Scratch Built ROV V2

[b-29er]

Some suggestions on a v3? ok
-use a way lower KV motor, like under or near 1000kv motor, maybe the motors the DJI quads use. Water is a lot more dense than air, and i would believe you want controllability over speed. Then you could still use your 4 or 5" props and get a lot of thrust at low RPM so you don't cavitate
-Figure out your cooling. You could cut a hole in your pipe to mount a metal plate to mount your ESCs to for direct passive cooling, or use boat ESCs and run a pump for active cooling.
-RUNCAM SPLIT so you get full HD recording and a live uplink without strapping a camera externally
-mount everything to a small boat with apm: boat so you can have a base station further out on the water that holds position by gps. Then use a regular fpv tx/controller to run from wherever you like.

 

[PsyBorg]

I was thinking on the first version that props for air may not be so effecti e in water.

Kinda why there are boat props of short diameter with huge disc area. You may end up using reduction gearing to drive that type of prop for effective propulsion. Spin the motors faster and let the gearing provide the torque to move higher disc area props.

I am how ever glad to hear you have made a significant stride towards your goal. Did you by chance dev any if the underwater footage?

 

[b-29er]

I was thinking on the first version that props for air may not be so effecti e in water.

Kinda why there are boat props of short diameter with huge disc area. You may end up using reduction gearing to drive that type of prop for effective propulsion. Spin the motors faster and let the gearing provide the torque to move higher disc area props.

I am how ever glad to hear you have made a significant stride towards your goal. Did you by chance dev any if the underwater footage?

I think that's a touch of a generalization. You are correct in saying on certain applications, boat motors use a wider prop that would look ridiculous in an aviation standard

But when you get to trolling motors, you get stuff like this

and then there's this guy, that literally uses a 6:1 gear ratio and a 14x17 aircraft prop

So what's my point here? there's a purpose for everything. The top is great for gas engines, where you have loads of torque, kind of like how you see bullnose props on racing quads. The goal is to put as much power into the air as possible, efficiency be damned. Then you have the trolling motor prop, for where the goal is just to go efficiently and slowly for a good while. Then you have the aircraft propeller mounted on this $1100 custom gearbox, whose sole purpose is to turn the minimal torque that a person pedaling a 6:1 gearbox will generate efficiently.

So lets look at what the options are. The current setup heats up because the motor is trying to spin too large of a prop, and while the motor can discipate the heat from the coils to the water expeditiously, the ESCs can't and overheat in the stagnant air in the tube. Putting a smaller diameter prop will solve this, but the higher RPMs will give a lot of pitch speed. In short, you turn the ROV into a speedboat with overcontrol and power consumption issues. And as much as a gearbox might be a good option, its complex, especially since there's not a lot of people jumping out of their seats to gear down a motor designed to be high RPM. So i would recommend go with something that uses the same bell diameter so it fits the current v2, but has a dramatically lower KV rating, like the b-pack motor or this emax from banggood. Both are 22mm motors, so roughly the same size, but 1/3 the rpm, which means more torque and less heat, and thus more runtime

either way, trade some RIPMs for some torque, water's a different ball game from air. Maybe go for a composite 5" two-blade prop for added efficiency and to reduce the blade flex and you should see a lot finer control, more power and less power consumption.

 

[Hai-Lee]

You could always use something like this: https://www.thingiverse.com/thing:707341

Have fun!

 

[Snarls]

Thanks as always for the input and comments everyone.

With the last tests I was able to prove to myself the feasibility of a DIY ROV build. Now that I have had the opportunity to try and retrieve my friends quad, I think it is time to slow down a bit and spend a lot more time on R&D (I have some other ideas for locating my friend's quad in turbulent water in the short term). I am thinking of developing different subsystems of the ROV instead of jumping in and building the complete V3 right away. Version 2 will probably be a test bed.

I have been reading a lot of the Bluerobotics documentation, as well as checking out the parts they sell. I think they have a lot of great ideas, so I will likely be borrowing some of their concepts for the V3 design. I have also been checking out Ardusub, the submersible version of Arducopter for Pixhawk. They have support for hooking up a Pixhawk and Raspberry Pi on the ROV, and connecting to a computer groundstation via ethernet on the surface. It looks very interesting as it supports a number of features like HD video uplink, support for multiple peripherals, and craft stabilization.


Here are some features I want to tackle with version 3.
-Upgrade from a 2" i.d. tube to a 3" i.d. tube to make maintenance easier and allow for more components.
-Use an o-ring end cap design like that of Bluerobotics to further work towards water tightness.
-Use the Bluerobotics wire penetrator concept on tube end caps to provide waterproof wire routing.
-Buy lower KV motors (<1000) for more efficient, safer, propulsion.
-Construct frame out of acrylic, UHMWPE, or any plastic heavier than water.
-Upgrade thether and control method (TBD).
-Upgrade groundstation monitor and controller (possible Nintendo Switch style controller).
-Increase LED light intensity and positioning (so lights don't blow out camera feed).
-More appropriate battery location and monitoring.
-Add more motors to improve control.
-Move vertical thruster(s) away from the bottom of the ROV to avoid stalling by debris.
-Develop and accurate CAD model to calculate buoyancy so that minimal ballast needs to be added.

