Here she is, in progress. Just finished with the main work on the float, still needs some more sanding and painting and whatnot, but it’s getting there. Thrusters haven’t arrived yet (waiting on the BlueESC ones to ship) but it’s ready to go when they do. It’s using OpenROV 2.7 electronics with additional batteries and Servo control Headers. The front pressure chamber is a Crustcrawler unit, and the rear one (which will house the batteries and the emergency ballast release) came off of an older video camera housing.
Thanks! I have a Lulzbot Taz 3d printer, so I can print 12" x 12" pieces. I figured I would design the biggest 3d printable ROV frame I could manage and go from there.
The floatation foam is actually homemade syntactic foam, Made up of 1.5 parts 3m S38 High strength microspheres to 1 part West system 105 epoxy resin and 205 hardner mix. There are some generic no name Carbon Macrospheres (about the size of a BB pellet) in the middle. Right now the float is theoretically depth rated way deeper than the rest of the ROV (except the thrusters) but it might come in handy to have something that can support those depths for a later build. The tan colored filler sections are bondo automotive filler mixed with leftover microspheres. The float itself seems to float pretty well in limited bathtub testing, but I haven’t got the whole unit submerged yet to see how close my calculations were.
In hindsight, a 2 part microsphere to 1 part resin mix might have been better for such a small float. I also did not perform the “winnowing” step, where you pour the microspheres into water, scoop out the ones that float and leave the rest to settle. I probably would get 3 or 4 % more floatyness that way, but i didn’t want to wait a week for them to finish drying.
Still need to print the front pressure hull retention brackets, but those should be done this week.
Definitly looking forward to getting the thrusters mounted!
Thanks! I’ll be sure and post more progress pictures.
I expect I’ll have about 22-26 lb/ft^3 worth of Boyancy, with early “stick weights ont he float until it sinks” testing seeming to be in line with that. I should have a theoretical depth rateing of between 3,000 and 6,000 meters for the float itself, though the plywood base would likely have issues. it’s encapsulated so it might survive longer than it would unprotected. The rest of my ROV structure (particularily the pressure hull tube) at this point is only rated for around 150 meters, so it is overkill at the moment. I have no idea how well the 3d printed parts are going to hold up to depth. They were printed at 100% infill and have no void spaces, so they should do pretty well. The robot claw servos probably won’t work past 30m at this point.
I made a built up mold with a permnanet piece of thin plywood as the base, and then wax paper lined cardboard for the sides. The inner thruster intake was 3d printed, and the mounting screw holes were made of several standoff and plumbing pieces bolted to the base. I also bolted an embedded section of 20mm T-slot and the front lift ring, as well as the handle from the former video housing that makes up the rear pressure hull to that plywood. Once everything was bolted in and level, I used Duct tape to reinforce the sides.
A couple lessons learned, and why i have so much bondo on that thing:
The heat of curing syntactic epoxy foam will melt thin wall 3d printed parts. A wall thickness of at least 3mm should be used, otherwise you will get some warping of the embedded part.
Duct tape reacts to Epoxy Fiberglass. Badly. Don’t use it where it might come into contact or it will basically become the surface of your casting at that point.
Wax paper Melts. Once it melts, the paper inside the wax will happily permanently stick to your fiberglass. Use a mold release agent instead.
Cardboard for form edging isn’t a great idea.
even though it looks set, don’t set your float on the top of a 5 gallon bucket in the middle and expect it to not warp slightly, inflicting hours of needless corrective work.
I have a left over bucket shaped foam float now too from what was left over that I didn’t use. So if anyone is building a ROV the exact diameter of a 5 gallon bucket, let me know.
We’ve used West Systems 105/206 a bit and it gets really hot during curing. That’s the only epoxy that I’ve seen heat up to the point of melting other things around it. I imagine there is a better alternative that won’t heat up as much.
I’m sure if you ever decide to built another, you will do things differently
It was basically all I could get locally in quantity. I toyed with the idea of doing a Urathene variant instead, but I haven’t ever tried casting anything with it, so I wasn’t sure how well it was going to work.
I may end up having to make it a bit larger in front to support the robot arm when it’s in it’s final form. That’s going to be an entertaining project for another day. I need to find a better source of half round edge profiles that can be bent before I get to that point
Thanks! It’s getting there. The main side panels are going to get some cleaning and painting as well, but that won’t be for a little while. Those lights are two Ikalite PCA dive lights, with a depth rating to 300 feet. They are about the chepest decent depth rating offering I could find. The brackets are a combination of 3d printed bits (which will be available for free download when I get them uploaded to the site) and standard dive camera strobe arms.
