Eventually the “wobbly” enclosures made their way back to my workbench, and upon opening them I discovered that they had all leaked. But determined to improve the design, I scraped off the surface rust and began the post-mortem. Pressure from the latch clamps had actually separated the o-ring seats from the main body, damaging the o-rings. But it looked like I might have damaged the adhesion well before that, with all the grinding I had done in my quest for more buoyancy.
So the first step was to make the end caps bigger, and move the upper latch clamp ring further away from the delicate o-ring seat. Then I lengthened the support struts along the bottom shell, carefully gluing them into place with clamps. Once the latch clamps were attached, I thought about how I was actually going to place the electronics inside housing. Using Lego blocks as a proxy for some of the parts, I experimented with many different configurations attached to knock-out caps which kept everything level inside the housing. In the process I made the happy discovery that you can solvent weld the ABS bricks together, and they will also bond reasonably well to the caps, provided you use “just enough” solvent to bond the surfaces without softening them. As they say: “Many Bothans died to bring us this information…”
In the end I decided to put the power block into the bottom of the unit, and use a second knockout cap as the platform holding the electronics in the top half of the unit, I would figure out how to connect them later via some kind of power plug.
I was really happy with the new housing, as the clips applied a nice even compression to the o-ring, and the overall unit just “felt right” to my divers hands. With the batteries held securely in place by a scaffold of Lego, the bucket buoyancy tests looked good, showing no torque from the uneven weight distribution that plagued the first builds. But it was riding pretty high in the water, and I need them to be “just barely positive” if they were going to respond well in low flow conditions. So I added a few ballast washers to both the top and bottom clam shells, and the Mark II housing was finally complete.
It was now somewhere past 2 AM, and my wife, who had been on some sort of Skype call to another time zone, came down to the workshop to suggest that we call it a day. But I was in pretty good spirits at at this point, and like a proud father, I started showing off my new baby as it bobbed up and down in the laundry tub. She was trying to smile, but the air was still pretty thick with the PVC/ABS solvent I had been using. And she couldn’t help but notice the small piles of half melted Lego scattered around the workbench. While I was babbling, she slid a phone out of her back pocket, and just before capturing the above photo, she quips “You know, other women loose their men to football, or video games, but I end up with one who hides down in the basement, playing with Lego!”
Addendum 2014-06-01: Just a note for any other underwater DIY’ers out there: my
housing design has changed significantly since this build. I have now moved away from metal latch clamps to a system of nylon bolts around the perimeter. See photos here.