Hi All,

I am sure that many of you have heard of the OpenROV Project which involves building your own ROV from a kit. I am an experienced model builder with plenty of experience and a reasonably equipped workshop, but I was a little taken aback at the complexity of assembling the OpenROV from their kit of parts. It seemed to me to be a high school to college level project requiring a small team of experienced modelers to successfully put the OpenROV together though I could be wrong.

It appeared to me that there were a number of critical assembly requirements that had to be performed spot on as there was no going back to realign or reseal once the assembly was completed. Therefore I have been looking around to see if there might be an easier way to design and build a ROV similar to the OpenROV without having to perform some of the critical assemblies required to complete the OpenROV.

One area of the OpenROV design that I took exception to was the thrusters. They were not going to last very long when immersed in seawater due to the exposed bearings and stator/armature. To me this was a kind of deal breaker as to the questionable lifespan of these thrusters. I know that commercial thrusters are not cheap (think Seabotics or CrustCrawler) and the OpenROV project was just trying to overcome the high price of commercial thrusters with their homebrew design. Still the unknown lifespan of the OpenROV design did not leave me with a warm feeling to say the least.

Then came along Blue Robotics and their T100 Thruster KickStarter Project Link. Their design approach looks good and they have actually tested and characterized prototypes of the thrusters they will deliver to the pledgers. With the addition of the in-thruster water cooled ESC this design becomes very attractive in that it reduces the number of wire penetrations in the Water Tight Compartment (WTC).

Great! Now we have a source of reasonably priced thrusters so what is next? Well, that is where you, the ArduBoat members, come in.

Let's start thinking about the WTC, navigation controller, communication, power, ballast, buoyancy, etc. and attempt to come up with a reasonably priced ROV that the average ArduBoat member might want to consider building.



Views: 20026

Replies to This Discussion

Hi Thomas,

This is exactly why we designed our thruster in the first place. There are tons of platforms and components for UAVs but there's almost nothing for marine robotics that is affordable to hobbyists.

We'll be happy to do whatever we can to assist. Here are my initial thoughts:

WTC: I think this is one of the most challenging parts because if it fails, you could lose hundreds of dollars of electronics. I think plastic tubes, like polycarbonate or PVC, are pretty easy to work with and you can use O-rings to seal them easily. Our BlueROV has a polycarbonate WTC with an aluminum backplate and O-ring to seal it.

Navigation: The APM or Pixhawk will work great for this. My only complaint is that 3DR combines the compass and GPS unit. GPS won't work underwater so it's not needed. The air pressure sensor could possibly be used to detect leaks in the WTC.

For an ROV, you generally want live video also. A Raspberry Pi or BeagleBone provide an Ethernet connection for communication and can send live video pretty easily.

Communication: We're using an Ethernet cable because they are cheap, readily available in long lengths, and provide a high-bandwidth connection. Unfortunately, underwater Ethernet connectors are really expensive ($150+) so the cable is potted directly through the vehicle's hull. A better solution for that would be great.

Power: We considered sending power over the tether, either directly or with PoE, but a long tether with small wires just isn't great for that. It's a lot easier to put a lipo in the WTC and have 1-2 hours endurance.

Ballast & Buoyancy: There's lot of options here. Syntactic foam is usually used for added buoyancy.

User Interface: I think the user interface is really important considering that it's really your only source of information once the ROV is in the water. OpenROV has a pretty cool one. I'm not sure how easy it would be to use with other ROV designs. We'd like need a MAVLink-esque communication standard.

I'll try to post more about our BlueROV and SolarSurfer in the coming week or two.


- Rusty

Hi Rusty,

You have identified some great discussion topics and I hope other ArduBoat members will chime in with their observations and interests.

As far as the User Interface is concerned, I purchased several of these inexpensive, compact modules to conduct bidirectional communication and control between my ROV and ROV control console: ROV Control System



Hi All,

I have some thoughts concerning the following ROV characteristics:

Power: There are two ways to go with providing power to the ROV: local power and remote power.

Installing onboard batteries simplifies powering the ROV, however they add to the weight of the ROV further complicating the buoyancy stability issue.

Using an onshore power supply has the drawback of requiring additional wires in the tether cable to carry the power to the ROV. Some builders have gotten around the heavy power cable issue by using 48vdc onshore power and converting the 48vdc back to 12vdc in the ROV. The use of the higher voltage allows the use of smaller power cables to transmit the same amount of power that a lower voltage higher current (bigger power cables) can transmit. However there is the issue of the power dissipation of the step down voltage converters at the ROV end that will add to the ROV thermal load.

WTC:  Many builders use pressure rated PVC pipe in various diameters with 6 inches appearing to be a popular diameter. The PVC pipe wall is usually 1/4" in thickness and the ends are usually sealed with flanges and O rings or acrylic windows. The 6 inch diameter of the tube allows plenty of room for a slide in tray to hold the navigation controller and other electronic accessories.



Hi Thomas,

There are thicker PVC tubes (usually called Sched 80) which we have used as a WTC up to 100 metres water depth.  Deeper usually requires Aluminum tubing, but there are many other new materials that can be used. For depths greater than 100 metres we usually use double o-ring seals and we install the ability to evacuate the WTC slightly so that the seals are compressed. 

In order to determine if there is a leak we use an "Aspirin" switch, which is a microswitch held closed by an aspirin.  The slightest presence of water will cause the aspirin to disinitegrate opening the switch.  The switch can control a master relay or just turn on an LED in camera view, (or switch the APM into surfacing mode).

For power, I have used on-board power, high voltage DC (600VDC), and regular 12V to 120 VDC supplies from the surface.  We had to custom build 600VDC to 12 or 24 VDC DC-DC converters, but that proved to be the best way for high power requirements.  We used DC because the induction loss in water is so high compared to DC, combined with the cable resistance loss.

