Build Your Own Undersea Robot

With a remotely operated underwater vehicle, you can explore the depths without getting wet

6 min read

Last year at about this time, crews in the Gulf of Mexico were working feverishly to bring BP's blown-out oil well under control. Some of the more spectacular parts of that effort, as you may recall, involved the use of remotely operated vehicles, or ROVs. Perhaps you had the same thought as I did—that it would be cool to build one.

To be sure, no garage-workbench hacker is going to build an undersea robot that operates a diamond saw or wrestles with a stuck blowout preventer. But those vehicles also monitored events on the seafloor and streamed some amazing video to the Web in real time. A small inspection-class unit—one that carries just a video camera around underwater—ought to be within the grasp of an avid DIYer.

A quick search of the Web revealed no shortage of home-built ROVs. There are even competitions, such as the one for students that the Marine Advanced Technology Education Center in Monterey, Calif., has been running for 10 years. In a similar vein, MIT's Sea Perch program trains teachers (who in turn train their students) to build a simple ROV as an educational exercise.

With all that going on, I thought it was high time for me to get my feet wet—again. As it happens, I've some background in this sort of thing, having worked briefly building underwater instruments at Columbia University's Lamont-Doherty Earth Observatory, just north of New York City.

I even built an ROV for fun in the late 1990s. Its underwater thrusters, like the ones employed by most DIYers today, used DC motors mounted in watertight housings. Flexible shaft seals prevented water from getting to the innards of the motors. It used trolling motors, the kind you see pushing small fishing boats around. Submersible bilge pumps are another popular solution.

The great thing about bilge-pump motors is that they are dirt cheap—perfect if all you want is something that can swim around at shallow depths. At greater depths, though, the pressure will cause the flexible seals to close down around the spinning shaft, sapping power and heating the seal. Ultimately, the seal fails and the motor floods.

I wanted my next-generation ROV to be able to go deep, so I took a different approach this time, which was to fill the motors with oil. Shaft seals are still required to keep the oil on the inside separated from the water on the outside. But the two sides of the seal are always at the same pressure, so the motors should be able to operate at any depth.

The wrinkle here is that brushed DC motors don't like to run in oil—the brushes foul. So I used brushless DC motors, which are electronically commutated and will spin as happily in oil as they do in air. These days such brushless motors and their controllers can be had for very little, because they are widely used in radio-controlled model cars and airplanes.

I wanted my ROV to be as small as possible—my first ROV was so big and heavy it was hard to lift—so the new model would need only small motors. I selected the HXT 2835 (US $20 each from Their controllers (HK-30A, also from are also surprisingly inexpensive, only $15. In fact, the tiny shaft seals (from McMaster-Carr) cost nearly as much as the controllers.

To hold each motor and to shield the attached prop, I used a 90-millimeter-diameter plastic shroud designed for an electric-ducted fan, the sort that powers some model airplanes ($2 from BP Hobbies). They aren't shaped like real Kort nozzles—the kind you might see on the U.S. Navy's Alvin or other submersibles—but they look the part. I cut down the propellers from some model-airplane props I had on hand. A small aluminum disk placed on the business end of the motor holds the shaft seal in place. Some vinyl tubing, a few stainless-steel hose clamps, and a plastic hose fitting from the local home center completed the unit, bringing the cost of each motor to about $70, which is astonishingly little for an underwater thruster that could, in principle, operate at the bottom of the Mariana Trench.

The rest of my ROV can't go anywhere near that deep, of course. But I expect it would have no trouble diving down as much as 60 meters or so. That's because I made its two hefty pressure housings from 10-centimeter-diameter polyvinyl chloride end caps ($17 each from a local plumbing supplier) intended for 3-inch schedule-80 pipe.

The rear pressure housing contains the three motor controllers, which just barely fit. The front housing was also stuffed rather full with a Panasonic GP-CX161 video camera (only $24 from, unfortunately now discontinued), a radio-control servo (a $3 no-name model) to tilt the camera up and down, a 0- to 100-psi pressure sensor to gauge depth (which I scored for just $28 on eBay), an HMC6352 digital compass ($35 from Sparkfun Electronics) to determine heading, and an Arduino Pro Mini microcontroller ($19, also from Sparkfun) to serve as the ROV's brain. Were I to do this again, I'd probably use 4- or 5-inch PVC end caps to make fitting everything in easier.

