Society of Robots - Robot Forum
General Misc => Misc => Topic started by: Admin on October 19, 2009, 06:01:02 PM
-
I just got emailed by a guy who does the model boat competitions (http://www.societyofrobots.com/robotforum/index.php?topic=9426.0). He tried out the SoR servo waterproofing methods (http://www.societyofrobots.com/actuators_waterproof_servo.shtml), and also varied some of the ideas. Figured it'd be helpful if I posted his email:
I'm including a picture of my model warship, performing fine even though badly listing and nearly underwater. Not a single servo twitch. She finally sunk in 5 foot of water and was there a good fifteen minutes before being raised and was still working fine after raising. This is the article I posted on the local warship group. Contained within are my process and some minior improvements I made to your technique. I then pointed them to your page as I didn't invent this method.
Thank you very much for your articles.
Walter Hansen
Why, yes, and what a wonderful picture indeed. She is so photogenic even when barely afloat with a dangerous list. Look at that cute wave pattern. Also I could not have fought on so long without well sealed decks as I think you will all agree. Too bad she was listing at this point, I still had a full load in those high point torpedo tubes ready to eat balsa.
As of just a few moments ago, I connected the radio and battery and they all seem to be doing fine after being sent to the bottom (5 feet) for fifteen minutes and going around the pond like this for a lot longer. The receiver was a two part 2.4ghz unit and I put both units inside condoms, put poster putty around the wires and zip tied them. I knew this wasn't perfect, but was a good time compromise with a radio delivered to the pond new. The small part with only two wires had no leakage. The large unit had some leakage, I probably didn't squish the putty into the wires good enough, but it collected at the bottom of the condom away from the radio proper. The radio appeared slightly splashed with water, and was post mortem treated with corrosionX. When brought up from the murky depths the radio, all servos and the ESC were working fine.
I will mention here that I have found in tests that although servo connections are fine with dunking, electrolysis happens between any positive and negative terminals quickly resulting in the positive terminal turning green. I don't think there is any way to avoid this, so I recommend sealing any connections well and/or waterproofing them all together like I did with the condoms. This is not a intimidate problem, but the corrosion will lead to an early demise for the components.
SERVO WATERPROOFING
Using this method I tested a servo under 6" of water for half an hour with no issues. I then put the same servo in a jar of water on my nightstand overnight. In the morning it worked fine still completely submerged. I have heard of these servos being reliable constantly functioning in ten feet of water and being tested to thirty feet. The servo I did first servo was controlling my forward gun for LMS. I did this process at work on breaks as it's a lot of work on it for a few minutes and set it aside for a few hours. I did all my servos this way in a couple work days. Nothing was really hard to do.
I can't claim to have invented this method. I will link to the guy I got it from, but I improved on it somewhat. He is a robotics guy and his target was waterproof servos that are meant to function underwater continuously. Previous to my work my servos would glitch if shot by a water gun. This method works by not only keeping water out of the servo, but also by replacing any places for water to get into with something other than water or air that will not harm the servo. Pressure should have little effect on this method as the servo will exist as a sold for the lower half and as a liquid for the upper half.
I used silicone II tub and shower, I understand that all manners of silicone and RTV will work. Also marine epoxy works and may actually be a superior solution, but just try un-doing it if you ever want to hack the servo or somehow mess up. I do question whether the silicone that is deepest inside ever really cures, however I can't come up with a reason this would ever be a problem anyway.
First dissemble the servo completely. I was using Futaba S-148s. Make sure you know how to put it back together (use a digital camera). There should be a top case, a bottom case and possibly a bottom cover. Now you should have a small motor, a potentiometer and a circuit board connecting them on the bottom. As mine were old servos and I had run into problems with their short leads, I de-soldered the old leads and soldered on fresh long ones. Be sure to test the servos each time you do something to make sure they work fine. I had one solder joint come undone and had to re-solder it midway through the process.
Now take your silicone product and squirt it from the inside of the servo out. Try to get it into each and every nook and cranny and around every component. Fill out until you have a silicone shape that looks like it will just fit back into your case. I covered the lower half of the potentiometer and covered up a few holes in the top of it. I also covered at least half of the motor case and covered a non-screw hole in the top of the motor. Also be sure that you seal where the motor meats the circuit board. I tried to force silicone in here. Now put the whole assembly into the lower case, some silicone should squirt out around the circuit board if you did it right.
