Designing home made ROV


#1

Hi Guys, First blog post but long time looker on OpenROV, awesome work.

Im lucky / unlucky enough to work with ROV's for a living in the North Sea in the Renewable Energy and now Oil and Gas in the UK, I work with ROV's as small as the Video Ray, to a Seaeye Falcon, to a Sub-Atlantic Mohawk up to a Perry XLS

I'm currently designing and getting the ground work down for a home made ROV to explore some of the shore based wrecks around Orkney and eventually something gutsy enough to explore the wrecks in Scapa Flow.

Currently thinking on making the first vehicle Battery powered (I've seen autonomous vehicles like AUV's and AIV's use LiPo's to great success) and using arduino with a shield for thruster control, that means learning arduino but keeping the camera to a simple mini CCTV at least for the beginning,

The need for micro-controlled thrusters is quite high, variable power thrusters brings an immense degree of control compared to on / off of thrusters so for the initial design that's the main design point. Preferably I would like 4 lateral vectored thrusters all working at the same time with one vertical thruster but I need to do my homework on the limitations on Arduino boards etc, worst comes to worst I'll just go for the 3 thrusters

I'll post on my findings etc and progress but any and all input is extremely welcome :)

cheers.


#2

Here's a question for the community...

Would it be possible to run/operate two beagle/bones capes in tandem , (through two twisted pairs), therefor providing the option of up to 8 servo channels & 2 video feeds? I suspect there would be some legwork required in modifying the control software in "cockpit"... Is this feasible?

Best regards,

Phil


#3

Hey David,

That sounds like an awesome project, and please let us know if we can help! I look forward to seeing what you come up with!

....For Phil- Yes! .. and not only could you run to BB/Cape systems on two twisted pairs, with two homeplug adapters on the ROV, you could run the set on only one twisted pair! Homeplug standard devices can be linked in parallel over the same line, so as long as you could write software for them to share tasks well and not conflict with each-other about addressing, you could theoretically get two feeds and eight servo outputs that way.

Eric


#4

Hi David,I share your enthusiasm. Depending on your current level of experience I would suggest a few books to help you out. A very good and basic text is http://www.marinetech.org/underwater_robotics/. Another one, which is more user enabling and cut and dry in regards to the mathematics and design/control is http://www.amazon.com/The-ROV-Manual-Observation-Remotely/dp/0750681489. If you have these, then awesome. You may also want topick up http://www.amazon.com/Advances-Unmanned-Marine-Vehicles-Control/dp/0863414508/ref=pd_sim_b_3 for some academic exploration into unmanned systems. But being a ROV pilot, your knowledge may already surpass these texts.

An aside, yet related. Myself and 2 others here in the states, Livermore CA, are building a pseudo-work class ROV on a maker budget. Very interesting and exciting, with the exact mission requirements that you state, plus some data gathering as well. (CTD, water sampling, ecology object tracking/counting).

One comment, and I have many more, but for time and space, you don't want to drive the motors with the brain boards if you're ROV is of significant size. Even with 12VDC bilge pump like thrusters, given load / ocean currents, etc, will overtax the amp capabilities of the arduino, etc. (but again, depending on size of ROV) My suggestion is to use the beagle bone or arduino as the main controller unit and have it talk to high voltage/current motor controllers with a separate, dedicated 12VDC or larger on-board power supply. (but in design this adds weight, causing the need for ore powerful motors...) you can see how the design spiral goes! I suggest brush-less DC motors, oil filled housing, with the motor controller also in an oil filled pressure housing.

In any event, start with a simple prototype, but keep the end requirements in mind and build to those requirements. The texts mentioned give good examples for the types of design considerations and mathematics that you should use to calculate loads/balllast/etc. And I am certainly available to chat about design, progress, or what not. OpenRov is a great resource and the developers here are happy to assist. I'm actually building one as a testbed as well.
One last link, http://www.homebuiltrovs.com/. Steve Thone has been building homebuilt rovs for over a decade. This site is responsible for many a-new-comer to ROVs from the garage...so to speak.

Have fun!