Olly Epsom explains how (and why!) he built his own 1kW electric outboard motor for under £600
It will almost always be possible to buy a petrol outboard for less. So why did I do it? Well, my motivations were as follows:
1. To keep myself occupied at home, alone, over the long nights of a Scottish winter (over previous winters I’ve constructed a navigation system, an autopilot and solar water heaters).
2. Because I wanted a system that was more powerful than the cheap trolling motors, but much cheaper than equivalent commercial units (an equivalent ePropulsion Spirit currently retails at around £1,500)
3. Because I have longer-term aims of making the system fully remote controlled (essentially a drone, perhaps for wildlife photography), and for that I needed complete mastery of the control system, which is achievable if I build it myself.
There may be other reasons to embark on such a project, and if so, fill your boots.
Preparation and design
To embark on a project like this requires a basic knowledge of electronics and mechanics, but should be within reach of most practical people with a willingness to learn.
The voltages are low enough that they should not present an electrocution risk but take all appropriate precautions when working with electricity and moving mechanical parts.
My project utilised computer control and a lot of new-to-me equipment such as touch screens, Bluetooth, Arduinos and oscilloscopes.
I also had to learn how to code the software.
With the learning and setbacks the project took me about six weeks of evenings.
You may wish to structure your project differently – a significant simplification could be made by discarding computer control.
I had the rough design in my head after an evening of research online.
I decided to use the bottom end of a conventional, second-hand outboard motor as the basis of the machine.
Above this, in place of the engine would be an electric motor, directly coupled to the driveshaft. Electronic control systems would be mounted alongside the motor.
The battery would be separately mounted and connected via flexible electric cable.
I settled on a power of about 1kW because it is significantly more than two rowers could achieve and would result in either a fast dinghy, or a dinghy with significant reserve power.
Batteries cannot be fully discharged without damage, so to power this for a minimum of 40 minutes at full power would require roughly 1kWh of batteries.
The motor, control systems and batteries are the main components, but figure on another £100-£200 to cover enclosures, transition pieces, wire, switches, connectors, and, of course, things going up in smoke.
This will bring the total budget to around £500-£600.
Plus weeks of your time. Suddenly rowing may seem less of a hardship!
But I am single (hard to believe I know!) and live alone and have nothing better to do with my time and money so onwards I went.
I located a 1974 Johnson 2hp outboard on ebay for £24. It cost me more in fuel to collect it.
I removed everything north of the driveshaft housing and re-sold those components for £5.
These old engines had very high specification mechanical components and the gearbox in particular was in excellent condition.
That said, the motor was forward-speed only, with no clutch.
It is possible to make an electric motor go backwards as well as forwards but be cautious doing this on an engine which originally had no reverse because the gears won’t be designed to operate backwards and there won’t be a thrust bearing to keep cogs in check.
I decided on a brushless DC motor because they are generally reliable and don’t require maintenance of brushes.
Usually they are slightly more efficient than permanent magnet DC motors as well, although in practice this depends entirely on the specific motors being compared.
Most motors in this power range run at 48V or above. While it’s possible to get 12V motors in this power range, they would draw over 90A at full power, requiring unacceptably heavy wires.
In addition there was a far wider range of motors available at 48V.
Likewise battery packs at 48V were more numerous. I settled on lithium-ion batteries due to their high capacity, low weight and high cycling ability.
A 1kWh lithium battery is roughly 20Ah at 48V – about the same amount of energy held in an 80Ah car battery at a quarter of the weight.
The downside to lithium-ion is the cost. There are various battery packs on ebay and Alibaba.
All have a built-in charge controller, and all require an external mains charger.
Batteries of this capacity range from £200-300 depending on vendor, delivery charges and taxes. The vendor also sold me a ‘matched’ charger which subsequently caught fire.
It would be very nice to have a high specification Parvalux or Maxim motor, but these things come at a significant cost, so I settled for a Chinese-made MAC12500-3A from MacMotor.
This is a 48V, sensored brushless DC motor with a top speed of around 4,000rpm and a power of 1kW, and it seemed like a reasonable match with the donor outboard’s original specification delivering 1.49kW at a top speed of around 4,000rpm.
MacMotor also sent me a speed controller, although that later proved to be less than spectacular and I had to replace it with a larger model.
The motor was around £100, the speed controller another £40.
Arduinos are tiny programmable computers with analogue and digital inputs and outputs. They run on 5V, consume an infinitesimal amount of power, and are of an inherently rugged, tough and reliable design.
