I’ve had the pleasure of collaborating with Gab707 for the Ichabod and Ichabod Jr., and we’ve also worked together on some shoots. I love working with Gab ‘cause he’s crazy smart AND an amazing pilot, and he pushes my designs further than I would on my own. So when he told me about a drone project for whale research I was totally excited to get involved.

Researchers wanted to gather information on Right Whales by collecting the mucous in their exhalations, to save them from extinction. Gab sent me this video from another team:

So it’d all been done already, but the drone in the video was unavailable, so the researchers needed their own. We couldn’t get too much info on the original drone, but we came up with our own wish list of specs:

It should be waterproof, because it’ll be flying through mist.

It should float. We don’t know if the original floated, the white parts under the arms might be styrofoam, so maybe it does. Regardless, we wanted ours to float so it could be saved in case of a crash.

The original looks like it had 9 or 10” props. We were told that they might want to launch and land from a small lifeboat, so a 9 or 10” quad sounded really scary. Gab had just gotten a Racer4 from DRL, so we had a good idea what a 5s 7” power train could lift. He decided that this should be a 7” quad so it could be as small and safe as possible.

And of course it needed the petri dish servo mech, and a Hero mount to record the action. We couldn’t think of a better way to hold down the petri dish, so we opted for same suction cup system in the video.

After some initial chats and sketches, the project went on hold as the client awaited funding approval through Canadian government agencies. Months went by. I secretly hoped it wouldn’t happen since I’d tried to make a waterproof floating quad once before and failed MISERABLY.

When the funding finally came through we were left with just 3 weeks or so. Given the short time frame we decided that Gab would order the carbon directly from Nick at CNC Madness, while I produced the printed parts here. I would never see the carbon or get to test fit the parts. Gab shipped me waterproof servos and suction cups for the petri dish mech. I had to work off drawings of the petri dish, since they had a minimum order of 1,000. Luckily the drawings were extremely detailed.

We knew heat dissipation would be an issue with an enclosed build, but we punted that problem down the line, hoping real life testing would reveal that it wasn’t as serious as we feared.

Given my previous failures with flotation, we decided that we’d use pool noodles on the arms to float the quad. Pool noodles are amazing. They’re cheap and super floaty, and if they’re not floaty enough you can just add more. I added an optional crossbar between the arms in case we needed more noodle.

Gab asked for a pure X layout, so that was simple enough. After laying out the frame, my next idea was to make a 5 sided TPU box to simplify waterproofing. I could make a lid with a channel to slip over the box, and integrate the servo mech and GoPro mount into the lid. There was no time for testing and iterations so I made it pretty robust.

The real challenge was the top. It had to hold the servo, the petri dish, and the GoPro. I lobbied for a pushrod mech (because it’s easy) for the petri dish, but Gab veto’d it because pushrods have varying torque, and if you have a big ol’ flap in the wind (i.e., the petri dish lid), you might want maximum torque at all times. Unable to argue against logic, I got working on a direct drive mech.

The splines on the servo are way small, so I took a pic of the servo and used that to draw an interface for the servo. I didn’t think fdm could resolve detail at that level, so I printed a little square with that spline pattern with my Form2. I was very pleasantly surprised when it slipped perfectly over the servo (chalk one up for dumb luck).

SLA is slow and expensive, so I left that part as is and drew a TPU part that wraps around it. That in turn screws into a carbon piece that has holes for four suction cups, which hold the petri dish lid. One side has a pivot, the other side locks right into the servo. I added tabs for the servo and the Hero, as well as guides to align the petri dish, and got this:

There’s a channel on the bottom to increase the contact surface area between the two pieces, to help with waterproofing.

The Form2 is incredibly precise, which allowed me to design superfine tiny details like this suction cup interface. Shove the little rubber piece down through that hole and it’s solidly mounted, with little or no play.

With the details worked out it was time to print the whole piece. I got durable resin from Formlabs just for this. The center section was sized so that the lid would just fit in the build volume at an angle. It was a big ol’ print.

Gab’s getting the lid holder cut in carbon himself, so I printed a stand-in just to see the whole thing in action. And yes, it was very satisfying.

With that, my job was done. I made enough parts for two builds and shipped them off to Gab. Heat did turn out to be an issue, and it was sorta solved by putting a conformal coated vtx on the outside, and laying the esc on a piece of aluminum foil. The foil didn’t transfer the heat outside of the pod, but it still did enough.

Gab enlisted Liam Olders to go on the expedition and fly the quad (it’s a multi-week commitment and Gab couldn’t go). One quad did failsafe and crash, but it was retrieved (yay noodles!). The other performed perfectly. I was beyond excited to see video of it in action.

A week later a Ted Talk was published, about the very expedition that was our original inspiration! It’s funny to learn all this AFTER starting the project, but it really added more depth to my understanding of this undertaking. It’s really worth watching.

And I finally got to see the entire quad!

Round Two

Even though the first expedition was a success, we still had to perfect the drones. Issues we wanted addressed were: weight, waterproofing, heat dissipation, and aesthetics.

The original TPU pod was more robust than necessary, so I thinned down its walls. The FC stack could also be shaken loose on hard landings. Gab had used so much glue to waterproof it he decided that a little more wouldn’t hurt, so we eliminated the bottom of the pod. That allows us to attach the FC stack to carbon rather than TPU. Gab also requested that all the wiring enter through the carbon on the bottom using one or two holes rather than individually through the TPU on the sides.

As for heat dissipation, we clearly needed a proper heat sink. Gab wanted something the FC stack could straddle, with the esc fets in direct contact. I suggested a shelf above it for the vtx. Fins could pierce the TPU pod to be cooled outside.

I got in touch with Mike Hornfeck (aka doingthedoings) to CNC the aluminum. Once he determined the bit sizes he needed for the cut, I redesigned the piece. It’s important to get those internal radii just right, since the TPU skirt has to match it for a good seal.

Mike redesigned the shelf so that it slots in more securely (he can turn and re-align the part to make cuts in different directions, so he’s more suited to make these changes). The result was really gorgeous. And yes, sliding the fins through the TPU skirt was…VERY SATISFYING.

We also opted to go with the DJI fpv system. The air unit is in a nice metal case that can be attached to the shelf, so that should take care of cooling for video and control.

As far as aesthetics, the noodles and zipties had to go. We figured floats could easily be hotwired foam that slides over the crossbars. I’d drawn the crossbars as totally straight pieces with no protrusions just for this purpose. Gab went ahead and ordered that from some friends.

1/21/20

It’s finally finished! Gab got floats made. They’re hotwire cut and slip over the carbon crossbar. He also made them a slick carrying case. Here’s some nice shots he sent along.

10/5/20 Update

Liam went on another expedition, here’s the video: