I’m going through a rough period right now. There’s an illness in my family and I’m having a hard time focusing on anything but worrying and trying to take care of health stuff. Everyone is going to be okay, but it’s going to be a difficult four or five months, and I really appreciate your patience and understanding. I’m going to keep putting up comics, but I don’t how much else I’ll be able to work on.

To anyone I’ve been corresponding with, I’m sorry that I may be even more tardy than usual. While davean (the xkcd sysadmin/business manager) monitors the press@xkcd.com address, I know he only forwards to me a fraction of the huge flood of mail that goes there. If you’re trying to reach me personally about something, you can write to me directly at xkcd@xkcd.com, but I’m afraid I won’t able to reply to most of it right now.

I know there haven’t been any posts here in a while. Since most of my projects are on hold right now, I thought I’d share some pictures from one that’s almost done: an underwater ROV. Exploring lakes and oceans has always fascinated me, and while I’ve spent a lot of time snorkeling and free diving, in the end I’m more interested in sending robots than going myself.

I tried to build a couple of ROVs in high school out of scavenged R/C cars and spare parts, but none of them ever worked very well. Last summer, I got interested again and picked up an Inventivity ROV-in-a-Box:

It’s a very basic kit designed to use off-the-shelf parts as much as possible, to encourage people to play with the design or expand on it. I’ve gotten a lot of help and some cool ideas from the company founder, Dr. Karen Suhm, who coaches robotics teams in ROV-building competitions and generally knows everything about ROVs. The kit comes with a good set of underwater motors and a sensitive camera, and this summer I started modifying it to use an Arduino and joystick control, running the whole thing over Cat-5 cable (which significantly lightened the tether). This will also let me add other equipment, like a still camera, depth gauge, compass, and sonar.

It’s very close to being finished—I just have a couple wires to reroute and a leak to seal—but for now, here are some pictures from construction and testing:

I made a coupler so the tether could be detached, and added a chamber to hold the Arduino, Ethernet shield, and motor control board. A Python script on the surface translates joystick values into motor speeds, and the Arduino has some code to listen to commands via the Ethernet and control the motors using three TLE-5206 H-bridges. The 5206s offer more protection than some other H-bridges—I initially used some smaller chips, and managed to blow out a couple. (Thank you to mpanetta of #sparkfun for hooking me up with the 5206s.)

A note to anyone who wants to build something like this: the Arduino isn’t actually capable of processing video, so you’ll need to either put an Ethernet camera and hub on the rover, or—if your camera isn’t digital—do what I did and divert two of the Cat-5’s twisted pairs to carry RCA video, running the Ethernet solely on the other half.

My friend Mike loaned his canoe for depth testing in Walden Pond, which is (according to data from the 1940s) the deepest lake in Massachusetts

At the bottom of Walden, there are close to three extra atmospheres of pressure.

The zip ties double as binary depth markers. This one is 14 meters.

This is the vacuum pump for sealing up wires passing into the sub (it’s sitting atop a draft of the online communities map). If you open up the exterior/water side of a cable and submerse it in a pool of marine epoxy, then apply suction to the dry interior of the sub, it sucks the epoxy through the cable, plugging it up completely. You can also use it to suck all the air out of a wine bottle with random objects inside. It’s fun to see how different materials react to a near-vacuum—particularly if you’ve just drunk a bottle of wine. I didn’t get much more done that day.

Lastly, here’s a clip of the bottom of Walden Pond, about 80 feet below the surface.  This was an unpowered pressure test—the sub was just dangling on a rope—so it’s not very exciting, but it was the only test where I could record the video feed:

The Walden lakebed is pretty dead—the material you’re seeing is flakes of debris stirred up by the sub. In other lakes, we’ve found cooler stuff.  In Seymour Pond on Cape Cod, we had huge catfish fish swim up to the camera and look at it, and we explored a sunken fishing boat on the bottom of Sheep Pond.  I’ve also learned that deck chairs apparently fall off docks all the time—the lakebed 20 feet below the dock on one lake was absolutely littered with them.  When I get a chance to send it to some more interesting places, I’ll be sure to share footage.

P.S. A belated thank-you to the NYC Makerbotters; after I posted comic #743, they fabricated and mailed to me an actual tiny open-source violin.