Brushless Motors


Many people (special thanks to elsinidentidad) have pointed out the potential advantages of using brushless motors for ROV thrusters as opposed to standard DC motors. Some of these advantages include

  • Minimal moving parts
  • Longer lifetime
  • Completely sealed electronics
  • Lower electrical noise
  • Greater efficiency
  • Greater controlability

This has been a topic of interest for quite a while, but there have been several hurtles with just going straight to using consumer-off-the-shelf (COTS) brushless motors for RC planes. Some of these concerns are

  • COTS brushless motors have higher speed and lower torque then the ideal specs for underwater vehicles (and gearboxes would negate many of the advantages of going brushless in the first place)
  • Most COTS brushless motors are not sealed to the point that they could be put in salt water out of the box so extra potting would likely be necessary, however the very small distance between the stators and magnets prohibits the addition of supplementary potting between the two
  • Most COTS brushless motors use bearings which would not last long in water

Many hobbyists build their own brushless motors, and in fact, there's quite a bit of information out there on just how to do it. The OpenROV community could really use a solid brushless motor design (either built from scratch or adapted from a COTS part) that is optimized for deep sea ROVs


I spent the last few days searching for a journal I started with a design I've been working on (off and on) for the last few months for a custom brushless motor for ROVs. The binder has much more detailed drawings then what I've posted below, but since I still can't find it, this will suffice.

I call this device a "Brushless Ring Thruster: The way it works is that a three phase signal controls 12 coils (four sets of three in series) configured as shown around a circular cavity which encapsulates a ring with 6 permanent magnets. Attached to the inside of this ring are six propeller blades which face inward. As the phases change, torque is imparted on the permanent magnets by the coils so the ring rotates and thrust is produced by the propeller blades.

Brushless Ring Thruster conceptual design drawing. Note that this design is released under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License. Please contribute, but share the love!

Here are some of the advantages of the Brushless Ring Thruster

  • Stators and magnets at high radius results in higher torque and lower RPM- ideal for underwater use
  • No center axle means there is nothing for debris to get caught around
  • Thin, versatile form factor
  • Relatively simple and low cost to manufacture once a process is developed
  • Inherent prevention of tip-vortices -->> greater efficency

I haven't built this design yet, as I anticipate it to be a >60 hour effort to develop the first working prototype, and there would likely be many iterations following the first build (I just don't have the time). Also (although I've developed a laser cutter-only version of the design) this would be best manufactured with an injection molding machine (which I have access to) but can not afford to make the die for. My intention has been to eventually fully develop the Brushless Ring Thruster and incorporate it into the OpenROV design (the motor box was designed with this in mind). If anyone else feels like taking this project on, it would be great to have access to this technology!


I'm new here but I know a bit about RC aircraft, software development and scuba diving. What is the ideal target rpm range for a thruster motor, no-load?


The sort answer is that it depends. The prop size and geometry greatly effect the motor sizing and vise-verse, but I generally picture rotational speeds on the order of a thousand or so RPM in water, so probably the low ones of thousands with no resistance. Again though, it can change greatly depending on the conditions.

For props that are meant to be used in air, I usually fudge it and assume that they should spin at about 1/15th the speed in water to match Raynolds numbers.



this is a very smart design. can't wait to see how it's taking shape.


I love this concept and though about developing a thruster liek this a few year back, but i found Southapton university had done this and i think patented the idea and created a company to take this concept to market.


I hope this helps, i gave up a went with a magnetic company as it was simpler and didn't overlap with there work.


I guess that confirms that it's a good concept.

About the patent, worth to look at what exactly they patented. Maybe it's still possible to sell parts of this or make it 3D printable.


Defiantly a good idea, i still like the use of a of the shelf brush-less motor, to keep the project simple. but i will soon be trying some 3d printer stuff for my self now i have mine up and running. just need to find some money to get the ideas moving. Time to search down the back of the sofa!


Found a few images of some design's i worked on

A Bressless DC motor with the magnets embedded into the Hub of the Propeller

An Exploded view of a rim driven thruster



in the design concept of the OP and your last image, would it be possible to use the impeller without any contact to the motor and without any bearing, just by staying centered through magnetic force? this would make it possible to seal the electric parts in epoxy and avoid any bearings that may rust or expensive ceramic versions.


