VIS Project Build — And why not to buy from Ebay.
Part 1: The proposal
VIS is a project I worked on with my classmate Eden. It was created as a conceptual design for an assessment in my degree — Computational Design.
Our prior assessments were about social media analysis as well as site analysis. The site we were assigned was the University Mall (UNSW Main walkway).
Initially, my partner and I had different ideas, I wanted to create a light installation that would be more low-key, and my partner wanted to do some sort of clicker count game with the tiles of the walkway.
Unfortunately, our teacher didn’t like my idea, it seemed like she wanted something more unique or innovative.
So as a joke, I proposed to Eden that we just fill the walkway with LEDs and light the whole thing up. Turns out our idea actually had some beneficial outcomes and we had some good ideas too. We both were fans of the movie Avatar, which is built on a world with unique bioluminescence that would respond to movement. So we decided to draw inspiration from the movie.
The proposal is to spread led roots throughout the whole walkway. Which will respond to movement and light up.
To keep things simple, I was in charge of the prototyping, as I have experience with electronics.
Part 2 : The Build
The idea for the prototype was to create something that people could actually stand on, and represent a small section of the large walkway. The original plan was to use camera trackers to detect movement and location, however due to time constraints it wasn’t viable, especially with the layout of the root system. Instead I opted for push buttons embedded below a platform, which will trigger depending on the tile. For lighting I went with WS2812B LED strips, which are addressable meaning I can control each LED’s three color channels.
The main parts of the build were actually ordered on Monday the 9th Sept, 3 weeks before the project was due. I chose to use Ebay since it generally has cheaper electronics.
As I didn’t want to wait for shipping from china, I chose items located in Australia, hoping I wouldn’t have 5 week shipping times.
The initial parts order was:
- 10 Push button switches
- 12V 8A Regulated power supply
- 10 meters of Low density WS2812B 12V
- 50 pcs of W2812B in a string 5V
- DC to DC Buck converter (To step down 12V to 5V)
Overall costs not looking too bad.
So after placing the order, all I could do was wait for the parts, or research some information about the project.
I found an Arduino library called FASTLED, which allows me to convert the led’s into an Array, and call functions for the color selections etc, this saved me a lot of time bit bashing.
After a few hours, I was greeted with this message.
Oh dear, this was for the power supply. Which was a shame as only they had 8A+ 12V power supply in Australia.
This was a bit of a problem, as the 8A was needed to power al 350 LEDS.
This was calculated with the following:
20ma per channel (RGB) = 60ma
60*350 = 21A
However I wasn’t going to use full brightness, nor will all the channels be on at the same time.
I divided it by 3 to get 7A (50% brightness and 1.5 channels on).
That would have given me 1A of headroom.
Instead I had to settle for this, a 12V 5A shoddy power supply that would never be allowed to run in a UNSW engineering faculty. But hey! We’re in the built environment faculty. They probably won’t care.
So with all that ordered, projected arrival time the same week on Thursday to the next week’s Wednesday, everything should be a breeze right? Well…
The power supply arrived, the DC buck converter arrived, the tiny buttons arrive but where were the LEDs?…
I messaged the seller and they told me they didn’t have tracking for it, great… These sellers are usually just Chinese distributors who drop ship them overseas, however they usually have warehouses in Australia which is why they put the item location is Australia. It seems their services are still extremely slow, granted marginally better than shipping them directly from overseas. (Have to pay the premium of course).
That’s okay I thought; I’ll just get started with the other stuff.
After having a long discussion with my classmate about the materials and how we were actually going to create the platforms, we settled on white arcrylic, or perspex, which would allow light to pass through if directly placed under. They had a 900 x 600mm acrylic sheet which was $55, however it was not large enough for the scale we wanted. I found a PVC palopaque, which I didn’t know if it was strong enough to step on however I wanted to have a look anyway in store.
So off to Bunnings I went.
And.. they didn’t have them.
Not the Acrylic or the Palopaque, in fact none in any store close by, I was panicking as that day was the only day that I could visit Bunnings. So looking around I found another material, Polycarbonate roofing.
I was unfamiliar with poly-carbonates, but the material was quite strong and when a light passed through it, there wasn’t a yellowing like some other materials. The material was also lightweight and extremely durable (Which I found out when trying to fit the 2.4 metre long roofing in a 2 metre Toyota Echo).
