water-proofing a servo

[Admin]

Does anyone have recommendations on how to water-proof a servo for full submersion?

I called both Futaba and Hitec, and both said "none of our servos are water-proof" but "many are water and dust resistant."

"They cannot be submerged."

I have also seen the balloon method, the fill the servo with grease method, and the plastidip method . . . but they all kinda suck . . .

[Gopher]

Could you possibly encase the servo inside a larger, water-tight case? the servo movement could be transmitted by a rod passing through a tight rubber neck. If it's a continuous servo, parts must exist, perhaps something scavanged from a cheap toy rc submarine?

Not sure if this will help, but it seems nobody else has any ideas...

[Admin]

So I have seen that method too, incasing it in a water tight box, using a rod, and something called a 'rubber boot.' It attaches between the rod and outer casing. I have seen pics of this rubber boot but cant find where to buy any . . . ideas?

Im thinking about just incasing the entire servo and push rod in this big blob of highly flexible liquid rubber . . .

[Gopher]

Not sure; the two closest things I can think of are rubber nipples (for baby bottles) and a kind of rubber switch/button covers I've seen used on boats; nipples you can get anywhere, and pretty cheap too. The others I'm not sure but a google search for boat electronic components might turn up something. Might be able to scavenge one from a generator, or other outdoor equipment. Hell, they might sell them at Home Depot, I've never looked for anything like that before.

[dunk]

so from much experience messing around in boats, any sort of rubber seal will always leak a little bit.
the further under the surface the greater the pressure so the more leakage.
this doesn't have to be a problem. you just have to remember to leave a small trap where the drips can form a pool away from your electrics.
with a boats propeller shaft the water just settles in the lowest part of the boat.

i'm curious what you have against the idea of filling a servo with grease?
admittedly the increased drag on the components inside the servo could be a problem.
many pleasure boat bilge pumps are constructed this way. the idea being they have to be able to operate underwater if you start sinking. if you open one up they are full of light grease.

hmm, this has me thinking, try looking at pleasure boat autohelm servos. i've seen these operate in some conditions that would describe as close to underwater. (they are definitely more water proof than any waterproof coat or leggings you can find.)
i imagine they would all be a bit too big for a small robot project but worth looking into for ideas on how to seal the things if nothing else.

dunk.

[Admin]

Ok I didnt mention these details before because I wanted to not overcomplicate the question . . .

So there will be 4 servos in a box. The robot will have a total of 4 of these boxes, for 16 servos. Whatever I do cant take too long to impliment, as even an hour per servo is 16 hours of work . . .

The boxes rotate about 100 degrees sinusoidally, and will be under water for about 3-4 hours total. I thought of your drip idea, but in this particular case it would be risky.

I am warry of the rubber boot idea because each servo needs 2 shafts exiting the box (one shaft for each end of the servo horn), so thats 8 rubber boots per box, for a total of 32 possible leak failures  The box itself is only 4"x1.5"x1.5" - I must fit 8 boots on just one side.

I think I will fill the servos with grease after all (or some other low viscous non-conductive oil or jelly) as a secondary waterproofer. I am pushing the limits on current draw and servo force, so Im a little worried about increased drag on the servos - might not have an option tho . . .

To add another complication, the water proofing shouldnt thermally insulate the servo too much as I dont want it to overheat and burn out . . .

The servo I am using is the Futaba S3110. I was falsely led to believe by multiple erronous websites out there that this servo (or any servo) was waterproof, so I already have it built and working. Now just gotta fix this itsy bitsy prob . . .

Ok I realize this is really over complicated . . . If I was just making a functional robot I would never have gone this direction . . . but there is good science research that requires this and its definitely pushing my limits as an engineer!

I think I will just test out my servo in a rubber blob idea and see what happens . . .

[dunk]

so i'm starting to doubt my previous post.
i'm wondering if taking a servo that is designed to run in air and filling it with grease will add too much friction to the little motor inside.

what about if you stick with your rubber boot idea but fill the space inside the boot with grease but leave the servo empty?
you said your servo was reasonably well sealed but not waterproof. i wonder if it would be well enough sealed to stop grease getting in?
this way the grease would form a barrier.
even if a small amount of water contaminated the grease you could service it occasionally, changing the grease.
if a small amount of grease found it's way into the servo there is no real harm.
a kind of messy solution but could work.

dunk.