Another approach would be to use surface power at low current rates to charge one (or 2) sets of batteries while another set is running the vehicle.

The GPS/Compass package would be useful if we use an INS for Navigation, (position updated by depth, Doppler etc) When at surface the position of the ROV on the surface is recorded, the INS data etc logged and upon return to the surface the position is taken again,  Post processing can be used to distribute errors in position throughout the mission.

The one thing to remember if an autonomous ROV is anticipated is that control of the motion has to use track, not the compass heading.

Ethernet connectors are expensive because the distance from cable to cable through the connector has to be as short as possible.  Fibre optic is the better way to go.  The downside being the cost of sliprings when using a winch for the cable. There are some tricks however if using a single mode cable to make a low low cost slipring.

Control functions can be transmitted from the RC transmitter to the receiver in an ROV using coaxial cable.  One of the first ROV's, called TROV had on-board batteries and had 2 coax cables.  One for the control, the other for the video signal to the surface. Then, the OSD was a board about 9 inches by 4 inches compared to the minOSD , <1 inch by <1.5 inches.


Ken McMillan (just joined the group)

Hi Ken,

Welcome to the Group!

You have provided some great observations and pertinent information concerning ROV construction.

Much appreciated.



Just new here, but not new to the world of ROVs.

MATE is a program where a wide range of students design, build and compete at regional and international competitions.

I have been working with this program for over 5 years now in various capacities. It really works to get students interested and focused in Science, Technology Engineering and Mathematics (STEM).

These students never cease to amaze me on what they come up with to solve increasingly complex problems the competition comes up with.

If you are interested check out www.marinetech.org

This years international competition is being held in St. John's Newfoundland at Memorial University and the Marine Institute. It will be streamed live through the marinetech.org web site with streaming facility provided by robotics.nasa.gov


Hi Patrick,

Welcome aboard so to speak:-)

I have followed the MATE competition in Robot Magazine and I too am amazed at the ingenuity and creativeness of the teams involved in that competition.



Good success with very long power delivery using Variable frequency 3 phase AC with active feedback.

A small onboard battery for surge loads that is recharged when the power load is low. (Capacitors can be used as well but a battery has a higher power density. (LiPo) The key here is the active control of the 3phase voltage and control of the power factor. Changing frequency can also help to mitigate Xc loss in water and help to reduce noise that can interfere with other signals. Also using ungrounded three phase is safer if one phase becomes shorted to water...

Think Brushless motor speed controller and a 3phase transformer and rectifier! (all off the shelf parts)!

I have had fun with this one and some good results powering small ROVs over very long, very small tether.

Remember Harmonics can be your best friend or your worst enemy...


Hi Patrick,

That is an exceptionally novel use of a brushless motor ESC for a DC to AC converter.

I assume that you set a specific PWM frequency for the most efficiency and lowest power factor for a given setup?

I know that MATE allows up to 48vdc on the ROV umbilical to help reduce the size of the cable that transmits the power to the ROV.




I agree with Tom, that's a really clever use of a brushless controller. I imagine it would be possible to use one on the vehicle side as an active rectifier?

Why did you choose 3-phase AC over high voltage DC? So that you can use a transformer?



Correct MATE Explorer class is allowed 48Vdc with current limit for power through the tether.

Ranger is limited to 12Vdc.

I have been working on Matt to allow AC in the tether circuit but limit it to 36v and apply a current limit as well, also require it to be ungrounded 3 phase. This would allow for very small tether. Most MATE explorer teams run the 48 to the ROV then use DC-DC buck converters to knock the voltage down for the various needs. This is not very efficient and requires additional hardware that needs to be in a pressure housing, cooled and can produce electrical noise that can effect the processors and other control and sensor circuits in the pressure can...

As for the PWM frequency that is what makes it a good solution! Very the frequency will allow to adjust around interference and find a good working frequency. Another nice thing is the ability to program the power curves in the ESC. On the down end (Subsea) use a 3 phase rectifier and a series of 1 chip regulators to get the voltages that are required. Keep in mind that there has to be feedback from the down side end that can control the "A-D" input of the ESC to keep the voltage at the subsea end constant as the load changes. A simple inexpensive PIC processor does that very well.

I have been working with Wild AC devices for many years as a hobby related to alternative energy production. DC transmission is expensive and needs to be heavy duty. AC is much easier to deal with, transmission is much more efficient.

I work on the  Large Work Class ROVs, 3P/ 3Kv in the range of 200Hp is very common. Oh did I say the Tether is 4.6cm  Dia. and is up to 4Km long. The same can be done for very small ROVs and the Model parts are there right off the shelf to do it.

Oh and a bit of Trivia:

On average 1 Cu Ft of:

Salt water weighs ~64.1 Lb.

Fresh water is ~ 62.4 Lb.

So a very basic, back of the match book buoyance equation is:

{Displaced volume (Cu Ft) * weight of Cu Ft of Water) - Weight of the object displacing the water = Buoyance.


If you go with active feedback then all you need is a 3P rectifier and a filter cap to convert back to DC. As the voltage drops the feedback loop "speeds up" the ESC and the Subsea output goes up. Very simple feedback. Converse when the DC voltage is above a set level the ESC throttle is backed off and the voltage drops. The DC voltage will very within limits of the sample rate and limits set in the feedback loop. But if you have the need for several DC voltages to run things subsea like 3.3, 5, 12, 36. Single chip regulators are available to bust down and Isolate the delivered DC. If you want to get really fancy you could use a 3p Multi drop output transformer and break it down that way. But that means you have to have the extra weight of the transformer subsea.

Besides a car alternator rectifier is a easy to get, off the shelf 3P Rectifier.




Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Vertical Horizontal one, is here

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service