The View Up Here

Random scribblings about kites, photography, machining, and anything else

  • Flickr Gallery

  • Advertisements

Underwater ROV Competition – More Ideas

Posted by Tom Benedict on 13/07/2010

Several years ago in another blog, I wrote down some thoughts on the underwater ROV competitions that are put on in middle schools, high schools, and colleges around the world each year.  Here in the US, the biggest one is the MATE challenge.  The regional competition here on the Big Island is the BIRR, or Big Island ROV Regional competition.  More information on the competition itself can be found on the MATE Website.

Technology has come a long way since I wrote down my thoughts on that other blog.  I’ll summarize them here, and then share some new thoughts I’ve had since:

  • First and foremost, have a good time.  It’s sad to see teams, especially mentors, get so wrapped up in the competition side of the ROV competition that they loose sight of the idea that their team is building an underwater exploration machine!  That’s cool!  That’s fun!
  • If you build specifically to the rules of the challenge, you leave yourself no wiggle-room to deal with the unexpected.  Remember that a 10″ tall prop may be 9.75″ high, and still be considered 10″ for all intents and purposes.  Stay flexible.
  • The operator will become disoriented.  I have yet to see an ROV team compete where this didn’t happen at some point in time.  The answer is to instrument your ROV so that the operator can get oriented again afterward.  More on this later.
  • Trust each other and delegate duties.  I saw this bite a FIRST Robotics team in the butt.  The team had two self-declared “stars” who wouldn’t trust each other.  In the end the two of them sabotaged their team out of sheer pride.  This happens with ROV competitions, too.  When you see the flip-side, where the members of a team trust each other implicitly, it’s a thing of beauty to see.  Want to compete at the top of the pack?  Build a team, not a star.
  • Test test test test test test test!  There’s nothing worse than seeing a team lower their ROV into the water, and then watching some major system, like one of their thrusters, fail.  No wait!  There is something worse: lowering your own ROV into the water, and finding out you didn’t test it first, and seeing some major system fail.  Test test test test test!
  • Train train train train train train train!  If I’m betting on a car race, I don’t put my money on the better car, I put it on the better driver.  During the heyday of Battlebots, someone made the observation that it wasn’t the weapons or the defenses that determined a win or a loss, it was how much time each operator had spent driving their robot.  The same holds true for ROV competition.  If a team has a $10,000 robot, but the operator only has thirty minutes of seat time on it, the team with the $250 robot with thirty hours of time at the controls will beat their pants off.  Beg, borrow, or steal as much pool time as you can get.  And when you’ve used that all up, toss it in a nearby lake and keep going!
  • Stay on task.  This really shouldn’t need saying, but it happened to all but one team the year I judged.  Kids procrastinated, mentors didn’t stay focused, teams lagged, and were still building their ROV the morning of competition.  It works on homework, and sort of works on tests.  It doesn’t work when undertaking an engineering project.  Stay on task!
  • Do your research…  (This is the rant…)

The Rant:

I’ve posted ideas for ROV competitions in the past.  I’ve talked to people who are building ROVs for competition.  I’ve offered pointers on physical layout, instrumentation, umbilical design, all kinds of things.  Every time I do this, I expect and hope that the person I’m talking to, be they a mentor or a student or a parent, or whatever, will say, “Yeah, we thought of that, and came up with a better idea.  Let me show you.”  Show me!

Instead I get, “Where did you learn all that stuff?”

I did my research!

Use Google.  It’s your friend.  If you don’t like Google, use another search engine.  Use the library.  There are books written on the subject of ROV design.  Some of them are for underwater design engineers.  Read them anyway, even if you’re in sixth grade!  Learn which issues apply only if you’re going down to 10,000′ of depth and which apply even if you’re under 1″ of water.  Others are written for students who are building ROVs for fun or for competition.  All information is useful.  Do your research!  And as soon as you think you’re done, open up a Google window again and start over!  Use your new-found knowledge to expand your search and get even more good ideas.

Before someone says that the Internet shouldn’t be used for research, please let me correct you. We use it all the time at work. This is how we find out what other people are doing, and if their methods are better than what we’re using. This is how we find out about new products that might make our workflow easier. This is how we get in contact with people who might eventually build stuff for us, or better yet design stuff for us. This is how a good percentage of my time is spent, and the benefits are enormous. If you want to teach kids how it’s done in the so-called real world, this really is how it’s done. It begins with a search.

This research can spare you months of wasted labor, because someone might already have tried your ideas out and either proved or disproved them. This research can spare you a blown competition because you might be able to read someone’s blog where their ROV of similar design failed for the following reasons.  This research can spare you a lot of your operating budget if you find out before buying parts that parts X, Y, and Z aren’t even waterproof! It’s worth having everyone on the team scour the Internet for ideas for a solid week before they ever set foot in the workshop. Fish for ideas. Scratch off the bad ones. Underline the good ones. Bring them in and brainstorm. Then go back out with your plan and see if there’s anything out there you can use to refine it.