[polar bear6]

i dont know if this helps, but you can have the servo electronics in another box than the servo, so you only have to wory about waterproofing the motor itself.

[Admin]

ok next question, anyone know of any non-conductive low-viscous grease i can use?
servo grease is too thick, i think . . .

just a weird off the wall idea, maybe i can fill the servo up with de-ionized (non-conductive) water to balance the fluidic pressure, and use a simple o-ring seal to prevent small leaks . . . fyi, de-ionized water is cheap and easy to get . . . too bad i cant just fill up an entire swimming pool with it . . .

as for the electronics, since its in a seperate compartment that i never need to open, im just going to superglue that end shut.

[Gopher]

Saw a computer mod where they tried filling the case with distilled water. It didn't work, and ultimately they used vegetable oil instead, which worked great. I believe they ultimately recomended motor oil instead of vegetable?

looking for the link... here it is. They describe their problems and solutions in detail, might be helpful.

[edit:originally said mineral water instead of distilled water - big difference!]

[dunk]

so the lightest grease i can ever remeber seeing was meant for high performance bicycle bearings.
i think it was referred to as "white lithium grease" and doing a search on the net seems to confirm this.

interestingly, doing a quick search for bicycle grease on the net i found some guy sugested partolium jelly (ie. vasaline) in his bicycle hubs.
that might just work for you as well. it's definalty very water resistant, non cunductive and not too thick.

dunk.

[megaman935]

If I need something, I always check ebay when I can't find just what I want. They usually have what I want....... Just a suggestion on the 'rubber boot' thing....

[Admin]

whoa good ideas!!!
man that cooking oil idea . . . totally whack, but i think its just what i want.

i have access to lithium grease too, as it comes in this spray can foamer thing to grease up tools for our CNC.

time to experiment 

maybe around christmas ill write a really good tutorial on waterproofing servos, after i test and try everything out.

so i just spoke to a friend of mine about the oil PC, and he said mineral oil would work better than cooking oil . . . but ill look at both

[Gopher]

Hope it works out

The guys that made the oil-cooled computer admitted vegetable oil was a poor choice, and that they used it in their test because it is cheap, readily available, and proved the concept well enough.

[Gopher]

Hey, you were asking about the snake robot vid because of your water-related project, and it got me thinking. Have you concidered not waterproofing your parts at all, but simply putting them together in a flexible, waterproof container? It doesn't even have to be a container, it could simply be a waterproof latex skin; you just build the bot ignoring the water, just giving it the right shape and movement, and then wrap it in a watertight skin. If you vacuum the air out with a pump when sealing it, it would keep water out far better than a rigid air-filled case.

No idea how you'd make a custom-shaped latex skin, but latex is just an example. I'm sure there's an appropriate textile out there which can be tailored into custom, water-tight shapes.

[Admin]

That was actually my first idea, but it had leaks. I was never able to make a perfect seal around the protruding parts. 

[Admin]

Browsing some more for ideas I found someone suggesting putting vasilene around the output shaft of the servo, followed by a coating of silicone oil, claimed to work 'for up to 5 foot of water.' Perhaps it can be filled with de-ionized water for extra measure . . .

So I cant think of any clean way to fill a servo with the oil, other than assembling it while submersed in a cup of it. Hmmm I also have to keep the oil from getting in the motor to avoid adversely affecting the torque . . . And I think the very first gear that comes out of the motor is extremely low torque but high velocity . . . so might be strongly affected too . . .

I think Ill just order a few servos and abuse them with all of the above mentioned ideas . . .

[Gopher]

You misunderstand, I was talking about not having any protuding parts. Just a fish-shaped bot, and wrap the whole thing up.  Still obstacles to overcome, but it'd be a lot easier to get a total seal at the opening if nothing is protruding through the opening. There might be a reason this is impossible for your application, though, if you have sensors or elements that must be in direct contact with the water...

[Admin]

It has a really wierd shape, like a fish. No container (or a wrapped sheet of rubber) can be easily made for that shape.

Anyone know what would be better to fill the servo with, silicone oil or mineral oil?
I cant seem to find these numbers for either - surface tension, viscosity, conductivity, and a hydrophobic comparison of both. I think either will work, I just want to quantitatively decide which is better.

