Technica Ball
This year, Technica was the world's largest all-women hackathon, and I thought it deserved some sort of cool piece, similar to what I built for Bitcamp. After a little brainstorming, we decided to iterate on an idea that began at Technica 1 in Fall 2015.
Towards the end of the event last year, a few other guys hung a large white china ball stuffed with LEDs over the venue. The LEDs played a spinning blue/pink animation, which was meant to be a spherical interpretation of Technica's gradient circle icon.
This was a neat concept, but had plenty of room to expand, and it would need to expand to fill the new massive venue the event was moving to (from Ritchie Coliseum in 2015 to Reckord Armory in 2016). The first iteration used a 30" china ball and maybe 150 or 300 RGB LEDs. My version would be bigger and brighter, with an 8' beach ball and 1,200 RGB LEDs. A big goal was to be bright enough that ambient light would not totally wash out the color, which was a problem with version 1.
My design consisted of a cardboard tube with the LEDs wrapped around it suspended inside of the beach ball. The tube would hang from a cap piece on the top of the beach ball. A pair of flanges would be installed on the top of the beach ball, which would create an opening that can be used to access the inside of the ball and insert the LED tube. The opening is sealed off by bolting on the cap, which features a filling valve and airtight power and USB passthroughs to run the LEDs and reprogram them on the fly if necessary. Three of the bolts used to attach the cap are eye bolts, which the entire assembly would hang from.
The air containment idea of this three piece cap assembly is as follows: two ring flanges clamp onto a hole in the ball, which creates the opening and interface for the cap. Their bolts are installed on the inside of the hole with an o-ring between them on the outside. This way, the only path for air to escape is past the o-ring. Air cannot escape through the bolt holes. The cap bolts onto the top flange, using bolts on the outside of a second o-ring, which creates the seal between the cap and the flange. By putting the bolts on the outside of the o-ring, again the only path for air to escape is past the o-ring. There are no bolt holes for air to escape through because they are on the outside of the ball.
LED work came along nicely in the weeks leading up to the event. This was the first project I've done with this many individually addressable LEDs working together, so I learned a bit. I used a 60A 5V power supply to run all 1200 LEDs. This was generally enough power to run the ball, but if you tried to turn all the LEDs on at full white at max power, you would run into a current draw ceiling. But under regular operating conditions I'd expect, 60A was sufficient. I was using four 300 LED strips, so I tapped power at the beginning and end of each strip, and didn't seem to have any power falloff issues.
I didn't initially realize that I would run out of memory on a typical Arduino, so this was also a learning process. I initially planned to use a Nano, but ended up with with a Teensy. It arrived a day before the event. One of the many points in this project where I very narrowly and luckily avoided disaster.
The Armory has ceiling rigging points that I intended to use, but due to late communications with facilities, we were not allowed to use them. I didn't find this out until three or four days before the event. So it's just a few days before the event and I have a ball that was designed to hang and nowhere to hang it from. The Amory's ceiling and walls are pretty featureless besides those rigging points. I thought about it a lot, and hanging the ball was just absolutely not an option. I tried to consider other installation options, but for a while I had no idea how to install this ball and I thought the entire project was busted. This was the first of many times that I was very prepared to give up.
Then the wild idea came: flip it upside down. This might seem like a simple solution, but I did not consider it for a long time because nothing about the design of the ball was meant to be supported from the bottom. The LED tube was supposed to hang from underneath the cap, which was meant to hang from underneath a rigging point above. Flipping the ball upside down would mean that these two connections which were previously simple cables in tension would have to be replaced with rigid supports in compression. Another complication was that the flange and cap parts were already ordered and printing, so redesigning them was not an option. I had to work with the cap and flange pieces I already had, which had interface holes that were designed only simply for cables to hang from. I had to rapidly design and implement a rigid support solution that would be able interface with these existing support points. The event started Saturday morning, and I had only Thursday and Friday to do much of the originally intended work on the ball and also make these modifications to invert the ball.