If everyone did this, the competitions would be out of this world.

Getting back to control and instrumentation, at the very least I would:

  • Include at least two cameras:  One should face forward, one should face down.  Ideally their fields of view should cross inside the work area of your manipulator.  Inexpensive cameras can be had from places like Synetlink for less than $30 apiece.  The more video cameras, the merrier.  Especially if you have a logical system of setting up your monitors.  This helps avoid operator disorientation!
  • Use nice video screens:  The increased use of LCD flat panels and cameras on commercial trucks for reversing means there are scads of inexpensive flat panels available online.  I picked up a two-channel LCD panel off Ebay for $20.  It runs off of 12VDC, so it can run off of the same system that powers the ROV.  These things are cheap, and give you lots of options for laying out a logical system for your viewports.
  • Make an artificial horizon / compass:  Get one of those ball compass keychains, knock the keychain off, and epoxy it onto your ROV inside the field of view of one of the cameras.  In addition to providing you with compass orientation, it gives you an artificial horizon so you can tell if you’re nose down, nose up, or tipped over sideways.  This helps avoid operator disorientation!
  • Make a depth gauge:  Get some clear plastic tubing, and put a resevoir on top.  As your ROV descends, the increased pressure will compress the air in the resevoir, and water will rise inside the tube.  You can even spend some pool time with your depth gauge and calibrate it using a grease pencil.  When you come out of the pool, convert your grease pencil marks to something more permanent.  Epoxy this to your ROV inside the field of view of one of the cameras.  Now you know how deep you are at all times.
  • Make your umbilical as flexible as possible:  I saw a number of teams use almost rigid wiring in their umbilical.  If your thrusters are having to fight your umbilical in addition to the water, you’ll never be able to maneuver.  Use the most flexible cable available to you.  I’ve used a CAT-5 network cable to send video signal.  Over short lengths, you don’t need to re-balance the signal.  Just solder the wires.  With one video signal per pair, a single CAT-5 cable can handle four cameras.  If you’re using CAT-5 for your video, it’s worth it to install four cameras and use every pair in the cable.
  • Make your umbilical neutrally buoyant:  Most teams did this, so I’m probably belaboring the point.  But if your umbilical sinks, it’ll force your ROV into a dive, typically at some oddball orientation.  Likewise if the umbilical floats.  This helps lead to operator disorientation!
  • Consider a more complex control scheme than the usual “three switches, three motors, six wires” approach.  (More on this in a bit.)
  • Consider making a small, compact ROV rather than a big behemoth:  There’s a commercial inspection ROV made in Scotland that measures under a foot on a side.  It’s easier to transport, easier to get in and out of the water, and it gets the job done.  (See?  I told you you should look at commercial ROV information during your research!)
  • Consider materials other than PVC pipe:  I know PVC pipe is a favorite at ROV competitions, but attaching components can be cumbersome.  Other materials like UHMW plastic or HDPE should be considered as well.  These often show up as plastic cutting boards that can be had for less than $10 apiece.  Three cutting boards, a drill press, and a saw, and you can make a really good ROV frame that’s got tons of surfaces you can drill out and stick bolts through.

Getting back to the idea of a more complex control scheme, this is one I’ve had a hard time convincing anyone of.  But it’s the one I’d most dearly love to see used in competition:

With one exception, every single ROV I’ve seen in competition used a control board with some TPDT switches cross-wired so that flipping the switch one direction will drive a single thruster forward, and flipping it back will drive that thruster in reverse.  The center position stops the thruster.  It’s dead simple to wire up, but it’s got two huge glaring issues:

The first is that it’s like having a car where you are either stomping on the gas, stomping on the brake, or doing nothing.  You’re either cranking the wheel hard left, hard right, or dead center.  There’s no finesse.  There’s no ability to fine-tune your position.  It’s binary.

The second is that for every thruster on the ROV, you’ve got two heavy gauge wires going down your umbilical.  Three thrusters is six wires.  Four thrusters, and your umbilical starts to take on a Herculean strength and weight.  Using thinner wires increases your line losses, and decreases the effectiveness of your thrusters.  Seems like a no-win?

There’s a better way:

The easiest is to add a dry box to your ROV and install a set of relays that tie into those switches on the surface.  Instead of switching high current at the surface, switch the high current at the ROV.  12VDC is brought down to the ROV by a single pair of heavy gauge wires, and the wires used to control the relays can be very light “signal” wires, like a CAT-5 cable.  It adds the complexity of a dry box on the ROV, but the umbilical becomes much lighter.  Instead of six or more heavy gauge wires, you’ve got two heavy wires and two CAT-5 cables, one for thruster signals and one for video.

But the real fix is to move away from the switches altogether and go with something much more maneuverable: proportional control.

