Homebuilt ROVs Forum

Hey all,

I new here and somewhat new to the world of ROVs. Since I was in 5th grade I have always wanted to build one but never had the necessary means to do so. For a science project in that grade I made a small submersible out of medical containers glued together that worked with a forward and reverse pump complete with ballast tanks and counterweight. Now since I am a robotics engineering student at Worcester Polytechnic Institute I now have the necessary means to build the ROV that I have always wanted to.

The purpose of this project is to force myself to improve my skill set, build an actual robot from non kit parts, gain recognition on campus in preparation for my MQP next year, and of course fun.

In order to keep the project moving fast and without delays I am going the route of just using simple bilge pumps for moving in Cartesian space. I plan to use four 1100 gpf bilge pumps at each corner attached with custom brackets to a PVC frame that I have already built.

Instead of using a neutrally buoyant design like most ROVs I would like to increase the complexity of the project by using a ballast tank like system, I will have two tanks made out of 6 in diameter PVC pipe and endcaps, inside will be smaller bilge pumps mounted and connected to valves. There will be two valves per tank actuated by solenoids (gonna get water proof solenoid valve combos) this will allow me to fill the tanks with water and flush the air out when descending to lower depths. The valves will also aid in retaining air put back into the tanks when the water is being flushed out for surfacing. I intend to have a compressed air supply on board made from canned air cans connect to a brass fitting setup, this rig will be connected to the solenoid valves on the two ballast tanks and be used to fill the tanks with air as the bilge pup works in reverse emptying the water. I will need some sort of sensor configuration to keep track of the ballast tanks but I have not decide what that is going to entail yet.

As for electrical control and features I am going to use high current relays to control the bilge pumps from a tether connected to my custom designed robot controller development board. The board will take command from a PC over serial com port using a RS232 shifter to communicate with the AVR Atmega 644P maser controller on my dev-board. I plan to control everything from a Java dashboard on my PC/Mac, in the dashboard I will have an option of camera feeds to choose from or I may have the camera I am putting on the ROV move in multiple directions. All the electronics are going to be housed in a PVC electrical junction box that will be sealed with silicone, is this a safe idea?

my dev board with ground planes turned off:


Uploaded with ImageShack.us

TO handle the camera feed I will be using a Netduino Plus with a USB Host shield to send the data from a USB web cam over ethernet to the computer where it will be decoded by the java dashboard for viewing by the pilot. In order to do this I will be writing custom software in C# that will run on the netduino to handle the routing of the USB camera data over ethernet. The reason for this is to avoid having to make a custom cable with op amps every 16 to feet to avoid signal being degraded for the USB camera.

Netduino Plus:


USB Host Shield:


This what I have in mind so far, in the future I am planning on adding a robotic arm with 3 degrees of freedom and a servo gripper, the arm will feature cameras on it to track where it is as well as encoders on each joint's actuator but that is long down the road when i get better with inverse kinematics. The initial goal here is to build the platform and get it working and get the vision system working.

Here is the frame I have build:


Uploaded with ImageShack.us

I am debating where or not reducing the height of the frame, its up in the air but I did not glue it together yet so its not finalized and open for change.

I still need to design brackets that will attach to the pipe at each corner to mount each pump, I am waiting until I get ther pumps so that I can get realistic dimensions off them before I design and manufacture the parts. Does anyone here recommend a material that is both strong and light as well as having the ability to be cut on a laser cutter?

More to come I welcome any and all criticism and suggestions!

Welcome to the forum and congrats on finally getting to put your ideas in motion. I know the feeling, I wanted to build one for a long time as well and these days two things came together to make it happen for me. Technology got cheaper and more available and that was met with me having extra cash to buy probably every PVC fitting ever made. lol.
I like your plans to go the ballast tank route. I think we all have this idea initially but are met with too many complications for the average joe to put the hardware together. Having said that, if I understand your design idea I might have some things to ponder. Now I'm NOT an engineer but here goes... I think you might be able to simplify your design of how to exchange water in your ballast. If you work with valving the canned air and can control it, I think you could eliminate the extra pumps by properly porting your ballast for water exchange. Idea 1: Provided your ROV design is balanced to always stay upright (not hard to do) you could vent the ballast to open water on the bottom, then have the controlled air valve at the top either "blow out" water to surface or reverse and allow ballast to vent to dive. Idea 2: This would involve one more bi-directional valve but would still work if the ROV becomes unstable or is knocked of axis by water current, gets attacked by a squid, etc. Set up like idea 1, but instead of just venting at the bottom of the tank, include one more bi-directional valve that would work in unison with the air in/out valve to allow water in/out. So both valves would actuate in tandem to let air in/water out, or air out/water in. This would eliminate the pumps and whatnot, but brings me back the usual dilemma, where do I find cool electronic waterproof valves with waterproof electrical connections that are waterproof to X amount of feet of water?? But maybe someone can help us with that?
On another topic, are you planning on having both RS232 and ethernet between topside and the ROV? Or did I misunderstand?
Sealing threaded PVC connections with RTV silicone that I didn't want to make absolutely permanent has worked fine. Keeps the water out and if I screw up, or something goes bad I can always twist them apart.
Oh, and a simple tip: Do NOT glue that frame together until you are SURE you like your design. If you share my luck you'll have a design change as soon as the solvent cures, every time. The fittings usually bite enough on their own to get from concept all the way through a pool test.
What do you think? Was this helpful or did I totally misunderstand your design theory? Hope this helps and have fun!

