Underwater Robotics Questions

[Greg Marra]

This is rather outside the usual scope of ChiefDelphi, but it's a forum full of smart robotically-minded people, so here I go:

I am taking a class at Olin College this semester, and our final project is to build a robotic fish. We’ve been building “robotic fish” all semester, but so far they’ve only been above water and haven’t seen any submersed action. (videos: http://vimeo.com/gregmarra/videos). The final project goes in a swimming pool.

We’re planning on building a watertight head compartment and using hydraulics to transfer power from inside the head to outside. The head will be rapid-prototyping-3D-printed, epoxied for watertightness, and have fittings to pass the hydraulic lines. I have high confidence this will work. What I’d really like to be able to do is use hobby servos outside of the head for controlling pectoral fins for swimming up/down. My research online suggests that waterproofing servos is a decent idea, but won’t really hold up that well. We may end up using more hydraulics here, but that's more complicated than I'd like.

Do you have any specific resources regarding building underwater electromechanical systems you could point me at, or any specific words of advice? My team doesn’t have that much planned out yet, but we’re working on a short timeline (4 weeks total!) so any advice or pointers at all would be extremely useful!

Thanks!

[sanddrag]

I've never done this but I had an idea.

You might consider keeping the servos inside the watertight enclosure/body, and running a music-wire linkage through the enclosure to the outside. Where the linkage passes through, I'd design a sealed grease box that basically has an O ring on each side where the linkage rod passes through, and the box is full of a very thick grease (maybe like wheel bearing grease). This grease box would not have to be very large. I'd think this would keep out the water.

EDIT: This is what I was thinking of http://en.wikipedia.org/wiki/Stuffing_box

[Tristan Lall]

My first thought was the exact same thing that sanddrag suggested.

Also, I'd imagine that you might be able to take the top cover off of a servo and overpack it with grease. Most of them place the potentiometer and motor in the lower compartment, so they would likely be safe enough as long as the grease doesn't flow too much. The viscosity shouldn't be too big a deal, if you use a powerful enough servo (and don't let it get too hot). In any case, once the gears spin a bit, you'll end up with a bit of a cavity surrounding them. The rest of the case should be pretty trivial to seal.

[Mikell Taylor]

Ask one of my former interns (Gui or Ana perhaps) about the robo shark they dissected -- I think the manufacturers used copious amounts of grease to "waterproof" that guy's servos.

[JesseK]

Greg,

I'm not sure of your monetary or time resources, yet I had an epiphany about something similar earlier this year during our SEAPerch development. I thought about using a single drive motor with a CAM shaft that moved the side fins up and down. Also, the 'fins' were made of a thin stainless steel skeleton with a flexible water permeable sheet of plastic/polymers/whatever instead of wood. I'm not sure if it's ever been done before, so who knows what the possibilities are.

If you did a single motor design, it's easy to encapsulate that into a watertight container.

Here is the project that 1885 did for our county in early 2008. We don't show how to make a watertight servo, but we do show how to make a watertight underwater motor. It stays water tight for 8-12 hours of operation, but if we used a higher-temp wax we could definitely see better longevity.
http://www.seaperchpwcs.org/

Go to Curriculum -> Build Manual -> Thrusters -> page 2-8. Hope this helps!

[Greg Marra]

Originally Posted by Mikell Taylor

Ask one of my former interns (Gui or Ana perhaps) about the robo shark they dissected -- I think the manufacturers used copious amounts of grease to "waterproof" that guy's servos.

I've also polled folks who have built swimmers for Olin's Design Nature course, and I think we've come up at the best DIY underwater servo solution possible on a budget.

1. Get a normal hobby servo
2. Open it up, coat circuit board in epoxy
3. Create stuffing-box-like seal with two O-rings between the output "horn" and the inside of the case
4. Reassemble servo in mineral oil
5. Dip servo is plasti-dip

I don't know anyone who has done all of these steps on a single servo, but doing a few of them gets you at least an hour of underwater runtime. Doing all of them hopefully gets you more.

Originally Posted by JesseK

If you did a single motor design, it's easy to encapsulate that into a watertight container.

The really tricky thing is we're going for a biomimetic swimming design, so just rotary motion from a motor isn't terribly useful. We need to wiggle like a real fish. I really like your link to how to waterproof motors though.

Ultimately it looks like we're going to go with a design that only has hydraulic components outside of our watertight shell, so maybe I won't need to run electronics underwater after all!

[EricVanWyk]

Originally Posted by Greg Marra

[*]Open it up, coat circuit board in epoxy

Conformal Coat works wonders. Let me know if you need some.