The Manta Drive: Proof-of-concept for an ROV Propulsion System.




About: Find me on Reddit, Tumblr and Twitter as @KitemanX. Buy my projects at

Every submersible vehicle has weaknesses. Everything that pierces the hull (door, cable) is a potential leak, and if something must both pierce the hull and move at the same time, the potential for leakage is multiplied.

This Instructable outlines a drive system that eliminates the need for drive-shafts to pierce the hull of an ROV ("Remotely Operated Vehicle" - a robot submarine controlled via wire), and also removes the very real possibility of rotating impellers getting tangled or jammed by underwater plants or hanging lines.

It could also give rise to vehicles that have a much less damaging effect on the habitats they are used to investigate, because of the lack of a "wash", and because the lack of rotating impellers will reduce the risk of injuring animals the Manta Drive encounters.

Step 1: The Concept.

The whole idea of the Manta Drive was inspired by a visit to an aquarium where members of the public got a chance to pilot small ROVs round an obstacle course. I got my first look at the ROVs and realised two things:

  • There were a lot of places for the water to get to the insides of the ROVs
  • The ROVs didn't look right - they were just boxes, and didn't look designed to swim. They lacked the elegance I associate with swimming animals.

Later cogitation also came to consider power - the high-revolution impellers used by the ROVs struck me as power-hungry. I may be wrong, and I have not tested the power consumption of the Manta Drive, but this is a secondary consideration.

As I wandered the aquarium, the ROVs played on my mind, and I found myself comparing them to each animal I saw. How did they compare? Could the animal's swimming motion be replicated elegantly, in a way that maintained hull integrity*?

Looking at fish like rays, sea cucumbers and stonefish, I realised that the most elegant propulsion method was the waving fin.

I also realised something important - fish don't leak. A rotating shaft needs to pierce the hull completely, working through a hole in the hull. On the other hand, a reciprocating motion (up-and-down) could work through a flexible, waterproof membrane which did could be fixed firmly around any moving parts without ripping.

I further realised that flexible membranes could wear out, but magnets don't, and magnets can act through any non-magnetic materials without restriction. Make the hull rigid, but non-magnetic, and the risk of leaks due to the drive system are completely eliminated.

* Oh, I went all Star Trek for a second there!

Step 2: Materials and Tools

All I actually bought for this project were the magnets - small, countersunk neodymium magnets from ebay.

The rest was made of material I already had stocked in my shed - scrap timber, bamboo skewers and a pair of dead ball-point pens.

Similarly, no specialist tools were required - a junior hacksaw with blades for wood and metal, a hot-glue gun, drill and my multi-tool.

Health and Safety

You'll be using hot things, sharp things and very whizzy things. Be careful.
Take particular care with neodymium magnets - they can nip painfully, and will shatter if allowed to fly together.

Step 3: The Frames

I cut two empty ball-point pens into five roughly-equal lengths each - three to take the manta's ribs, two to space them out.

The frame itself is made of three lengths cut from scrap timber - the base is about 10cm long, the end sections are about 3cm long and drilled near the top, using a twist-bit the same diameter as the bamboo skewers.

I hot-glued the timber together, then threaded bamboo through the holes and the pieces of pen.

Step 4: The Ribs

The actual propulsion of the Manta Drive is carried by simple ribs. These are coupled to the drive mechanism by the magnets.


I threaded bamboo skewers into the holes of the magnets and hot-glued them in place, then glued the bamboo to three of the pieces of pen on the frame.

Step 5: The Actual Drive

The ribs are connected, via magnetic forces, to the drive mechanism.

In a finished ROV, the internal magnets would probably be moved by motors or servos. In this model, I just used more levers, shortened versions of the ribs.

Step 6: Connection and Drive

The drive isn't intended for the magnets to be in direct contact, and it defeats the object anyway.

In the final ROV, there will be a non-magnetic hull between the ribs and the drive. Non-magnetic air does the same thing, so all I needed was a set of spacers to hold the two sets of magnets apart. More scrap timber (6cm long, if you're interested), with pieces of bamboo to stop it slipping to one side.

Step 7: Working the Model

Operation is, in principle, very simple: when the levers move inside the ROV, the spines move on the outside.

The trick is to move the ribs in a useful sequence.

In this video, I made a simple "bracket" from more bamboo, slipped it over the drive levers and used it to move the levers in a basic wave sequence.

In the final ROV, the levers would be moved simply by a cam-shaft driven by a single motor. For more control, allowing "waves" of different length and frequency, each lever could be moved individually by a micro-processor-controlled servo-motor.

Step 8: Future Steps

Obviously, the model as presented in step 7 will drive nothing.

A finished ROV will have a row of ribs down each side of the hull, significantly more ribs than three. Between the ribs, the ROV will have either a single membrane, so that ripples in the membrane will provide the propulsive force.