On Thursday afternoon, I paid a visit to Bill Brandwein, the stage operations manager at the Clarice, where the school's theatre department is. I used to work there building sets. I can't say enough good things about Bill. I tracked him down in the middle of his work day, started to explain my dilemma, and before I even got the chance to ask him for anything, he offered, "So want to borrow a boom base and pipe?", which is exactly what I wanted to borrow! How did he know! A boom base is a heavy cast iron base that a vertical 2" pipe screws into. Usually, it's used to support up to a few hundred pounds of lighting or sound equipment. It's black, subtle, and minimal, which I thought would be a perfect solution. It was also able to hold the ball 10 feet in the air, so the ball wouldn't obstruct any sightlines along the ground, even if placed in the middle of the show. Bill was incredibly generous and willing to loan out whatever I needed. In addition to the boom base and pipe, I also borrowed a flange to thread onto the top of the pipe to attach to the ball, and two 50lb sandbags to weigh down the base and make sure the ball doesn't fall on anybody.
To connect the flange on the top of the pipe to the inverted cap on the bottom of the ball, I replaced the three eye bolts that were originally meant to hang the ball with six 5" bolts. The bolts extended down to a plywood disk with the pipe flange bolted on the other side. I used 3.5" steel tubes over each bolt as spacers to keep the cap and the plywood disk separated. I cut a hole in the middle of the plywood disk so that the power cable could be run up through the inside of the pipe.
So that took care of replacing the hanging cables with an upside down rigid connection. The other connection I needed to invert and make rigid was the connection between the cap and the LED tube on the inside of the ball. The cap had three horizontal holes on the bottom (now top) of the cap, which were meant for three small shackles to connect to. Below is the cap, printed in a resin on an Objet500 Connex3. The cap came off the printer later that day, Thursday evening. In the image below, it has the USB passthrough and valve threads installed, both sealed in with epoxy.
I made three upside-down U brackets that would fit over and bolt to each of the three shackle tabs. Each bracket also had a top hole to bolt to a plywood disk. This effectively just moved the plane of the cap up and turned it into plywood, a material I could work off of. I made a few holes in this plywood disk to let power and USB freely pass through. From there, I screwed three L brackets to the plywood disk, and connected three scrap shelf brackets vertically to the L brackets. These three uprights fit snugly inside of the LED tube. I used three bolts through the tube to secure it to the uprights.
In the original design, the power supply would have been inside of the cardboard tube. This made the tube several times heavier, so I opted to move it down and bolt it to one of the uprights. This was a much more sturdy means of mounting it, and it made the tube itself pretty featherweight. It also lowered the center of gravity of the entire thing. I was most concerned with the moment when we would tip the assembled ball upside down and thread it on top of the support pipe. When the tube is horizontal, if not well supported, it could be exerting a lot of force on those small shackle tabs on the cap, so I was worried about them breaking off at this point during assembly. Moving the power supply down meant that the tube would exert a little less of a bending moment on the tabs.
Printing was a big complication. I should have gotten the cap and flange parts weeks in advance, but due to bad communications and complications, I did not get the cap until Thursday evening and flange parts until early Friday morning, the day before the event. The flanges printed on a Fortus 400mc, and finished around 4 AM, and as soon as they were complete, I pulled them off the printer and started installing them on the ball. I worked through the night getting the inversion modifications done and installing the flange rings on the ball.
During my initial inflation tests with the cap installed, I found that I was not able to fill the ball to a high enough pressure for the ball to support itself upside down. It would sag off of the base in one direction or another. This was the second moment where I was very close to giving up. I had already spent a full day building the rigid mounts to invert the ball, and after these tests it was looking like the ball just could not be supported from the bottom like I had hoped, and that this whole idea would have only worked if the ball was hanging.
There wasn't much time left, but I spent some time considering. I opted to build a support structure to keep the ball from falling, which I initially really wanted to avoid doing for two reasons: the first was that it would obstruct visibility of the ball, since the ball would mostly be viewed from the bottom. The second reason was that I thought it would be complicated and difficult and I just would not have time. But there wasn't time to hesitate and I couldn't think of any other option, so I went to Home Depot and looked around for ideas. What I came up with addressed both of my concerns. It only took a bit more than an hour to implement, and it was an extremely minimal visible obstruction. It was simply three aluminum U channels bent up to support the ball. These three supports bolted to the bottom plywood disk, which also had the pipe flange and cap bolted to it.