They are programmed from a PC using a language derived from C (you’ll need to know a bit about programming to know what that means). They come in various models including the Nano that costs just £3.
While simple, they are more than powerful enough for a task such as this.
As a display I initially intended to use a cheap, rugged, reliable two-line, 16-character LCD display of the sort that are well known to work with Arduinos.
I then noticed the Nextion range of graphical touch screen displays, which retail at around £16 for a 3.2in version.
Whilst not as rugged or tough I was seduced by their sexy colour screens, with the ability to display multiple dials, graphs and buttons.
Finally I located a cheap electric bike twist-grip throttle for around £10 on ebay.
The one I bought came with a battery voltage indicator and miscellaneous switch that I figured might be useful.
The project was constructed in my garage with hand tools.
That said, the better equipped your workshop the better the finished product is going to be and, in addition, costs will reduce because I had to have a few parts manufactured that could be made at home in a good enough workshop.
At the bare minimum you’ll need a good toolkit, hacksaw, files, drills, power/battery drill, lighting, workbench, soldering iron and good multimeter.
An oscilloscope is also useful, I used a PicoScope USB-type that plugs into my laptop.
The first hurdle was to mechanically connect the electric motor to the donor outboard engine. This connection needs to be secure and accurate.
I measured up all mounting holes on top of the engine, and the mounting holes on the electric motor, and sketched up a suitable transition plate on the free online CAD software Onshape.
I then sent this to a laser cutting company, who cut it out of 10mm thick aluminium for around £25. I could have sent it for 3D printing out of plastic.
To connect the motor to the driveshaft I asked a friend with a lathe to turn down the end of the approx 11mm spline and put an 8mm thread on it.
I did consider collet-type connections or finding a matching female spline but the lathe option was easiest for me.
The main disadvantage of this, however, is that every time the motor is removed it brings the driveshaft with it, which makes reassembly quite a faff.
The lathed spline screwed nicely into the 8mm female thread on the motor shaft, and a lock-nut completed the union.
At this point the motor, donor engine and driveshaft are all connected, so to avoid accidents with a fast-spinning prop it’s a good idea to remove the propeller and not replace it until you’re going to use the motor in water.
You’re now able to electrically connect the components and conduct a simple test.
1. Connect the three thick phase wires from the motor to the appropriate colours on the motor controller. These are usually coloured green, yellow and blue.
2. Connect the sensor plug on the controller to the sensor socket on the motor. This will have at least five wires. Usually red and black (5V and ground) and green, yellow and blue (which take the signals from the three hall sensors in the motor). Hopefully plug/socket from the hall sensors on the motor will match the plug/socket on the controller hall sensor line, otherwise go figure it out.
3. Connect the red wire on the throttle to 5V (usually available on one of the numerous outputs of the motor controller), the black wire on the throttle controller to ground on the motor controller, and the third wire (could be any colour) to the speed input of the controller, which could also be any colour.
Finding which wire does what in a manual-free wasteland is very challenging. Some controllers may not have a 5V supply for the throttle and you may need to provide this separately via a 48V-5V DC converter.
This should not daunt you as they are easily available. In addition most controllers have a reverse function which requires connection of a switch. One such arrangement is shown in the schematic diagram (above) of the simplest possible control system.
Now is the moment of truth. Connect the battery positive and negative to the motor controller positive and negative, and give the throttle a twist.
At this point, assuming the motor turns, you are essentially home and dry. All that remains is as many weeks of refinement as you care to deploy.
And the inevitable problems that emerge, some of which I’ll detail in my next article.
If it doesn’t go check your connections and wiring. These motors hardly ever fail, so if it’s not working it’s most likely the controller and the parts connected to it that are the culprits.
A word of warning: neither the motor, controller, Nextion display, charger or battery came with a shred of documentation. I used a multimeter and figured it out for myself.
The Nextion display in particular was impenetrable as they hadn’t been around long enough for a comprehensive internet community to form – this is changing for the better.
About the author
Olly Epsom is a chartered engineer specialising in renewable energy, particularly wave power. He bought Radioactivity, a UFO 27 which he cruises and races, in 2015. Based in Scotland, Olly enjoys the outdoors, promoting science and engineering at local schools, and running his own consultancy company, Moose Pyll Engineering.
- Not only is Onshape free (www.onshape.com/products/free), but its drawings are public and PBO readers can download my designs directly.
Part 1: the basics – as published in the December 2018 issue of PBO.
Watch this space for part 2: programming the motor’s electronic controls.