So all these goals are a lot to ask for, and version 3 of the ROV won't just pop out of nowhere. Probably the development over the next months will be testing of different subsystems on version 2. So really version 2 is not done with, it is just now a test bed.

I'm struggling between having an end goal of making an affordable, functional, DIY ROV for others to build, or an end goal of building the fanciest ROV I can make. Perhaps I will reach the former after some development, and then push further to reach the latter. Uh oh, did I just trap myself in another hobby?

 

[cranialrectosis]

Uh oh, did I just trap myself in another hobby?

Hmmmm SinkTest.. DiveTest.. AquaTest.. SubTest.. is Kapper reading this?

 

[b-29er]

Thanks as always for the input and comments everyone.

With the last tests I was able to prove to myself the feasibility of a DIY ROV build. Now that I have had the opportunity to try and retrieve my friends quad, I think it is time to slow down a bit and spend a lot more time on R&D (I have some other ideas for locating my friend's quad in turbulent water in the short term). I am thinking of developing different subsystems of the ROV instead of jumping in and building the complete V3 right away. Version 2 will probably be a test bed.

I have been reading a lot of the Bluerobotics documentation, as well as checking out the parts they sell. I think they have a lot of great ideas, so I will likely be borrowing some of their concepts for the V3 design. I have also been checking out Ardusub, the submersible version of Arducopter for Pixhawk. They have support for hooking up a Pixhawk and Raspberry Pi on the ROV, and connecting to a computer groundstation via ethernet on the surface. It looks very interesting as it supports a number of features like HD video uplink, support for multiple peripherals, and craft stabilization.


Here are some features I want to tackle with version 3.
-Upgrade from a 2" i.d. tube to a 3" i.d. tube to make maintenance easier and allow for more components.
-Use an o-ring end cap design like that of Bluerobotics to further work towards water tightness.
-Use the Bluerobotics wire penetrator concept on tube end caps to provide waterproof wire routing.
-Buy lower KV motors (<1000) for more efficient, safer, propulsion.
-Construct frame out of acrylic, UHMWPE, or any plastic heavier than water.
-Upgrade thether and control method (TBD).
-Upgrade groundstation monitor and controller (possible Nintendo Switch style controller).
-Increase LED light intensity and positioning (so lights don't blow out camera feed).
-More appropriate battery location and monitoring.
-Add more motors to improve control.
-Move vertical thruster(s) away from the bottom of the ROV to avoid stalling by debris.
-Develop and accurate CAD model to calculate buoyancy so that minimal ballast needs to be added.

So all these goals are a lot to ask for, and version 3 of the ROV won't just pop out of nowhere. Probably the development over the next months will be testing of different subsystems on version 2. So really version 2 is not done with, it is just now a test bed.

I'm struggling between having an end goal of making an affordable, functional, DIY ROV for others to build, or an end goal of building the fanciest ROV I can make. Perhaps I will reach the former after some development, and then push further to reach the latter. Uh oh, did I just trap myself in another hobby?

Some thoughts on this:
-remember, a larger tube holds more air. If your goal is to use less ballast, minimize tube diameter. Unless you need room for like a battery or something.
-o-rings are good, but prob a little hard to implement unless you feel like cutting the grove into the tube for the o-ring. That plug you were using is a pretty decent solution, just make sure to lube with vaseline. If you were really trying to keep the water out, you could use a bike pump and bike valve to pressurize the inside of the vehicle before usage to pre-equalize your pressure to your projected max depth.
-if a pre-engineered solution is available for the wire penetrators, may not be a terrible idea to use it. Epoxy/superglue may also be your friend here
-This b-pack motor or this emax from banggood are both exactly what you are looking for. Use the same mounting holes as your current motors, 1/3 of the KV, and should run off your ESCs. Pair up with some smaller props like these and you should be good
-You can always incorporate holes into the frame to put weights if you are good with a drill
-fan shrouds are always good, especially when you can model them and print off
-Everything about this, even with an accurate cad model, is going to require some level of actual sit down and test.

As i've said before, i may be interested in building one of these, and i have access to a CNC, so if you want some parts cut or something, let me know.

 

[Earthchill]

There are ESCs with pipes for water cooling. Not sure if it worth it though, considering the extra complexity with pipes and perhaps some sort of pump.

 

[Snarls]

There's a nice makerspace nearby my apartment that has a small CNC mill, laser cutter, and lather among other things. Not sure how effective they'll be so I'll keep your services in mind b-29er.

Active cooled parts are probably my last resort, just do to the complexity of adding pumps and pipes, and creating more leak sources. My main plan of attach is to use low kv motors with small props, and maybe to get higher amp rated ESCs. I could also add a heatsink which interfaces between the inside and outside of the tube, or find an aluminum tube which would conduct heat better.

 

[Hai-Lee]

There is a simple way to reduce pressure water leaks. As there is no need to maintain atmospheric pressure inside the ROV you can use pressure equalisation. This, (in its simplest form), is to have a bladder, (balloon), inside the craft's electronics area which is open to the water.