biglakeexile wrote:Welcome to the forum and congrats on finally getting to put your ideas in motion. I know the feeling, I wanted to build one for a long time as well and these days two things came together to make it happen for me. Technology got cheaper and more available and that was met with me having extra cash to buy probably every PVC fitting ever made. lol.
I like your plans to go the ballast tank route. I think we all have this idea initially but are met with too many complications for the average joe to put the hardware together. Having said that, if I understand your design idea I might have some things to ponder. Now I'm NOT an engineer but here goes... I think you might be able to simplify your design of how to exchange water in your ballast. If you work with valving the canned air and can control it, I think you could eliminate the extra pumps by properly porting your ballast for water exchange. Idea 1: Provided your ROV design is balanced to always stay upright (not hard to do) you could vent the ballast to open water on the bottom, then have the controlled air valve at the top either "blow out" water to surface or reverse and allow ballast to vent to dive. Idea 2: This would involve one more bi-directional valve but would still work if the ROV becomes unstable or is knocked of axis by water current, gets attacked by a squid, etc. Set up like idea 1, but instead of just venting at the bottom of the tank, include one more bi-directional valve that would work in unison with the air in/out valve to allow water in/out. So both valves would actuate in tandem to let air in/water out, or air out/water in. This would eliminate the pumps and whatnot, but brings me back the usual dilemma, where do I find cool electronic waterproof valves with waterproof electrical connections that are waterproof to X amount of feet of water?? But maybe someone can help us with that?
On another topic, are you planning on having both RS232 and ethernet between topside and the ROV? Or did I misunderstand?
Sealing threaded PVC connections with RTV silicone that I didn't want to make absolutely permanent has worked fine. Keeps the water out and if I screw up, or something goes bad I can always twist them apart.
Oh, and a simple tip: Do NOT glue that frame together until you are SURE you like your design. If you share my luck you'll have a design change as soon as the solvent cures, every time. The fittings usually bite enough on their own to get from concept all the way through a pool test.
What do you think? Was this helpful or did I totally misunderstand your design theory? Hope this helps and have fun!

Thanks for the lengthy response ! I appreciate your critiques and advice!

I appreciate your comments on the ballast system, after having a few days to think about it I have some design changes somewhat bases on what you said. I plan to have a larger pump be used for the purpose of filling both the tanks simultaneously to facilitate diving but still have smaller pumps inside each tank with valves to facilitate the flushing of the tanks when its time to surface. The internal pumps will be placed with their outputs downward coming out through holes drilled into the tank so the output is poking through the tank, on the end of the output there will be a solenoid controlled valve, that will be opened and closed as needed, it will be closed during diving and be opened when it is time to surface and pump out the water, when the water is completely flushed the valve will close. there will also be a valve for the pump that is filling the tanks, again it will be a solenoid valve and will be actuated as needed. Finally there will be two more valves on top of the tanks to let the air out as the tanks are filled with water during diving. For surfacing I plan to use paintball gun compressed air tanks because they are light weight and can hold a lot of air, they will be hooked up with a series of brass fittings, I am still looking for a solenoid valve that can withstand the pressure of compressed air, I think I may just use a standard metal valve and some sort of water proof actuator. There will be plenty more to come on the ballast system.

On the electrical side yest I plan on using both RS232 serial and ethernet, but the ethernet idea for the webcam is iffy because the netduino does not have enough internal memory yo pass video fram through, I am more than likely gonna use a bird watching camera with AV output and convert it on the other end. I still may use the Netduino to record the Webcam to SD card if it permits me to write directly to the SD as if it were memory. The RS232 will be going to my Custom controller board from my computer. On board the ROV will be a relay control board with a general IO breakout for all other functions so all the electronics will be wired to the same board in a waterproof PVC electrical junction box, wring up the ROV will be as simple as hooking a cable onto it and the other end to the control board.


UPDATES:
Since I was smart and did not glue my PVC (still haven't) I changed the frame design, its a bit longer now so that I can accommodate larger ballast tanks and have more space to house all the systems I will need for the ballast tanks to work.

The Netduino web cam routing plan has been scrapped and I am looking for a good AV output cam at a good price to act as my vision system, I am open to recommendations

With the changes to the frame I am now going the rout of modding the bilge pumps into thruster , there will be one on the right and left on the front for turning and two parallel to eachother on the back for propulsion and assisting in turning.