[Admin]

After spending hours today searching online, I came up with this info:

Mineral Oil
surface tension: 25mNm-30mNm
viscosity: 30cSt-100cSt

Silicone Oil
surface tension: 20.5mNm
viscosity: 50cSt-500cSt

I have decided on using mineral oil because:
Mineral oil is more like to be used in power transformers as opposed to silicone oil, so I reason someone has a reason for that . . . Mineral oil has a higher surface tension, meaning it is less likely to leak through tiny gaps. It also usually has a lower viscosity, so less torque resistance on the motor and gears. And lastly, mcmaster sells mineral oil but not silicone oil.

I couldnt find any info on how hydrophobic either were . . .

[Arislan]

Try asking in the Subcommittee.com forum, they generally answer quite quickly.

[Admin]

Hmmm useful forum for underwater robots! Will use that for research now 

Unfortunately couldnt find anything on waterproofing servos . . .

Anyway I ordered a gallon of oil and some soon to be victimized servos, will test that as soon as it all comes in.

[Admin]

It worked!

So I put it in oil, used O-rings, superglued it all shut, and it worked!

I got videos and step by step images too 

Ok so I didnt fully test it . . . put it in only 2ft of water and ran it for only 5 minutes . . . Ill give it more brutal tests when I put it on my robot . . . I estimate it to handle like 50ft+ depth, since oil filled motors are used for extreme depth deep sea diving vehicles.

A very thin oil slick appeared to float at the top surface of water . . . possibly a leak, or just leftover oil I didnt clean up when I was done . . . Slow oil leaks are my only worry for failure . . .

Ill write up a tutorial on it when I can, but might take several weeks . . .

[dunk]

cool. so that must be a bit of a worry off your mind.
i'm curious, so did you fill the whole servo with the oil in the end or just part of it?

i was once toying with the idea of making an ocean going solar powered bot, using GPS to navigate that would send a text message with it's location when it got close enough to shore.
the idea has never got any further than being an idea i throw around my brain from time to time but making motors watertight for extended periods was always one of the stumbling blocks.
nice to hear it might be easier than i thought.

dunk.
(too impatient to wait for the tutorial.)

[Admin]

i reassembled the servo in a jar of oil, fully submersed. lemme tell you how hard it is to reassemble a servo in a light refracting jar of messy slippery oil while wearing gloves . . . lol damn thing took about 30+ minutes to put back together haha

didnt help when all the tiny gears decided to leap out of the servo . . . or that screwdrivers are impossible to rotate when your hand is oiled up . . .

then the oil kept leaking out so used lotsa super glue . . .

i guess it was a 2 hour process, plus a month trying to figure it out . . . but i think you can do it in like 30 min following the tutorial i will write up . . . ill point out all the mistakes i made so others wont make em . . .

but generally, im very happy with the final product. couldnt have done it without everyones suggestions! 

ocean going solar powered bot, using GPS to navigate that would send a text message with it's location when it got close enough to shore

So did I! (minus the text messaging part) Meant to do it years ago . . . tried recently on my new robot boat i made a few weeks ago but alas the chassis was too unstable for large waves and high winds . . .

[dunk]

so i planned on making the buoyancy very low so it was practically submerged to help avoid this problem.
this was why i was so concerned about any water leaking in.

there are a few projects out there that basically drift in ocean currents.
i wanted something that could actually make way against the weather.
a sailing craft would be nice as it could still make way through the night but solar power and electric motor driven would be far easier to implement.

let me know if you ever need any input on navigation strategies.
i have a lot of small craft experience in tidal waters and ocean crossings.

dunk.

[JesseWelling]

I bet if you drilled a hole through the srewdiver handle and put a screw or pin or some thing else to press your thumb against through it it would generaly be easier to use with oil. Just a quick thought I had if he's going to make alot of these.....

[Admin]

Okie its done:

http://www.societyofrobots.com/actuators_waterproof_servo.shtml

my longest tutorial yet . . .

[Dosbomber]

I'm wondering if it would be prudent to add a dab of white lithium grease to the gears (if that other stuff dissolves) before oil-filling the servos..

Great write-up.  Plenty of details, and always good to hear about the mistakes.  It's good to learn from your mistakes, but cheaper to learn from someone else's.   

[JonHylands]

Care to post some pictures or at least a more in-depth explanation of what you're doing?

I have built an AUV, and have a pretty well-documented site about it:

http://www.huv.com

- Jon