On 2 April 2011, MATE held the Monterey Bay Regional underwater ROV competition, one of 20 regional competitions they run (and the oldest of them!). There were about a dozen “RANGER” teams (all high-school teams, I believe) and 30 or 40 “SCOUT” teams (mostly middle-school groups). The team I was coaching was one of the smallest (3 boys, only 2 of whom could make it to the contest). This was their first time doing the contest, and my first time coaching, and they started decided very late on entering the contest. As is usual for engineers, they underestimated the amount of work that would be involved in building and debugging the design.
As a coach, I tried very hard to give them good advice but let them make all their own decisions, even when I thought they were making a mistake. This is surprisingly hard to do—harder even than essentially the same task supervising graduate research students, as 15-year-olds are much more likely to pick up a brainstorming suggestion that is intended as a strawman argument and take it as a command. For many of the building sessions I had to let the kids work alone, just checking on them every 15-to-20 minutes, to avoid kibbitzing. After each coaching session I did check privately with my son to see if I had been providing the right level of support for them to keep on track but make their own decisions.
They managed to get their underwater ROV fully assembled and in the water just the day before the contest. I was not able to be present for this launching, but I understand that the original foam flotation was not sufficient, so they MacGyvered on a couple of empty soda cans, using electrical tape to seal the holes and fasten the cans to the frame. We had still not come up with a solution to the problem of providing a pool-side TV, so they were not able to test the cameras.
The underwater ROV ashore on the day of the contest. Note the last-minute addition of soda cans for extra floatation and the very lightweight tether. There are 4 motors (two for front and back and yaw-steering, 2 for up and down and pitch). The right motor and propeller is clearly visible, but the other 3 are mostly obscured by the framework in this picture. The propellers were a good find: they cost less than $3 each from Amazon and mount directly on the motor shafts with just a setscrew (no adapter needed). They are made by Traxxas and are intended as replacement propellers for radio-controlled toy boats (their Villain EX model).
Because the team was so small, the drive down to Monterey was done in one minivan. One of the team members did not show up at the pickup point, and we had no way to contact him, so we had to leave after 20 minutes without him. We went past our exit on Highway 1 (my fault—I should have been navigating, but I was in the back seat and not paying enough attention to the signs) and had to turn around at the next exit.
The team did have to do some repairs, as some of the tether wires had come loose from the terminal block in the controller, and there was a cold-soldered joint that had shaken loose on one switch. They fixed these problems and got their machine set up for their first run. We were all pleased to see that it floated.
ROV floating in the pool. All 4 motors are visible in this view, as are the two camera mounts for the navigation and tool cameras.
This was the first time the team had access to a poolside tv screen, and the monitor that had been provided at their station did not work well. They did not know whether this was a problem with the monitor or the cameras, but eventually a new monitor was brought over, and the problem was determined to be the monitor provided by the contest.
The team switching monitors. The black-framed one worked much better than older one with the cream-colored frame. I have to find out what those black monitors cost and where we can get them—they look much more feasible for taking to pool practices than junk TVs from thrift stores, which is what we were planning to get (but had not yet gotten around to acquiring).
After switching monitors, the boys did get some flying time using the cameras. Most of the teams had spent weeks learning to pilot their rigs and tweaking propulsion and tool systems, but this team was flying for the first time at the contest itself. Their design was simple and robust enough that they were able to maneuver down to the props, but they did not have fine enough control to actually perform the missions with the tools they had built. Still we were all stoked that they could see through the cameras and do some steering, even if not to the fine precision they had dreamed of.
ROV flying down to the props. Propulsion worked well, but steering was not as fine as the designers had hoped, and the camera view was somewhat confusing.
They got another half hour run later in the day, but it was pretty much a repeat of the first run, though the monitor worked this time for the whole run. One problem the team had was that they could not tell which set of props they were looking at and wasted a lot of time trying to reach the next station over (a large identifying number on the underwater stations would have helped). A down-pointing camera so that they could see the bottom of the pool might have helped with navigation.
The boys were happy that everything worked as well as it did, given that they had not been able to test ahead of time. They did not have any illusions of winning (they knew that they were going up against teams that had been working on their vehicles for 4 or 5 years), but they were hoping that they could get a little further on the mission tasks.
The boys gave a fine engineering presentation of the features of their design, but I suspect that they lost some points for safety, as their propellers were unshrouded, and all the other designs had some sort of mesh box or ducting around the propellers to protect fingers. A very popular design (because of the ease of construction) was to put a metal mesh gutter guard around the motor and propeller fastened on with a large hose clamp. This is something I should have coached the boys about, as safety was one of my main concerns, but I had been content with just admonishing them not to get near the motors while the power was connected. The props were all inside the frame and they had been careful to wire the motors so that they couldn’t foul the props with the motor wires. They had also been very good about disconnecting the power when doing any work on the vehicle. Next year, they’ll have the passive safety shields around the props.
Yes, it was fun enough for them that they want to do it again next year. They were even talking about what to do on Club Day next fall to recruit a bigger team. They are also trying to figure out how to get adequate pool access, as it was clear to them that one of their biggest constraints this year was the very limited practice and debugging time.
They are not proposing any major changes to the mechanical design (adding shields around the props, maybe moving the left and right motors further apart), but they have some major ambitions for the control system. They want to move to electronic controls in the vehicle, so that they can use a longer lightweight tether, but not have the large resistance drops that come from having many small wires and communicating through power switching. They want the tether to have just one pair of power conductors, the camera wires, and a few wires for a serial connection. They got some ideas for how to make connections to a pressure-proof box (from a team that had worked on that, but not managed to finish their vehicle) and are going to spend this spring learning how to program microcontrollers to control DC motors using PWM and H-bridges.
Already in half a day, my son has a working Python program for doing serial communication with an Arduino microcontroller over a USB link from a Macintosh. I did have to help debug, as it took us a while to realize that opening the serial connection to the Arduino did a reset and started the program on the Arduino, so we had to add a handshaking signal to the Arduino code that the Python program waited for, to make sure that the Arduino did not lose the beginning of the communication.
They’ll build a motor shield for the Arduino for controlling some small motors (since I already have a kit), but we’ll have to look for some higher-power H-bridges, since the chip used on the motor shields can only deliver 600 milliamps per motor, and they’ll need over 4 times that. There do seem to be some reasonably priced 3 Amp H-bridges around using the LMD18200 chip, but I think it would be good for them to practice with low-power designs first. The 3 Amp H-bridges will require some big heat sinks, and I’m not sure how they’re going to keep four of them cool enough in a watertight box. I think that will have to wait until the summer.
- High school Robotics Club started (gasstationwithoutpumps.wordpress.com)
- Waterproofing cameras for underwater ROVs (gasstationwithoutpumps.wordpress.com)
- National Robotics Week April 9 – 17, 2011 (gasstationwithoutpumps.wordpress.com)