If you’re regularly hauling your ass around the golf course or the local retirement community, you’ve probably got a golf cart. Many rely on primitive electric drivetrains with lead-acid batteries, while others rely on gasoline. But what if you combined both to build a weird hybrid? Jimbo, the man behind the Robot Cantina YouTube channel, is doing just that.
Golf carts are a funny sort of vehicle. They’re like cars, only slower and with fewer consequences. They’re also a great platform for weird builds and hacking; at least if you buy a crappy, beat-up one. Running and driving examples tend to hold their value pretty well, but those with dead batteries or mechanical issues can sometimes be had for a song.
Jimbo was recently able to secure an old 3-wheeled Cushman golf cart with a completely dead set of batteries. His plan was simple—replace the failed cells with a “petroleum battery” instead.
If you’ve never heard of a “petroleum battery,” it’s a bit of a joke on Jimbo’s part. The golf cart had a functional electric motor, it was just the batteries that were shot. Rather than replacing the heavy old lead-acid cells, Jimbo wondered if a small gasoline engine could be substituted instead. Only, rather than having it drive the wheels, he’d hook it up with an alternator. This would then generate electricity to run the original drive motor. “The generator device is something we’re calling the petroleum battery, because it operates on petroleum, and replaces the batteries,” says Jimbo.
This setup promised a couple of benefits. A gasoline engine, alternator, and some supporting electrical hardware would be much lighter than the original lead-acid batteries. It would also be quicker to refuel—much faster than spending hours hooked up to a battery charger. The drawback is that this setup is not as efficient as just using an engine. That’s because turning the engine’s mechanical energy into electricity via a generator, and then back into mechanical energy in a motor, incurs losses thanks to the laws of thermodynamics. It’s a topic we’ve covered in depth on this site. Jimbo isn’t fussed about the losses though. “We don’t care, because this is a fun project,” says Jimbo.
Overall, the idea is basically the same as a diesel-electric locomotive. In that application, the diesel engine runs a generator, with the electricity produced used to turn the traction motors that drive the train. Jimbo’s design just substitutes a gasoline engine for a diesel one.
Jimbo chose a 212cc Predator engine to serve as the power plant for this build. He’s had plenty of experience with these small gas engines, having previously swapped one into an early Honda Insight. To use the engine to generate electricity, he hooked it up to an alternator with a belt drive. When spun up by the engine, the alternator produces electricity to run the golf cart’s original drive motor. Even better, Jimbo had a neat idea for how to use the alternator to control the speed of the drive motor, too.
Normally, the alternator has a regulator which controls its voltage output by varying the voltage on its rotor windings. The regulator does this to keep the car’s electrical system and the battery at the right voltage.
In this setup, the regulator is gone, and its place, Jimbo hooked up a variable voltage supply to the alternator’s rotor windings instead. The higher the voltage passed to the alternator’s rotor windings, the stronger the magnetic field generated in the rotor, and thus the output voltage of the alternator. Since the alternator is hooked up directly to the golf cart’s drive motor, the higher its output voltage, the faster the cart goes. The variable voltage supply runs off a single 12-volt battery and doesn’t need to deliver a lot of power. It’s just there to supply power to the rotor to create a magnetic field that controls the output of the alternator.
This system avoids the need for any smart motor control electronics, or even any facility to vary the speed of the main engine. If the alternator’s rotor voltage is set to zero, it outputs nothing, and no power goes to the drive motor. Ramping up that rotor voltage essentially acts like a throttle, ramping up the alternator output and the drive motor speed in turn. The Predator engine itself is just left running at a constant speed of around 2,500 to 3,000 RPM most of the time. However, for maximum power, the engine can be throttled up to maximum to spin the alternator as fast as possible.
Turning this into a practical throttle is slightly more complex. In testing, the rotor voltage was controlled with a knob on a small variable voltage module that was sourced from Amazon. Jimbo had to hack a solution together to make it work with the Cushman’s original throttle pedal instead.
The variable voltage module effectively acts as the throttle, and was originally controlled via knob turning a 100 kiloohm potentiometer. While Jimbo could have grabbed an e-throttle pedal from just about any junkyard to replace it, this would cause a problem. Most e-throttles are rated at 2 to 5 kiloohms or so, as are most throttle position sensors. This would force him to heavily modify the circuit of the variable voltage module.
Instead, he ended up cobbling together a complex 3D-printed solution to couple the cart’s accelerator pedal to a simple 100 kiloohm linear slide potentiometer. The linear potentiometer was then spliced into the circuit of the variable voltage module in place of the original knob. It’s a messy solution, but we’re told a better one is already in the works.
For safety, there’s also an emergency stop in the loop. It’s set up to cut power to the alternator’s rotor windings when it’s pressed. The brake pedal on the cart also triggers a relay which cuts the power, too, providing multiple ways to bring a halt to proceedings if things get out of control. This is a smart call whenever you’re building your own throttle mechanisms.
With the kooky junkyard drivetrain in place, Jimbo set about determining the performance of his effectively-hybrid golf cart. The Cushman’s rear differential has a 12.7:1 ratio. Initial assumptions were that the golf cart’s motor topped out around 4,000 rpm. Calculations suggested a top speed of around 16 mph. However, in a short road test, the cart was able to hit 21 mph quite easily, suggesting there was some headroom on that figure. At that speed, the golf cart motor was running at 5,200 rpm, with the gasoline Predator engine spinning at 3,600 rpm to provide the necessary power.
