Jonathan Warren

Wires Expert GitHub YouTube Blogger LinkedIn

Balloon Pixels

Summer 2018

The project I'm most proud of. The scope and scale of it, along with the preposterously short timeline (1 month from concept to execution) meant a lot of sleepless nights, but I'm so happy with the result! The ask was to build a wireless outdoor light show that people could both control (i.e. control the light show) and direcly experience walking through. It had to be able to travel around to different festivals and events for the rest of the year.

The Approach

There wasn't enough time available in this project to build anything custom - everything had to be bought off-the-shelf and integrated cleanly together. I quickly determined that wireless DMX (a common command protocol used to drive lights and devices) would be the best way to go, as there was enough consumer hardware available to support what we needed to do. I identified the following components and requirements for the technical hardware:

While initially we had wanted to be able to drive full RGB colouring for each pixel, strict brand colouring requirements, a short timeline, and a tighter project budget quickly ruled that out. Instead, we focused on finding balloons of the right colour and robustness and would light them from a singular bright white LED. SO much work happened around selecting the right colours, size, shape, etc of each pixel, the physical posts to mount them into the ground so they wouldn't tip over, managing winds and rain, physical layout in the field for the animations, and so much more was tackled by the rest of the team, and this couldn't have happened without all their hard work.

We determined over 300 pixels would be required to meet the client's request that from above the light show will be in the shape of Canada. That meant that the first priority was quickly validating components and placing orders from vendors in China. Amazon and AliExpress were our friends. Then, while these things were shipping, we had to have a plan to hit the ground running hard the moment the equipmernt arrived.


Since I was able to buy small quantities of the products we bulk-ordered from China, I was able to move forward with a lighting control strategy. I set up 25 pixels in our conference room, and was relieved to see it working well, not too laggy. However, I knew the difference between a confined office space and a giant field, with tons of electromagnetic noise, could still be a very different story. Unfortunately with this project, we'd have to cross that bridge when we got to it.

The month was a blur. So many different things had to come together. The first "custom" integration I had to consider was the battery packs to drive each pixel. Due to the scale of this project, there was no way we could easily recharge 300+ battery packs every night, and I'm sorry to the planet to say the cheapest and most effective option we had was to use packs of standard AA batteries.

Like I mentioned earlier, all the physical components of this project were huge undertakings of their own. The client had wanted large "balloons" for each pixel, however we quickly found that standard balloons really liked to pop after being out in the sun for an hour or two. About 10% would pop per hour, and at a scale of 300+ that was just not an option. We settled on what were effectively small exercise balls. My next task was illuminating them! The original plan was just to mount a light source underneath the balloon to shine up into it. However, the thick and slightly reflective of these "balloons" meant that the lights were super dim. So, I quickly spun up a small PCB design that could integrate into the air plug on the balloon. The power cable to the LED would also act as a tether to hold the balloon down on its post


We effectively had no way to know for sure if this was going to work on the first day it was due to be deployed. We spent 10 hours inflating these balloons using 5 air compressors on a generator in a field, and as darkness fell my stomach was in knots. The client was there throughout, and to my surprise and initial deep concern, one of them pulled out a drone to film the VERY FIRST ATTEMPT I was making to drive the light show. No pressure or anything.

I very nearly cried with happiness. There were a few pixels addressed incorrectly, but holy moly it looked great!! Naturally there were still number of challenges we had to work through: wireless interference requiring me to hack a signal booster antenna onto the original DMX transmitter, and laggy code requiring a quick swap from a Raspbery Pi to a full desktop computer. But by the end of the first full event, this setup was ready to travel across Canada and provide a fun experience for many.

Setup     DroneShot