Those initial inflation tests also revealed a second problem. The ball would not hold air. I spent some time trying to find and seal leaks, but it was a hopeless pursuit. This is the third time I wanted to give up. The only solution seemed like a hack to me, but again I saw no other option. We'd have to constantly refill the ball during the event to make up for lost air. So in that trip to Home Depot we also bought an air compressor, air hose, and air hose connection hardware. The Armory is nothing but thousands of square feet of hard surfaces, so running an air compressor during the event might have been a sound problem, so we got the quietest air compressor we could find, which was impressively quiet.
I drilled and tapped a hole in the plastic cap for the air hose connector. I was terrified that I would crack the cap, but it worked. The air hose also ran through the inside of the support pipe alongside the power cable, so the setup was still clean and the air hose didn't add any extra clutter.
At this point, all I had was hope that the air compressor was the last hack that would be needed to make the ball work. We moved everything into the armory at around 11:00 PM Friday night and set it up. Setup went pretty smoothly. The shackle connections in the cap held up, despite at one point accidentally allowing it to support the entire tube horizontally in cantilever.
The ball is filled through a screw-in valve using a shop vac. The valve was original to the beach ball, I just pulled out the threads from the ball and sealed it into the cap. The valve had a rubber gasket that provided a solid seal against the cap.
The air compressor almost worked. It was not fast enough. The ball was leaking at a faster rate than the air compressor was able to replenish. So we'd have to go and buy another, bigger, louder air compressor. But at this point, it was 11:30 and Home Depot was closed. I arranged to have some other event organizers go and pick up an air compressor early the next morning. I wouldn't be able to go myself because I did not sleep the night before and I had very little to eat that day and at that point was barely functional. I was feeling incredibly sick. I threw up on the way home. I wouldn't recommend working this way. Sleep treated me well though, and I was feeling great in the morning.
I got up around 10:00 and headed to the Armory, where there was a brand new air compressor waiting for me. It was loud. Fortunately, we had just enough air hose to put the compressor outside and run the hose and power out through a window. Also it was fortunate that it didn't rain that weekend.
The compressor ran constantly for the next 24 hours at a duty cycle of about 60 seconds on/30 seconds off. It may have been a little intense, but that little guy powered through it.
And it worked. Nothing else went wrong. After nearly every possible design factor failed, the project somehow pulled through, and the organizers loved it.
In retrospect, it's very fortunate that we were not able to hang the ball as initially planned. If it was in the air, there would have been no conceivable way to access it with an air hose. We would have rigged it, inflated it, and within an hour it would have been a very sad glowing sack hanging from the ceiling, and it would be totally inaccessible so it would remain that way for the entire event. Or, if I was smart enough to do inflation tests beforehand, I would have realized it was leaking, but I likely would not have found a solution and it would not have occurred to me to not hang it. Another advantage of it being on the ground was that it was closer to the people and more visible. It would have been easy to miss up on the ceiling, but only 10' off the ground, people were sure to see it and were able to interact with it around the base.
This project was the most intense exercise in high pressure problem solving I've ever undertaken. If there are two things I've learned from this project, they would be that it's critical to be more organized with communication, and if your project plans on containing air, you should probably plan on your plan not working.
When the show was over, I disassembled the ball and laid out the parts. It was the end of an incredibly difficult, but rewarding process. I opted to strip the hardware and valuable components and throw out the rest. It wasn't emotionally easy to do, but there were many reasons that reusing it would not be a good idea. It would be difficult to store for a year until Technica 2017, attaining a boom base and pipe again wouldn't necessarily be guaranteed, the ball was full of holes and leaks, and I don't think I will be able to make it to Technica 2017 to resurrect it. If they want a ball for 2017, it would be better to revise the design with all the insight I've gained and make something that's actually solid and not hacked together 48 hours before the event. Bringing back this flawed and hacked up design back wouldn't really make sense.
So the ball is gone. Intense preparation in anticipation of of a brief but exciting show, then you throw it away. Such is the nature of showbusiness. I'll remember it with a mix of fond and stressed emotions.