The hull is just about ready for buoyancy testing. I’m designing and printing all the internal bits for that crustcrawler housing right now. So far It’s about 80% complete. I still need to finish the camera servo arm and servo mounts, and the enclosure for the control boards and primary battery mount plate. That and get some thrusters mounted.
At some point i’m going to need to punch two more cable holes through the other pressure hull endcap. Since I have the BlueESC variant thrusters on the way for at least the rear motors, I’ll need to pass a bunch of cables into the hull. One of the next challenges is going to be funding. I need to get a T-200 on order for the upper thruster, and then figure out how to fund the 4 $120 each Seacon passthrough cable assemblies. Once I have all that, dive date will be about a month out. I figure I’ll need to find about another $700 to put into it. All told so far I’m about $1200 into it, not counting the laptop for the control station which I already had on hand. The Thrusters and the pressure hulls were a good chunk of that.
Finally got the inside tube mounts done, so I can mount the brain and electrical bits. Batteries at this point are likely going to be in the rear pressure hull, but most of the rest of the electronics will be in there. There will also be a lid for the beaglebone box. I’m not intending to pot that, I just want to be able to work around it safer. It’s getting there!
Looks great. I’m still amazed with how big and thick your 3d printed parts are. How heavy is the whole ROV?
Electronics tube looks good.
One thing you might consider is consolidating the motor wires (power, gnd, and pwm for each) into just one or two connectors. For example, you could connect the power wires together and pass all though a single connector and pass all of the pwm signal wires through another connector. This may save money on connectors and holes in the enclosure.
It’s about 45 pounds without thrusters or batteries. 10 of that is the tube endcaps alone for the main pressure hull. The rear pressure hull is another 7 or so. The float is around 15 lbs .There is probably around 7 pounds of plastic +/- 2 at this point, but the plastic pieces also float.
The electronics is probably my weakest area of knowledge. If I combined the wires what would happen if I tried to use two or more of the motors at the same time? Using less connectors is good if I can make it work. Those things are expensive.
2 more frame structural pieces to print (the two tube holder arms), then just the internal Z axis motor mount and the tether management parts. After that it just needs a little paint and putty and it will be structurally done and ready for the first float test.
With the power wires, the main concern is the current rating of the connectors. Each T100 thruster can draw over 13 amps at full throttle, so you’d want a connector that is rated to handle that. You can consider using a 4 pin connector with two grounds, two power to double the current rating.
For the signal wires, each PWM signal needs its own pin. There is also a separate ground wire in the signal cable but technically it is connected to the same ground as the power ground wire, so it does not have to be used if you want to minimize wires. The signal requires almost no current, so you can use any connector. You could get a 6 pin connector and connect 6 signal wires.
Hmm, hadn’t considered the full amperage. Looks like i’m going to have to use a different pass through connector for the motor leads anyways. IN that case i’ll ddefinitly look at sharing the connections.
Got my first 2 thrusters today! I switched to the Non- Blue ESC ones so I could get them a little quicker, and the bluerobotics folk were happy to accommodate that. Here’s one mocked up to the side of the frame
Yeah, I do actually have an unreleased piece to adapt 2 motors per sideplate for the sides. I may have to go with 4 rear motors if it’s too sluggish with just 2 of them. I’m hopeing I can get by with just the 2 rear and the one T-200 top for now, and then add more later as funds permit. Aside from the float and the pressure hulls, the rest of the frame is actually pretty light comparativly, so it might handle ok.
Another option would be for me to lose the back pressure hull. mount the motor on the front section and make a shorter float.
Have you calculated your weight in water yet? Granted it is going to be like going through any fluid … a sports car will have a better drag coeff. then a school bus, but the close you are to neutral on weight that will help out a lot.
It should be slightly positivly boyant at this point. The plastic pieces all float, as does the float itself. The two pressure hulls should be about neutral with their respective equipment. I plan on adding a bank of fishing weights underneith with an emergancy “low battery, open the servo and drop the weights” module to return the unit to the surface in the event of controll loss.
Once I get the sideplates painted this weekend, I’m taking it to a friend’s house who has a pool for testing to see if the reality matches my math.
