Fail Of The Week: Alternator Powered Electric Go-Kart Doesn’t Go

What do you give a six-year-old who loves going fast but doesn’t like loud noises? Convert a gas go-kart to electric of course! (Video, embedded below.) That goal started [Robert Dunn] of Aging Wheels down a long path toward a go-kart that almost, but doesn’t quite… work.

If you’ve watched any of [Robert’s] videos, you know he doesn’t take the easy path. The man owns a Trabant and Reliant Robin after all. Rather than buy a battery pack, he built his own 5S24P pack from individual LiFePO4 cells. Those cells generally are spot welded, so [Robert] built an Arduino-controlled heirloom-quality spot welder. Now while the welder could handle thin nickel strips, it wasn’t up the task of welding high current nickel-plated copper. When attempts at a solution failed, [Robert] built a system of clamped copper bus bars to handle the high current connections for the batteries.

If batteries weren’t hard enough, [Robert] also decided he wasn’t going to use an off-the-shelf motor for this project. He converted a car alternator to operate as a brushless motor. We’ve covered projects using this sort of conversion before. Our own [Jenny List] even wrote a tutorial on it. [Robert] unfortunately has had no end of trouble with his build.

In the first video, [Robert] works with a common no-name motor controller and tries to run the motor sensorless. Like others before him, he found that the motor would run great on the bench, but wouldn’t start under load. Give the go-kart a rolling start, and it would run. This is a common with sensorless motors — the motor controller uses unpowered coils as sensors, and can sometimes have trouble starting a motor in low speed/high torque situations. The solution is to add hall effect sensors to the motor. [Robert] did all that, tested all the 36 combinations of hall effect and coil wiring, and it STILL didn’t work.

Fast forward to [Robert’s] second video on this project. He switched over to a better-documented motor controller, built a great-looking battery box (twice), and put everything back together. The motor worked great on the bench and would start just fine. But any time the kart hit half throttle, the motor controller would cut out with a “hall sensor error”.

It seems like [Robert] built his motor correctly – the hall sensors are mounted in the middle of the coils much like [austiwawa] discovered in his e-bike build. While not shown in the YouTube video, [Robert] even tried isolating and shielding his hall wires, but nothing changed. At this point, [Robert] is taking a step back from the project for a few weeks, or at least until his six-year-old convinces him to fix it.

We have to hand it to [Robert] for not only posting his successes, but also his failures. The go-kart and spot welder are both examples of great workmanship. We’re sure with a bit more tinkering he’ll get his alternator-converted motor to work.  Have an idea what the problem is? Let us know in the comments!

21 thoughts on “Fail Of The Week: Alternator Powered Electric Go-Kart Doesn’t Go

  1. I think he messed up in the installation of his Hall sensors.

    By placing them equidistant around internal circumference of the stator, they will all trigger on the same phase–which you don’t want. Evidence for this error is his own video showing all the sensors being triggered in unison.

    I submit the sensors should be staggered–with a phase offset between each–such that one fires…then the next…then the next… as the rotor is turned.

    His only challenge after that wouldl be to properly correlate each drive phase (power lead) with the appropriate sensor. If the controller isn’t smart enough through firmware to figure this out for him, he should be able to arrive at the correct relationship in <=6 attempts.

    1. in first video at second attempt you can see how he relocate the sensors to the correct position. at least the leds turn on in the correct sequence. what I don’t understand is how he decide the possible combinations are 32 instead of 6 like you comment correctly.

      1. Me neither. 6 possible alternator lead combinations and 6 possible Hall sensor combinations, make 36 combinations.
        But, the only thing that matters is the Hall sensor positions /relative to/ the stator lead combinations, which reduces to 6 (three “forward” and three “backward”) combinations, and I can’t see a smart motor controller like this not finding out the difference between those by some sensing (even the dumb one seemed to do that)

    2. I did the same thing with the alternator in a 1500 watt brushless motor controller on a Schwinn Stealth 1000 scooter and at first when I would get going really fast or it would get under any stress of the motor like up a hill or in soft sand it would start cutting in and out and making a high-pitched screeching noise so what I did was I isolated the brushes to their own 12 volt battery that was not hooked up to anything else and then I run four 12-volt 12 amp hour batteries to the motor controller without any type of Hall sensor I just did not even use that plug all I use is the three motor control wires the power and ground wires for the motor controller itself and the start wire for it and then the green black and red for turning the motor controller on through my handlebar throttle control and I also have a switch because my motor controller has three separate speeds on it and to be honest with you that alternator motor has more power then my 125 cc four wheeler