The tether connecting the ROV with the surface is split into two parts: a power cable and a data cable, both bright yellow because it looks so spiffy in the water. I strapped on short lengths of foam pipe insulation attached at intervals; a neutrally buoyant tether would have been great, but pricey.

The cheapest power cable I could find to send 12 volts DC down to the ROV was a 30-meter outdoor extension cord ($39 on eBay). To reduce resistance, I wired two of its three 14-gauge conductors together. The 30-meter CAT5e data cable (another eBay bargain, $10) contains four twisted pairs of conductors. One pair carries a salvaged RC transmitter's pulse-position-modulated control signals from the surface down to the ROV. Another pair carries heading and depth values up to the surface at 9600 baud, where they are read by a second Arduino microcontroller, which displays the data on a serial-enabled LCD screen (Sparkfun, $25). The control signals and data telemetry pass through Maxim 490 line driver/receiver chips at both ends, which will allow the ROV to operate with a much longer data cable if desired. A third twisted pair conveys the video signal up to the surface, with video baluns ($20 each from at each end to make that work properly.

Although I tried to design this little ROV to be easy to build, making the end caps for the two pressure housings watertight took some machining on a lathe. For the rear housing, I used aluminum, which acts as a heat sink for the three motor controllers mounted behind it. The front pressure housing includes a ¼-inch-thick (6-mm) acrylic dome ($30 from EZ Tops World Wide) that allows the camera to tilt up and down without introducing optical distortion. The dome also looks slick. But I had to machine an acrylic disk and attach it to the back of the dome so that it could be fitted with an O-ring (about $1 from McMaster-Carr). The machined aluminum cap on the rear housing is similarly outfitted with an O-ring for a high-pressure seal.

One particular challenge is making the many electrical connections that pass through the two PVC housings—you really don't want those to leak under pressure. Were money no object, I'd buy fittings designed for this purpose from Sea Con or a similar company. Money was an object, however, so I made my own bulkhead connectors, using plastic pipe fittings, copper wire, and epoxy. I tested one of these creations under pressure using a short length of water-filled steel pipe and a bicycle pump, and it seemed to work just fine. I covered these and all other wiring splices exposed to water with a generous amount of 3M's Scotchkote Electrical Coating—an important trick I learned at Lamont-Doherty.

VIDEO: Taking the ROV out for a couple test-dives.

I tested my ROV at a flooded quarry that local scuba divers frequent, where the water was crystal clear. All systems worked pretty much as intended. The little sub went up and down briskly when I operated the vertical thruster, and it proved very responsive in the horizontal plane, too.

At some point, I'd like to take this little ROV out for some wreck diving in the ocean off my home state of North Carolina, a goal that drove many of my design decisions. Some fascinating sites, including the remains of U-352, a German submarine that was depth-charged in 1942, are located not far away. Sure, you can visit that wreck, which lies in about 30 meters of water, by donning conventional scuba gear. But that requires no motors, no batteries, no cables, no video camera, no electronics, no microcontrollers, and no programming. Where's the fun in that?

This article originally appeared in print as "Build Your Own Robosub".

To Probe Further

If you're interested in building your own ROV, download the source code and schematics [.zip].

The Conversation (0)

From WinZips to Cat GIFs, Jacob Ziv’s Algorithms Have Powered Decades of Compression

The lossless-compression pioneer received the 2021 IEEE Medal of Honor

11 min read
Photo: Rami Shlush

Lossless data compression seems a bit like a magic trick. Its cousin, lossy compression, is easier to comprehend. Lossy algorithms are used to get music into the popular MP3 format and turn a digital image into a standard JPEG file. They do this by selectively removing bits, taking what scientists know about the way we see and hear to determine which bits we'd least miss. But no one can make the case that the resulting file is a perfect replica of the original.

Not so with lossless data compression. Bits do disappear, making the data file dramatically smaller and thus easier to store and transmit. The important difference is that the bits reappear on command. It's as if the bits are rabbits in a magician's act, disappearing and then reappearing from inside a hat at the wave of a wand.

Keep Reading ↓ Show less