Stories and Failures
Unsurprisingly, in the quest to build flying machines (when we didn't know what we were doing at first), failure was always an option. Here are some stories from the quest to get to the Flyt 16.
When I first started this project, it didn't have a name. That's why the first several prototypes were simply called the Mark X. Let's face it - I'm not creative this way. One day while driving with several friends in a car someone suggested that I call my prototype the 'AnselCopter' (my name is Ansel, so you can see where this name came from). Now this is clearly a terrible name, mostly because the majority of people in the world don't know me and so this name means nothing to them. Also, I've never wanted to name things after myself because it seems egotistical and narcissistic.
My friends, however, thought this was a great (aka hilarious) name and to this day they ask me how the AnselCopter is going no matter how many times I ask them not to. Sigh.
Let it snow (packing peanuts)
Another memorable event was during a test of the Mark III when I was throttling it up and a large garbage bag suddenly got sucked into a propeller from above. Somehow in the attic storage space in the garage there was a large bag full of packing peanuts that came loose in the wind and was obliterated. It looked like a sudden snowstorm and took over 2 hours to clean up the mess.
The uncontrollable toilet bowl
The term 'toilet bowl' here is not referring to anything porcelain, but rather the action of a vehicle spinning as though it were being flushed. This was the second test of the Mark IV and the PID values were waaaaaaay off, leading it to bounce off the ground and in a frantic attempt to stabilize itself it went crazy. Arguably this was the most terrifying test we ever conducted. Luckily, given our focus on safety we had it tied down to weights and limited its max power output to prevent it from flying off uncontrollably.
During testing of the Mark II, I would test each side independently because there was only enough space in the garage to really test one side at a time (#startuplife). We had dramatically over-estimated the propeller pitch needed (ok - in all honesty I didn't actually understand how a propeller worked back then and was mostly guessing), and so each test we would slowly bring the engine to full throttle and it would get to a higher RPM, but only by a couple hundred each time. The goal was to get the engine to full power at 6100 RPM to deliver maximum thrust.
Each time I knew there was a chance of liftoff on the active side (so we had safety precautions in place so it could only lift ~2" off the ground), but after running something like 50 tests I didn't really expect it to. Eventually it did lift off rather unexpectedly and in the video you can see the surprise in my eyes:
The exploding duct
No, not an exploding duck (that would be horrible), but an exploding duct (the big round thing that went around the propeller). If built right, a duct can improve the efficiency of a propeller (meaning more thrust for the same amount of power). However, it's very tricky to shape it correctly, especially when you're a guy trying to figure this all out by yourself in a garage. Let me also add that building ducts was the bane of my existence. It involved either cutting Styrofoam or using an expanding foam in a 3D printed mold (both labor-intensive and messy processes), and then gluing all the segments together, followed by fiberglass or carbon fiber cloth and epoxy coats, and finally letting it dry for 24+ hours. In all, it took anywhere between 1.5-3 weeks to build a set of ducts. Yuck.
So, in a (misguided) attempt at making a better duct I thought I could 3D print the segments and be able to try better shapes (and lighter weight). I bought two 3D printers and had them literally printing around the clock (including waking up in the middle of the night to remove a finished print and start the next one). After a week and a half of printing I then glued the segments together into a finished duct. I then mounted it and ran a test to see if it would increase the amount of thrust generated. Well, it worked great until the Mark II hit full speed and two of the glue joints failed (on the longest unsupported segment). You can see what happens in the video...
Battery cage fail
The Mark III had ~100lbs of batteries carefully housed in the bottom. On one of the first full tests I discovered that there was insufficient cross-bracing on the battery holders to support the battery weight. They simply collapsed on themselves and I'm incredibly lucky that none of the batteries were crushed/punctured, which could have led them to either catch fire or explode.
Goodbye bath towel
During the testing of the Mark II occasionally items would be sucked through a propeller (like a loose plastic bag), so when it came time to test both engines at the same time we cleared out as much as we could from the garage and used tape to prevent the rest of the stuff from flying off shelves. The first time the whole vehicle got off the ground it put out so much wind it managed to suck a full-sized bath towel into a propeller and shred it to tiny pieces that flew everywhere:
The belt shredder
The initial prototype, the Mark I, had problems at higher speeds with belts. It used belts to transmit the power from the engine to the propellers, which isn't done very often in aerospace applications and I soon discovered why. Specifically, it would destroy belts and shred them to pieces once the power got to ~50% or higher. At ~$50 per belt, I wasn't too happy about this. It turns out I had installed the engine motor mounts incorrectly so they weren't actually dampening the engine vibrations among other design flaws. You can see it happen below:
The only major injury we ever had was on the Mark I, where I was being particularly stupid. I used to pull on the upper belt connecting the propellers to the gear box by hand to get the propellers to spin and test the friction in the system (ok, I also wanted to show off how well the propellers would spin to people). This worked great, until I installed a tension pulley so the belt couldn't slip, and in trying to make it go fast I didn't release my hand fast enough before it got to a bushing. Now, the bushing is where the timing belt slots into a series of grooves that helps it transmit the power, and with the tension pulley installed the belt no longer had flexibility. The end result was two of my fingers getting crushed between the belt and bushing, which was the equivalent of putting your fingers through some gears.
I was literally screaming in pain for a good 45 minutes while running my fingers under water to wash the blood away. Amazing, there was no permanent damage and all that happened was I lost two fingernails. One of the most painful events of my life, but luckily the only time someone was seriously hurt during construction and testing of these prototypes. I figured it would be better not to include pictures for this story.
See the safety section for out commitment to safety.
When the Mark III was finally complete and ready for flight testing we ran into a surprising problem. It turns out the steel frame seriously messed with the compass and frequently the remote control system would refuse to arm. We ended up physically lowering the compass to get it further away from the metal and constantly re-calibrating it, but it was incredibly annoying to deal with.