As the water pressure increases the water inflates the balloon and the internal/external pressures are equalized automatically. This will mean a slight loss of buoyancy as you descend but normally nothing that the thrusters cannot overcome! Should the loss of buoyancy be a serious concern then a number of small floats on the tether cable can bring the buoyancy back to neutral, (in steps)! Without serious pressure differential between inside the ROV and the outside will reduce the pressure leaks which you have had in the past.

Just an idea!

Have fun!

 

[LitterBug]

Thanks as always for the input and comments everyone.

With the last tests I was able to prove to myself the feasibility of a DIY ROV build. Now that I have had the opportunity to try and retrieve my friends quad, I think it is time to slow down a bit and spend a lot more time on R&D (I have some other ideas for locating my friend's quad in turbulent water in the short term). I am thinking of developing different subsystems of the ROV instead of jumping in and building the complete V3 right away. Version 2 will probably be a test bed.

I have been reading a lot of the Bluerobotics documentation, as well as checking out the parts they sell. I think they have a lot of great ideas, so I will likely be borrowing some of their concepts for the V3 design. I have also been checking out Ardusub, the submersible version of Arducopter for Pixhawk. They have support for hooking up a Pixhawk and Raspberry Pi on the ROV, and connecting to a computer groundstation via ethernet on the surface. It looks very interesting as it supports a number of features like HD video uplink, support for multiple peripherals, and craft stabilization.


Here are some features I want to tackle with version 3.
-Upgrade from a 2" i.d. tube to a 3" i.d. tube to make maintenance easier and allow for more components.
-Use an o-ring end cap design like that of Bluerobotics to further work towards water tightness.
-Use the Bluerobotics wire penetrator concept on tube end caps to provide waterproof wire routing.
-Buy lower KV motors (<1000) for more efficient, safer, propulsion.
-Construct frame out of acrylic, UHMWPE, or any plastic heavier than water.
-Upgrade thether and control method (TBD).
-Upgrade groundstation monitor and controller (possible Nintendo Switch style controller).
-Increase LED light intensity and positioning (so lights don't blow out camera feed).
-More appropriate battery location and monitoring.
-Add more motors to improve control.
-Move vertical thruster(s) away from the bottom of the ROV to avoid stalling by debris.
-Develop and accurate CAD model to calculate buoyancy so that minimal ballast needs to be added.

So all these goals are a lot to ask for, and version 3 of the ROV won't just pop out of nowhere. Probably the development over the next months will be testing of different subsystems on version 2. So really version 2 is not done with, it is just now a test bed.

I'm struggling between having an end goal of making an affordable, functional, DIY ROV for others to build, or an end goal of building the fanciest ROV I can make. Perhaps I will reach the former after some development, and then push further to reach the latter. Uh oh, did I just trap myself in another hobby?

Still Been thinking about tackling one of these myself for the quarry where I kayak. Puting a gopro on a stick and hoping for good footage has proven fruitless. I've seen a few interesting ideas out there, but not many good DIY designs. Using an F4 controller and Ardusub : https://www.ardusub.com/ looks like a somewhat inexpensive way to go. Using FrSky sbus s.port over the teather would be an interesting way to get all the data in and out with a very small number of wires. I saw one guy using that method tethered with a buoy that housed the RX/VTX so he could remotely roam all around the lake. i've seen a few Thingi's that use 2212 motors for the thrusters, and have quite a few of these already https://www.buddyrc.com/sunnysky-x2212-13-980kv-ii.html .

Have you done any work on V3???

Cheers!
LitterBug

 

[LitterBug]

Printed up one of these to test fit on some 2212 980kv motors I have.

May need to make different props. These aren't the best looking and don't look too efficient compared to what could go in there.

Cheers!
LitterBug

 

[Snarls]

Hey Litterbug. I am still working on the V3 slowly but surely. Right now everything is still really in the design phase, with every detail being worked out in the CAD before I plunge full force into the build. That said, I have prototyped the LED lights and the thrusters (without duct) and both are working as intended. Like you, I have chosen to use the same Sunnysky 2212 980kv motors. For the props, I 3D printed the prop design that BlueRobotics provides, with some modifications to fit the Sunneskys. I will post some pics and details when I make the ROV v3 thread (hopefully soon!).

I've been doing some research into ArduSub, and have made some unfortunately conclusions for you and me. For one, I have seen no support for ArduSub on F4 boards. There was maybe one guy who tried it, but there was some additional work that needed to be done to make it fully work. Next thing is that ArduSub does not support Sbus, or any RC input for that matter. It only communicates via Mavlink, which requires you to connect it to a computer or a device that talks in Mavlink.

Now my original intention was to use just an Arduino and analog joysticks on the surface to talk to an Arduino on the sub which I would program myself. I could still do that, but the control power of ArduSub is enticing. To get that though, I will likely have to use a computer and game controller on the surface. Which is ok, but I will likely not be using it from a kayak or other potentially wet scenarios. I'm still not sure which route I will go.