Take an handy tool and evenly spread the silicone that squired out over the exposed circuit board. If you need more add it as you want too much once again. Once that's done put the bottom cover on and once again silicone should squirt out the seams. Spread this silicone evenly about a quarter inch all the way around the servo to further seal the seam. If your servo doesn't have a bottom cover you'll have to figure out how best to do all this. Spread some silicone on the wires for a good 1/4" or so to seal them better and prevent wear.
On my servos you now have to screw two little screws back into the motor from the top. Once done I checked the servo for functionality. At this point I found that some of my servos had two little holes that went from top to bottom and I filled these with silicone and smoothed off the top. I then put the top on and screwed it together to let it dry, overnight is best, but a couple hours seemed to do.
For the top part of the servo we're going to use a liquid to keep water out as the parts in there have to move a lot. I chose marine RC racing grease as I had it on hand and it was about right. I tested some with a multimeter and found it non-conductive. The original guy used mineral oil. Mineral oil is more prone to leaking but otherwise works well. Marine grease is heavy and will considerably slow the servo. However for our uses it really doesn't seem to be much of an issue. I did not notice any appreciable delay in the short throws that we use. It was most apparent on the rudder where it took a bit longer to throw it from side to side, but I don't think this affected anything at all.
Once dry, take the screws back out and remove the top cover. You probably have to pick some silicone out of the screws. At this point you can also create a continuous rotation servo if you like (instructions not included here). Now put marine grease under the gears, on top of the case and under the case cover. You want enough grease in here that some squishes out when you put it back together. You'll also want an O ring to fit around the exposed servo spindle. Now put it carefully back together with all the gears and pins in the right place. The cover should easily fit back on, don't force it or you might break something. I made a continuous rotation servo accidentally by forcing it. Sometimes rocking back and forth will help get things in place. If not take a break and try later, getting frustrated is bad. Most of my servos went together easily, one took half an hour of sweat and many tries to get it aligned.
Once screwed together test the servo and verify that it works right. Now clean the grease that squired out with alcohol, but leave the grease that squired out the spindle hole (if some did). Now take your handy dandy silicone and lay a bead all around the case to seal the seam. Flatten it nicely, I liked mine about 1/4" wide. Let it dry again at least until it's hard to the touch.
Take your O ring and give it a nice coat of grease and put it onto the servo spindle. Then shoot some grease down the hole in the servo spindle. Yes a screw goes here, but the grease will seal the treads and on my servos that hole goes all the way down. Now put whatever arm your using on and screw it down and test and see how well it works. As a final touch I spread some silicone on the screw heads holding the case together. It's probably best to let everything thoroughly dry for a day or more before water testing (if you can wait).
USING OIL - FASTER SERVOS
Using mineral oil is more difficult, here's what I suggest (and what I actually did for my first servo with grease). Drill a small hole in the servo top case. Make sure to put it in a non critical location. Put all the gears in, put the top case on and test to make sure it works, then seal the seam with silicone or perhaps CA and then silicone. Dry completely.
Now use a syringe and put your oil into the top case, you'll probably have to swish it around a bit to get all the oil in and the air out. Once you have it filled to the brim, clean the top case, so its not oily. Now you want to plug your hole. I used just silicone, but that was with grease. You may want to use CA if you can and then silicone on top. The smaller the hole you can use to squirt the stuff in the better. Then oil the O ring and put it together like normal. I'd still use grease to seal the spindle threads though. This method should not significantly slow the servo, but will be more prone to leakage.
Here is the guy I learned all this from. Note that he starts from not so good ideas and progresses to better ideas towards the bottom of the page:
http://www.societyofrobots.com/actuators_waterproof_servo.shtml (http://www.societyofrobots.com/actuators_waterproof_servo.shtml)
I'll have to forward this to him.
(imagine a cute or dirty tagline here)
On Mon 19/10/09 1:02 PM , Rob Wood [email protected] sent:
> In yesterday's LMS contest, we saw two good examples of just how
> critical it is to waterproof servos. In the last round, when only 3 combat
> ships were left in the cage, the Spahkreuzer had to be pulled because its
> steering servo got soaked. At the same time, Walter was able to continue to
> steer his Haguro, even though its stern was underwater.
>
> Later, Walter explained how he waterproofed his servos, and his approach
> obviously works.
>
> Walter, could you please explain the procedure you used to waterproof
> your servos?
>
> Rob