Homebrew gasoline-electric golf carts are now a proven technology. The cart was able to haul two adults without complaint. None of the electronics got excessively hot or released any magic smoke, and the engine held up just fine as well. Is it more complicated than a traditional battery-powered or gas-powered golf cart? Yes. Does it work, though? Also yes.
Jimbo notes that he could get a permit to drive the rig on public roads in his local area, too. However, more work remains to be done. As it stands, the Predator engine is particularly noisy—both due to exhaust and vibrations. The throttle setup is also due for an update. The first revision is quite mechanically complicated, and the heavy return springs make it a pain to use.
This build was fundamentally a fun experiment performed on a dead golf cart that was otherwise headed for the scrap heap. To that end, it’s a great success. It’s always wonderful to see a beleaguered vehicle brought back to life, particularly in an amusingly unique way like this.
It’s a great payoff to see the thing work so go please give Robot Cantina a watch and follow if you’re into these types of builds.
Image credits: Robot Cantina via YouTube screenshot
I used a similar setup to control the PWM that looks very much like the controller he’s using on my sit inside e-kayak. I used a Minn Kota 35, which comes with a step resistor throttle and I wanted fine remote throttle control and greater efficiency, so I went with a PWM. I used a bicycle bar end shifter for the throttle control pulling a Bowden cable connected to an arm in the rear compartment like he has, but I attached the arm to a spur gear that meshes with a smaller gear on the PWM knob to multiply the the short pull distance at about a 1:3 ratio. I also used a tension spring for return to neutral in a similar fashion. His is a lot cleaner as I had quite the packaging challenge. I did it for about the same reason: just to try it out. Turns out to be a lot more fun than I thought it would be.
I’ve been watching Robot Cantina’s videos for years now. His Kubota diesel powered Honda Insight and Saturn SC1 are most excellent builds. They’re slow, but usable, and his fuel economy for both are in the 60 mpg range.
My personal favorite is the Predator V-twin 1969 Renault with both a CVT and a manual transmission with the CVT acting as the clutch! The Kubota diesel cars are awesome though, makes me want a small diesel hypermiling project of my own.
Anyway, happy to see Robot Cantina getting recognition here, such a unique and fun kind of project!
On ecomodder, someone actually built a Triumph Spitfire-based Centurion from R.Q. Riley’s plans, running the same Kubota diesel engine. He can get 100 mpg as advertised, but has to do a steady 45 mph or so and use all the hypermiling techniques in the book to achieve it. A larger, heavier, and more powerful VW TDI engine is far more thermally efficient.
Interesting! I remember seeing those Centurion plans, as well as the economy trikes like the Tri-magnum. Part of me feels like what I really want to see is the Kubota diesel in the Renault… although the aerodynamics of that car would be counterproductive towards hypermiling. It’d also be really interesting to pair that engine with a CVT, since it’s an industrial engine made to run efficiently and reliably at a constant RPM all day. Would also be curious about powering a car with a tractor engine and CVT for similar reasons.
YES! Robot Cantina is excellent! For us Autopians, he also has a Renault 10 (how rare are those in the US!!??) that had a dead engine. He rescued it and installed a Predator V-Twin….goes pretty good! He also has a Honda Insight with a Kubota 3-cyl diesel swap!
It’s not really a hybrid?
It only has one source of motivation. It’s a petroleum powered vehicle with an electric transmission. Like a train engine or mining dump truck. We don’t tend to consider those hybrids.
If you added some batteries back, then it would be a hybrid.
That’s what I was thinking. If he left in one set of batteries to make the original voltage and used the alternator rewound (or a different model completely) to charge that voltage, then it would be a series hybrid.
How very interesting. Kudos to Jimbo, even if this isn’t the most practical golf cart, it’s a fun engineering exercise. I’d be curious what’s the MPG compared to a normal gas powered cart.
I don’t know a normal cart, maybe someone else can help. But a basic predator engine can run for about 3hr/gal at 3600 RPM.
So looks like maybe 60mpg for this build.
I never get tired of people doing stuff just because, with enough effort, they can.
Bring on the Mk II version!
Predator 212 engine from Loncin via HF has become the “LS” of small engine swaps. I’ve put two of them on old go karts and they’re sweet. I noticed the new Toro snowblower I bought has one also (rebranded),
Sometimes I wonder about making a radial engine out of 5-7 barrels/heads/pistons from those guys.
Same guy put a diesel in a First Gen Honda Insight.
Lewin, dude, I hate to pedantic (okay, no I don’t) but this butchering of the word “then” has got to stop. Remember it like then = when. The phrase “rather then” is a legitimate expression, but it’s very rare that it would be appropriate. You’re almost always needing to use “rather than”.
Every time I read an article or post that uses “then” when it should be “than”, I pretty much stop reading and go on to something else. I don’t know why it bothers me so, but it does.
Here’s an example of why it matters: I’d rather eat glass than have a hot poker shoved up my nose, versus I’d rather eat glass then have a hot poker shoved up my nose.
See? There’s a difference.
I don’t know why, but in my mind I read it as “a hot poker shoved up my @ss”. That would be an even better example, haha.
Not here to kink shame
That was the original word, but I switched it to nose. Didn’t want to be too much of a, you know, @ss.