This turned out to be a great idea, as the perspex probably wouldn’t have been able to handle heavy weights, nor let light pass through unless it was directly pressed against the material (The led strips had a strange way of bending sideways). I also purchased a long piece of timber, for the platform supports.
So after getting home and kindly requesting the car to be parked outside the garage for a while, I got to work. Bringing out the 12 year old shoddy table saw bought at some discount store, I started cutting the tiles and the timber. The tiles cut fine, but the timber chipped a lot and I got lazy with the measurements (A huge mistake).
I ended up with 16 tiles and 35 supports, I decided to use only 12 tiles, but if you do the maths you’ll realise that it still isn’t enough for 4 supports per tile.
I ended up deciding to use 2 supports on the left and right side, and 4 supports to the center row, leaving 3 supports for step detection.
So all that was done, still had to wait for the electronics however. I setup the DC to DC step down and everything but the led’s arrived.
But they didn’t, I started panicking and I ended up buying a high density 5M led strip with express shipping (Which costed a fortune). It would still have the same amount of led’s but in a tighter area. I was also faced with another problem… the new LED strip is 5V,, meaning I had to go to Jaycar and buy one of their DC buck converters.. This was okay but the 5 converters are only supposed to handle 2A, with 3A being it’s limit. This might have been a problem but I had to take the risk anyway.
They didn’t arrive that week either. I was stressed throughout the entire weekend, but I was preparing another assessment for Monday 23/10.
Finally it was Monday, the postal service was alive and I could do nothing but hope. After getting on the train home, I messaged my brother, who told me a package had come, but it wasn’t the LEDS.. I had almost lost hope but I received a text to my phone saying they have a parcel pickup from Killara. The strange thing was that I don’t live in Killara, so it was strange that the tracking made me pick it up there, I went to the post office and turns out it wasn’t my package either. I headed home for Seppuku.
At 4:30PM I heard a motorcycle. It was the mailman, who usually delivered before 3PM. I don’t think I’ve ever felt so relieved about a package arriving before. They were the 10M of leds from 3 weeks ago.
So I began to build.
Unfortunately, the 5A power limit still existed, so I decided to only use 5M of the low density 12v LEDS. I cut them into various lengths and soldered them on. Fortunately for me, I found out later that the led’s can run in parallel if soldered on the end of another strip, meaning they branch out and still inherit data from the parent LED.
So one night before the project was due to be presented, I started working on it.
First thing was the power supply, I cut an extension cord and attached it on to the power supply, just hoping I got the right earth/neutral/hot connections. I connected it and hoped it didn’t explode. Success.
The DC buck converter was also reading 5v from the 12v step down. Success.
I then had to solder the led strips, which was a real pain considering how close the copper contact points were, but with some persistence, I managed to do it. Then again for the other LED set. Success.
Now to setup the buttons for Arduino.. however I realised that my buttons I ordered were so small it was almost impossible to solder.
The issue was that I ran out of copper cabling so I ended up using garden wire fasteners for long cabling, which was a rigid material that liked to break the solder every time I touch it. Instead I tried something else out, which turned out to be an abomination
It broke almost instantly as the aluminium was way too weak, it didn’t hold a great connection either. I ended up making these;
Which I soldered with tiny copper connections. All seemed fine but then… The platforms were too heavy, when they pressed down on the button they wouldn’t come back up. So that plan was foiled…..
I ended up using some large buttons that came with the Arduino kit, and to attach them on the wooden support I had to carve in a ditch (Chiseled using a screw driver) and drill a hole in the middle to offset it below the surface.
They were also much easier to solder with, and it seemed like a success.
After some more programming, I managed to get the chain effect I wanted, and it was morning already.
I arrived an hour early to setup, originally there was going to be another element to attach the 50 chain LEDs to, however they short circuited every time I connected them to the power supply, and made the DC converters smell like burnt silicon, so that ended up being omitted.
After finally being able to route the Buttons correctly, wire up the power supply and re-solder many connections that broke. It was complete.
Thanks to our tutor who allowed us to present last so we could do more preparation and let us present the project where it was setup.
Here’s a video of the prototype in action:
Other issues throughout the project:
- 5v + 12v Grounding problem.
- Really terrible tree
- Lack of sleep
- the RGB channels of one of the led strips actually being RBG.
- Not wiring a push button correctly and making a distance sensor by mistake.