    3. I would mount a Faulharber BLDC motor…to the shaft concentric to main prime mover…your alternator..that can be clocked and mounted with 360 degree rotation..the 244 series…and many other motors have (3) 120 spaced apart hall devices with a buffer circuitry. The hall sensors work off 5 VDC. They are phased correctly to the 3 phase BLDC motor. This little motor is capable of 25K rpm…so it can sense the phases of your alternator. Using an oscilloscope you can set the sensor BLDC 3 phase motor position to be correct with the main prime mover 3 phases…
      Now the generic BLDC motor controller will have the correct signals…for it will have the rotor position…from the hall sensors in the 244 BLDC little motor. Its rated for like 25 watts…and relatively tiny. Then the huge DC buss supply created from your batteries can feed the IGFET or MOSFET final 6 power transistors inside your motor controller and drive you car alternator. The little motor acting would be like a rotor position sensor supplying your controller with intelligence wrt rotor position.I wouldn’t think these alternators could produce say their old DC current rating x nominal voltage…like maybe 80 amps x 13.6 VDC…say 1080 watts continous…about 1.5 horse power. But if the DC buss voltage is alot higher the old alternator being driven as a motor could perhaps double or create more power for short periods. The alternator DC field winding thru the slip rings could be driven harder or weakened to make the most power. Isn’t nice to talk about something I never…but it seems it could work. One could even use the 3 phase power leads to drive the 6 power transistors…with signal conditioning and lots of EE work….good luck

    1. “a six-year-old who loves going fast but doesn’t like loud noises? ”

      But, you _DO_ want some noise coming from your kids toys…
      So you know if they’ve left the yard, or are up to something else!

        1. I knew a couple who had four very young boys. One evening while at their place watching some movies, I heard… nothing. So I peeked into their bedroom. No boys in the beds but their bathroom light was on. So I took a look in the bathroom and there were the four of them, covered in band-aids with liquid soap and shampoo *everywhere*, even on the ceiling. Their mother was not at all happy with her progeny that evening.

          Silent children are either sleeping or up to no good!

  2. As many others have mentioned, his approach to the sensors is very, um, well, confident? It is kind of like he put them in there, so they should work. Little to no concern over alignment, electrical noise, stray magnetic flux, etc.

      1. I agree, but in addition, another possible error is the asynchronous nature. Terrible brushes, worn brush contacts, and who knows what voltage and current should be used. Too many unknowns there for my taste. I’d at least make certain the rotor is staying energized, and maybe add a cheap optical encoder to verify the rotor position is where it’s supposed to be.

        Also, the wiring combinations could have been sorted out by energizing the rotor, and one winding of the stator. The rotor would align, and the hall sensor of the corresponding coil would be set high.

        One more issue, when does the hall sensor trigger? Could be triggering too short of time in the rotation, or too long. Sloppy timing should work at low torque.

        In short, I do not think the hall sensor error is useless, but it pointing directly to the possible problems.

    1. Yeah, he is not putting them in right. It you look at hall diagrams for most drives they overlap.

      Really yje simplest thing to do here is add an encoder with hall tracks. You do need the correct encoder for the pole count which is who knows what for an alternator. Probably 8 pole or something.

  3. That’s quite an impressive project! But perhaps a bit much for a 6-year-old? Maybe scale it down to something like a Powerwheels Jeep with improvements?

    I’ve also experimented with running an alternator as a motor. I found I could run it sensorless as a big stepper motor; but it worked better to add a shaft encoder. Mine was home-made; a disk made of PCB material, with slots sensed by opto-interrupters.

    Another thing I learned about alternators; these are high-RPM, low-torque motors. He may need a lot more reduction than just the single-stage chain I see in the photos.

    Oh, and their efficiency is really terrible. Nobody knows how to make inefficient motors like the auto companies. For my project (an electric lawn tractor), I gave up on the alternator and used an old auto generator as a motor. :-)

  4. This might be the answer to something else that was bugging me.
    Recently, a company has started renting out electric scooters round my town. Apparently, you can only engage the motor once the scooter is already moving (ie not from a standing start). Maybe the reason is that they’ve cheaped out on their motors?

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.