Reversing the direction of the wave reverses thrust.

I intend this Instructable to be freely available for others to use to build their own ROVs far more cheaply than the professional devices currently available. Using the magnetic-coupled drive, the hull could be easy to source, and easy to make water-tight.

I imagine that it would work nicely with a length of large-diameter plastic sewer pipe as the hull. Matching compression-fittings can easily close off the ends of the pipe. Modifications to allow a camera to see out, or a control cable to pass in can be made watertight very easily, because they won't need to allow motion.

For actual use, ROVs powered by the Manta Drive will, I expect, be mainly hobby vehicles, used to explore the mysteries of the local swimming pool or canal. However, I would hope that the drive could be taken up by "serious" researchers, as it could be used to make ROVs more stealthy - with a hull suitably shaped and coloured, a Manta Drive ROV could be disguised as a large stonefish, or even an actual Manta ray. This would allow them to interact with living fish more naturally, in a similar way to the BBC's Roboshark or Draper Laboratory's Robot Tuna, but with fewer technological hurdles to leap (and much more cheaply!)

Instructables and RoboGames Robot Contest

Second Prize in the
Instructables and RoboGames Robot Contest



    • Arduino Contest 2019

      Arduino Contest 2019
    • Trash to Treasure

      Trash to Treasure
    • Tape Contest

      Tape Contest

    147 Discussions


    1 year ago

    did you ever finish it? or has anyone else taken up this mantel? it is such a simple, yet increadible idea. The way that you laid it out in this ible, makes sense even to me.

    3 replies

    Reply 1 year ago

    Wow, flashback comment!

    No, I never did take this any further - at the time, I lacked the skill / resources / money to build a real ROV, and I guess I just sort of forgot about it.

    Please, feel free to use the idea in your own projects!


    Reply 1 year ago

    I will be sure to share this with my students especially as we are building a new ROV/Robotics program. If anyone ever grabs onto your idea and runs with it, I will have them credit you.

    I on the other hand see many great costuming potentials with it... must play...


    Reply 6 years ago on Introduction

    The video should be fine - it's working as of this message's time-stamp, and there was no reason for it to be down.


    Reply 6 years ago on Introduction

    I mean the first embedded video on step1, according to YouTube, the user closed their account.


    5 years ago on Step 8

    Anechoic rubber-clad foam forms a decent "skin" for this type of underwater propulsion. With a long enough transit between each of what you termed "ribs" you need only limited flexibility in line with the ribs, rather than a fully elastic membrane as a whole. You can also drive this with electromagnets above/below the permanent magnet at the end of each rib to eliminate moving parts in the body of the vessel and allow more asymmetric variations in movement between each side and in different operational modes. I tried a similar design and then used Nitinol SMA wire on either side to draw each spar upwards and downwards in sequence, reducing the external EM signature of the ROV.


    5 years ago on Step 6

    One thing though,If the magnets are too strong, they could interfere with the motors.
    AMAZING design though.


    6 years ago on Step 8

    I love this idea! I think, however, that only the front rib needs to be driven, and the others can be free-moving. Perhaps the rear rib (and the front) can be engaged or released, allowing reversal of thrust. Also, the magnetic force will decrease as the ribs reach e tops and bottoms of their paths. Perhaps a straight sided vessel with sliding magnets would reduce this problem (a similar mechanism on the outside would reduce it further).

    I might try building something like this. A vertical fin in the rear (and perhaps front) could be driven the same way and make steering easy.

    1 reply

    Did you think about the resistance of water? Magnet system could be unefficent under water. But if you use 2 crank shafts -one for both sides- you can easily control your ROV's movements ,for example turning, just by reversing one of the motors or making it move faster than the other one. Great instructable though!

    3 replies

    Reply 6 years ago on Introduction

    Water is non-magnetic, and so does not affect the efficiency of the system in that way.

    If you look at step 8, at the bottom-right part of my sketches, you'll see I already thought of the two-crankshaft idea.

    Thanks, by the way.


    Reply 6 years ago on Introduction

    I was talking about the resistance force of the water. I mean, you are testing the system in air and there is much less resistance in air than it is in water. Are your magnets powerful enough to beat that force?


    Reply 6 years ago on Introduction


    As far as I know, it should be OK as long as things go steadily - too quick, and the magnets could get jolted out of synch.

    The Cartographer

    6 years ago on Step 7

    This is such a good idea, if I understand it correctly you'd have 3 different part propelling the submersible forward and because of the design they'd be each helping to move the other parts within. And because it's using magnets and not gears in would be easier to stop one side physically to turn and such if it where being used as a manned submersible.


    7 years ago on Introduction

    I love this project, although to simplify the drive mechanism I think having the crankshaft outside the hull would be easier to build.