I have recently joined a RoboSub team as a freshman in college, and I am conducting research into pneumatic systems that function underwater. I was wondering if anyone had any insight on creating a pneumatic system that can still function while submerged for extended periods of time.
So far in my research, I have found a paintball gun tank I want to use for storage (Tippmann 20oz CO2 Tank) that I believe will still function perfectly normal underwater.
The biggest questions I have yet to find an answer to is if there are 4-way 2-position solonoids or valves, pressure regulators, and pneumatic pistons that function underwater. I’ve been talking with an upperclassman ME student, and we agree that pneumatic pistons and pressure regulators have the potential to function as normal underwater without any special features.
Any insight at all into underwater pneumatic systems would be greatly appreciated.
Ours didn’t work too well when we tried it in 2007…but I think the Desert Wave team (at ASU in the Phoenix area) built a pneumatic system for their robot this year, might look to see if you can find out what they did.
Thanks! I wish there was something as comprehensive or all knowing as chief delphi for RoboSub My understanding is that there isn’t much centralized community knowledge.
yeah, I don’t think the teams talk to each other much. My son helped start a team that ran for several years, they seemed to always be off in their own world.
fun trivia about the movie Spare Parts…the MIT robot was actually the Carl Hayden RoboSub robot from the 2013(?) competition
I ran NC State’s RoboSub team from 2019-2020. Unfortunately I did more electrical work so I can’t give you the exact details, but our solenoid manifold was inside an IP68-rated COTS box that we put grommets in for the pneumatic tubing and either a BlueRobotics penetrator or Fischer waterproof connector to connect to our pneumatics PCB (basically just a microcontroller connected to some transistors to switch on and off each of the solenoids)
Hope this is helpful - there may also be some good resources on the Robosub forums.
Thanks for the tip! I have found that locking everything in a waterproof box is a potential solution, but something that I would like to use as a last resort.
My first suggestion would be to put all in in a positive pressure sealed box with the rod going through the side wall with O-ring seals (and grease).
If you have your heart set on all of it in the water then you have many more issues.
Cylinders – you’d want to keep positive backpressure in the cylinder because “everything leaks”. By keeping a little back pressure air will escape through any orifice hopefullt preventing water intrusion.
Pressure regulators are more of an issue as most use a bleed orifice to regulate downstream pressure. this is a water intrusion point and will need a positive pressure check valve and maybe a tube for a air bubble seal.
The same is true for the exhaust port on the solenoid valves.
In all cases, you want to always keep the internal pressure of your pneumatic system higher than the external water pressure so that any leak point leaks air out of your system and not water into your system.
I don’t envy having to deal with fluctuating hydro static pressure in this application. good luck!
The one critique I have here (based off gut feel and un-substantiated) it to go for a double (or triple) lip seal vs an o ring. There may be a better seal I do not know of for linear motion on a ground shaft in an aquatic environment. Or better yet avoid the seal and go with bellows, water can be a pain in the butt to keep out of things.
Just keep in mind an o-ring is not a “fits all” solution. A single o-ring can hold enormous pressures, but only when it is seated properly, this rapidly (instantly?) becomes an issue with motion (as Jim Forbes said)
You are absolutely correct.
Triple o-ring seal with ~25% compression is as close to optimal as possible.
Grease helps with keeping the seal on moving (smooth bore) parts
Check out the MATE Explorer class Technical Reports. These are extensive technical documentation reports from event winners over the past several years: Archives . MATE is a different but similar underwater competition, and the Explorer class is intended for college-level programs, so you’ll learn a lot there that’s useful for you. (Fun fact: the MATE Ranger class now refers to itself as “watergame” , in response to FRC student calls for the same.)
I haven’t tried pneumatics underwater, but it’s interesting and has a lot of potential capabilities. Reducing and relieving air backpressure may be your big challenge: on FRC robot control systems the relieved air is simply vented to atmosphere, but underwater this causes problems. First, the vented backpressure air (say, from when you move a piston, and you allow the air from the compressing side of the piston to escape through the control solenoid) will simply go into the ambient of your sealed underwater control enclosure; this will build up pressure in the sealed enclosure until the “backpressure” zone has nearly as much pressure as your control air, to the point that your pistons can no longer move, or have significantly reduced power. Second, vented air from pressure regulators (part of the regulation process) will also build up in your control enclosure, causing the same problem as cited above – or perhaps even bursting of the enclosure box. You’ll need methods to vent the control box to release excess pressure without allowing water inside; this could be a vent tube that goes all the way to the surface through the control tether. Or something more innovative.
Keep in mind that for the Robosub competition mentioned by OP, the ROV is completely self contained. There is no tether.
Typically there aren’t enough pneumatic actuations during the course of a run to require venting of a self-contained pnumatics enclosure, since there are only ~5 actuations maximum during a run with very low volume cylinders. I believe I have seen some teams using venting valves on their pneumatic enclosures to vent extra pressure to the water. Keep in mind the water pressure at the depths these robots operate is ~5-10psi.
It’s probably just the way we’ve built things in the past, but dual/triple o-rings always bug me. One correctly seated o-ring will do the job correctly. If you think one will fail, then focus should be on installing it properly, not installing another one or two that might also fail.
Thanks to everyone for the input! I think I’ve settled on containing as much as possible within a sealed container. Pretty much everything will be contained within the container except for the actual pistons.
Can you explain what you mean by this? For context, I’m will be using a spring extend, air return piston.
Hi I’ve been looking into similar systems for my teams robosub vehicle. So i’ve done a good bit of research/prelim designs/looking at how other teams have built there systems.
I would probably rethink this decision and use double acting cylinders instead. Spring return cylinders typically function by having a small hole that is open to atmosphere. When you put your system underwater the spring half of the cylinder would flood with water. The small hole will restrict flow of water in/out and would slow down the cylinder motion. You could also run into issues with rust inside your cylinder depending on materials. Using a double acting cylinder is better in this application.
From working on MATE vehicles I’m fairly confident that for the run time and depth for robosub it’s fine to just have the cylinders be in the water and run the pressure tube into a sealed enclosure with your solenoids. I’ve seen some teams use pressure vents like these for overpressure safety in enclosures. I haven’t had a chance yet to test that for a pneumatic enclosure but I’ve seen it used with battery enclosures as a safety system to prevent over pressurization in case of a battery expanding.
Thanks for the tip! I was planning on using that so I could halve the number of tubes I need leaving my enclosure, but you raise some valuable points about switching to a dual acting.
What did your research find for what tanks teams were using? I really like the idea of using a paintball gun tank, but I’m not sure how to interface that into my loop.
From what I can tell a lot of teams seem to be using paintball HPA tanks. I haven’t seen information on how they are doing regulation from that tank. Some teams seem to be using metal clippard tanks like these or similar parts. I’d guess those tanks are running much closer to how an FRC system is setup.