyping of Led diffusers

the cheapest and best way to do this was with thermoforming process. Then
by a lot of trial and error experiments with thickness and types of plastics we
came out with the solution: "vikureen plaat" of 1 mm thickness.



Here is a 3D concept of the way we want to make and connect  the LD’s:

We make a real study of how the light duffused with this LD:



Now the BIG CHALLENGE is to make 400 LD’s. We will try to do 100 times 4 LD, because the
method to have a non-folded diffuser requires precision and 2 people.

Digital Mock-up

Been working on a digital mock-up of the installation. Trying to get a feel for the size and shape of the display and the frame that will support it.



Yesterday we made some prototypes of the led diffusers. The diffusers were formed out of plastic sheets by vacuum forming. We made different samples which we varied in height, fillet and draft angle. 

It was hard to release the molds and the sides of the cubes were folded. But the top of the cube, the most important side, looked pretty good. Our next step is playing with the height-dept relation till we get no folds at the side anymore.


© 2011 TU Delft