It is possible. We've prototyped the magnetically coupled drive your describing. It seems like it would work fine, but we want to really put the exposed DC brushless motors to the test. For cost reasons...

Short answer: yes.


I was just thinking about wanting an off the shelf solution for a rim driven drive system, could a stepper motor provide the outer ring of slators and inderting a new set of perminant magnets with a internal prop? if a stepper motor with 12(6,9,12,15 ect would work) coil could be found a standard RC motor controller might work to drive it :) does that make sense?

Im tempted to try it but most data sheets dont let you know how many coils they have. would it need to be a 6 wire stepper?



What about brushless computer fans with sleeves instead of bearings? I'd suspect torque would be an issue. I don't know about under water anything - what sorts of RPM and tip velocities would you consider maximum? what sorts of torque and/or power envelopes are you working with?



As i read of efficiency in Diver Propulsion Vehicule (DPV) rpm around 7-850 are the most efficient in water.

Brushless are generaly way to fast for this purpose.


ok, I dropped a small 80x80mm computer fan into the tub. specs: "Aspire Model DF1208SM 2600 RPM 28 CFM 12V 0.25A" It ran really slow (not surprising) eyeballing, probably about 3-4 rotations per second, or about 180-240 RPM. I think that by definition the 12V 0.25A is clearly not powerful enough; not enough torque to push water around.

I never got around to doing any real measurements (I do have a tach) because it died in less than an hour. it came apart pretty easy. after removing the !plastic! c-clip, the fan (and attached magnets) came right out no problem. poking around the (unprotected) circuit board, there was a tiny bit of corrosion, until I came to this one 4-lead transistor-like device labeled "277 oB4" (note the odd capitalization). the one lead connected to +v simply disintegrated into dust; it was completely gone from board to chip; no way I could reconnect or salvage anything.

The fan itself is awfully simple. there appear to be 4 coils/armatures, a circular ceramic magnet attached to the fan blades that rotate around them, and this one transistor-like device and two 3.3uF electrolytic caps.

I'm going to try with another larger more powerful fan, and see if this one was just a fluke. While I now realize that computer fans are probably not viable (not enough torque, non-reversible, and frequently not easy to dis- & re-assemble if you wanted to waterproof (conformal coating) the electronics.


Ok, after disassembling another fan and reverse engineering the circuitry a bit, a few more thoughts:

No way that leaving a circuit board unprotected is going to work. It's performing electrolysis of the water, with the component leads acting as electrodes. of course they'd disintegrate (corrode) in water.

Motor torque is proportional to the number of poles. The two fans that I've disassembled so far are 2-pole motors. that is there are basically two windings both with one side wired to ground (common) and the other two wires labeled L1, and L2 (L=inductor). There are four wound electromagnets, and it appears that they're alternately wound in opposite directions, i.e. NSNS facing outward. I haven't yet run a fan on a bench yet to see how they're powered, but I suspect that they're just powered t=0: winding L1 turned on only, t=1 L2 turned on only, t=2 L1 turned on only... only one winding at a time. I need to get one on the bench to figure out how a 3 ohm coil isn't pulling more than the rated 1.34A (12V / 3ohm = 4A).

Conclusion: Computer fans as direct propulsion units probably isn't going to work. the needed torque just isn't present. but this may help provide some guidance to choosing a n-pole brushless motor.


Just ran across this:


Eric this is really cool, I'd like to prototype something like this if I find the time. Do you have optimization calcs for the prop and windings?


We are now building the motors described on this page, primarily for higher-speed applications. The motors are ironless, therefore are thermally limited - there is virtually no inductance to limit the torque, so long as the coils can be cooled.

The motor work perfectly underwater (see video at and can be wound for higher torque at lower speeds..

Our website is Commercial sales will begin in approximately 180 days to legitimate OEM customers.

A typical rotor is shown here, with a "spider" installed for testing. Wall thickness for a 4kW motor @ 7,000 rpm is approximately 3/8", and weight about 2.5 lbs.


Really old bump, bud did you ever get this thing up and running?

I am looking for something like this. Options for small ROV’s are very limited.