The R/C world has come up with all sorts of wonderful new products in the past decade.  One of my favorites is the electronic speed controller.  These are devices that plug in, just like a servo, and control the speed of a DC motor.  Want to control a thruster?  Plug it into a 12V ESC, plug that into a remote receiver, and turn it on.  Move the joystick on your R/C transmitter forward, and the motor gradually throttles up to full forward.  Slowly pull it back, and the motor ramps up to full reverse.  All of a sudden you get proportional speed, steering, etc.  The operator gains the ability to finesse the ROV into the tiniest spaces, to maneuver the manipulator down to the tiniest fraction of an inch.  The difference is night and day.

There’s another device from the R/C world that’s worth considering: a servo mixer.  Most ROVs use “tank-style” steering.  Two main thrusters, two main sticks.  Push them both forward, you go forward.  Pull both back, you go back.  Shove one forward and one back, and you turn.  A mixer does this for you, and makes for more logical control:  Push a single stick forward, and you go forward.  Lean it left, and you turn left.  Lean it right, you turn right.  Pull back, and it goes into reverse.  The mixer takes care of telling each ESC how much thrust to apply.  And this way you only use one joystick of your transmitter to do most of the control on your ROV.

“But radio transmitters don’t go underwater!”  Not by themselves, no.  But with a little help they can.  A few years ago I picked up a 2.4GHz radio system for about $50.  A quick Google search will find you one for about that price or better.  The transmitter antenna screws onto the top of the radio via a little reverse-polarized SMA connector.  If you’re using it to control an R/C airplane, that’s what you want to do.  If you want your signal to go underwater, there’s one more step:

RP-SMA cables can be bought from almost every electronics supply house in the world.  If you’re building a 50′ umbilical, get 60′ of RP-SMA cable.  Get a male connector on one end, and a female connector on the other.  Make this part of your umbilical.  These things are tiny, maybe 1/8″ diameter or less.  And they’re incredibly flexible, so your umbilical won’t suffer because of it.  Connect the dry-side end to your transmitter, and stick the transmitter’s antenna on the underwater side, inside the dry box.  Put the R/C receiver nearby.  Voila, for about $70 you have an R/C link to your ROV.  At this point you can add ESCs and servos to your heart’s content.

So put all this together and here’s what you get:

  • Four video cameras, four monitors.  One faces forward, one straight down, and one to either side.  The monitors are arranged the same way.  The cameras are aimed so that as an object slides out of the field of view of one camera, it slides right into the field of view of the next.  It feels like you’re looking out of the windows in the nose of the ROV.
  • Artificial horizon / compass and depth gauge, located in the field of view of the forward camera.  Even if you lose a camera, you always know where you’re pointing.
  • A 2.4GHz R/C transmitter provides proportional control.  The right stick drives the fore/aft/turn thrusters, and the left stick provides up/down thrust.  All other channels on the radio are available to drive the manipulator.  (Yep, a powered manipulator!)
  • A hefty metal geared servo drives a pushrod that goes through the side of the dry box.  Look up “prop shaft stuffing” on Google to see how this is done.  This opens and closes the jaws of the manipulator.
  • The umbilical consists of two 14ga fine stranded copper wires, one CAT-5 network cable, and one 2.4GHz RP-SMA coax cable.  Total size is less than 3/8″ in diameter.
  • The ROV itself is built small and compact, with the thrusters as far out toward the edges as possible.
  • The dry box uses Bulgin Buccaneer underwater connectors for the umbilical, each of the thrusters, and each of the cameras.  These are about $7 apiece.
  • The team has spare thrusters, already cabled to a Bulgin Buccaneer connector, so swapping out a dead thruster takes less than a minute.
  • The team has spare cameras, already cabled to a Bulgin Buccaneer connector, so swapping out a dead camera takes less than a minute.
  • The team has a spare umbilical, which they have tested in advance prior to competition.  In the event of an umbilical failure, they swap and move on while the other is inspected.
  • Everything, and I do mean EVERYTHING will fit in a largeish ice chest for transport.

Easy to take to a pool for testing, easy to pack up when you’re done.  Easy to unplug individual components to be worked on while the driver continues getting seat time.  Easy to service during competition in the event that a camera, thruster, manipulator, or umbilical is damaged.

And driving it?  Super-easy since it offers proportional control.

This is likely the last time I offer up these ideas on designing and building ROVs for competition.  My daughter starts middle school next year.  My son is two years behind her.  I don’t think she’s interested in doing ROV competition, but I know he is.  We’ve talked about a lot of this, and my son is on-board with the ideas I’ve presented.  And before anyone gets the impression that I’m one of those parents who will build the thing for the kids, drive all the decisions, and be one of those jerk parents/mentors who sucks the fun out of the thing, think again.  My son has been using my shop tools for years.  He hasn’t spent any time on the mill yet, but he’s spent time on the lathe and is a fair hand on the drill press.  He’s pretty skilled at driving R/C cars and helicopters, and will likely take to an ROV like a fish to water.

So no, I haven’t built an ROV like the one I described.  And chances are I never will.  But I’m looking forward to seeing him build it and use it in competition.  So here’s your chance to beat him to the punch.

– Tom


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: