Please don’t judge [Alan] on his choice of vests. This project is from 1999 when it was common to see people rockin’ these threads. Anyone who has ever spent time on the University of Minnesota campus in Minneapolis, Minnesota will know that parking is at a premium. [Alan] had a 12-15 minute walk from his parking garage to his office and was considering a cheaper parking location that would balloon that to 20-25 minutes. But engineers don’t see problems, they see project ideas. He started work on a tiny electric scooter that could slim down his commute. Obviously he did find some success, but it was interspersed with failures that make his scooter the Fail of the Week.
After doing some research on existing electric scooters [Alan] headed to Ax-Man Surplus, a hacker’s paradise of engineering supplies spanning the electronic, mechanical, and chemistry disciplines. He picked up a 130V DC motor which draws 1A with a 1/8 HP rating. He built a rectifying DC power supply to test it and was laying down rubber in the basement of his building. But you can’t get very far running on an extension cord so the natural progression of the project moved to prototyping a battery for the unit; a dozen 10.8V 1700 mAh NiCad batteries. He figured it would be good to add in a fuse. His “testing” of a 20A fuse was done by shorting the connections which is how he discovered that the cells themselves have an over-temp cutoff built-in. We’d call that fail number 1.
The next issue resulted in all the cells in the battery going dead. There’s a great description of this that culminates in [Alan] realizing he probably charged them with reverse polarity. He swapped out for another pack and took the finished scooter for a rather scary test drive. He hit the accelerator and the thing took off.. with no way to stop it. Flipping the switch did nothing as it heated up and began to melt. And he only had the front brake installed which was no match for that powerful motor. We start shaking our heads when [Alan] kept repeating the same mistake, moving to a mains light switch with the same results, then adding a flyback diode also with the same results. So what’s going on with this fail? Why were the switches unable to cut power even after he moved them to the off position? Let us know what you think by leaving a comment.
We love the idea of building your own electric transportation. But electric scooters are getting to be a bit too common. We’d recommend a commuting tricycle. You might think that you’ll be made fun of, but not with the way this thing tears up the track!
Fail of the Week is a Hackaday column which runs every Wednesday. Help keep the fun rolling by writing about your past failures and sending us a link to the story — or sending in links to fail write ups you find in your Internet travels.
38 thoughts on “Fail Of The Week: Commute-Shortening Electric Scooter”
The switches didn’t work because he was using AC switches for inductive DC loads. He said he only measured a maximum of 4 amp load, but I bet when you switch that sucker for the first time with a person standing on it, it pulls at least 10x that at a stall. The rule of thumbs I’ve read recommend derating AC switches by at least 5x the voltage when being used for AC. Or, you know, use the right switch. 120 VDC is not friendly.
I use AC switches all day long on DC. it doesnt matter. what DOES matter is that he actually needed a switch capable of 100A as that was most likely what the inrush current was when he flipped the switch.
Switches can break AC during the zero-crossing. Switches have to break DC under full load. That’s why every switch you can buy has two separate, and very different, ratings for AC and DC use.
And the DC number is much smaller.
inrush current welds the contactors together. gotta have a soft-start functionality or an absurdly overrated DC switch.
Or a motor controller. Little Chinese EV ones aren’t more than $100 nowadays. Kelly Controller comes to mind. Cruddy product, nice price.
Or build one with a 555 and a mongo switching tranny.
in 1999 (when this was done), they were expensive or weren’t available.
Off the top of my head I think Mike and jordan are probably correct – drastically underrated switch for the inrush current that welded the contacts together. There should probably be a kill switch. Motorcycles have those too for exactly the same reason he encountered. On a car you just turn the ignition off. Neat writeup though!
Did he not think of, y’know, just lifting the rear tyre off the ground? Should have at least stopped it running off.
Otherwise, yeah, use a bigger switch. And maybe a removable link, fastened to his belt, to cut the power. Could be something like a car fuse, or even a bolt between some nicely-designed contacts.
I work w/ Alan and, believe it or not, he’s rocking the SAME vest this very day.
Good to hear he’s alive, I thought he’d died after the updates stopped…
Hey, if you find a look that works for you…
He had to use some mosfet for switching.
I’ve gotta say though as well, testing a fuse by shorting a very powerful battery seems like a bad idea. He had light bulbs and knew the current draw, wouldn’t have taken too much to create an array to overload the fuse without shorting things out.
Current limiting for charging via light bulbs also doesn’t sound like a great idea, I’m no expert here, but IIRC the resistance changes as the filament heats up, so especially if it’s not very bright it’s not going to be constant. I would think that if you’re going to do something like this then a proper charging circuit is pretty important.
I’d also say that after having one near accident with a bad switch, one should probably mitigate against the same problem (helmet, lights, disconnect) before just replacing the switch and trying again, then doing it a third time :-)
I had to kind of chuckle at his comment about mandatory disconnects. He thought that a rule made by a professional organization was silly (his words).
Perhaps the biggest lesson here is safety first, and following industry standard practices :-)
Incandescent light bulbs make nice soft current limiters—when you do all the math, they basically work out to be something like P=I³k, or V=I²k. What deviation there is mostly works in your favor in this case (primarily the resistance when the filament is cold).
But there are serious problems with charging using one: the comparatively slow thermal time constant means you don’t protect yourself against surges, much like a slow-blow fuse. And you don’t get to choose k, you have to determine k experimentally, although it is at least a function of (bulb type, power rating, voltage rating).
After reading a lot, we decided that the reason the switch failed even with a flyback diode is that AC switches don’t open very far because they assume there’s going to be a zero-current event soon to stop any arcing. DC doesn’t do that. So DC switches open much farther to prevent the arc from continuing. (And yes, I do have that vest on right now.)
So your dead-switch postmortem didn’t show that the contacts were welded together? Interesting… that must have been one heck of a continuous arc!
Just thinking… carbon doesn’t weld. Maybe some big enough carbon contacts with a nice slidey-lever assembly could handle the current, and shouldn’t present much resistance, after all they use it for brushes in high-power motors.
It might, conceivably, also act to slow down the starting inrush current, since the otherwise negligible resistance will have more effect, the greater the current draw.
There might even be some combination solution, involving a 3-stage switch, off -> resistor -> on. For a nice simple mechanical soft-start. Must be able to get some chunky carbon cylinders from somewhere.
Also would help the lifetime of the motor, as well as all the other electric bits, if he gave it a push-off before he started the motor. Or let it start with the wheel lifted off the pavement spinning free, with no load.
Hi Alan, been a while. Hope things are good where you’re at now! Man, I still remember when this happened! Mind immediately drifted to your welded switch and your “exciting ride”. So much experience garnered back then! Good thing too… Not sure how well we’d take the “experience” that comes with bangs and bruises 14+ years later!
Got out of my scooter phase when I promptly wrecked one… and my thumb… and my face… Oops!
Good to see that Alan (or his ghost) is still posting. Hindsight being 20-20, I think I’d have at least a big cartridge fuse/bus bar that could be pulled loose mechanically.
Kudos to even considering this. FAIL for me would be considering riding a scooter around Minneapolis other than in the summer.
Chains for a scooter would be a good mechanical hacker project. :-)
There are ways to limit inrush current and then switch over to full load.
One way is a BIG power resistor (or the light bulbs mentioned above even) in series with a switch, and then after a set amount of time (or monitor current and when it is low enough) switch in a parallel switch to take over the full load so you aren’t wasting energy as heat. I see that trick a lot as a soft start.
I’ve yet to see an e-scooter that is better than a bicycle. But any thing that includes “FUFF noise, and sparks, and a blue ball of plasma” can not be a fail.
It’s probably easier and cheaper to throw together a scooter than to motorize a bike. And a scooter is a LOT easier to transport and park than a bike. He still had to drive to/from home.
When I was at the U, I rode my unicycle.
It’s 2013 – doesn’t that area have MTA between the parking and the campus?
And the project dates from 1999
We had an electric motorbike when we were younger, and used a solenoid from a car starting circuit for switching current to the motor. They still did fry eventually, but were fine for a good while.
Ask the crew of Apollo 13 about switches that aren’t rated properly. He welded the metal and likely melted any plastic that was used to physically move the contacts. I work on my RV, and when it comes to that 12VDC battery, I’ve got physical disconnects that totally isolate the power, just in case. DC is not to be trifled with.
Glad he survived….
Why only a front brake? That would cause him to flip over at speed.
Hindsight is always 20-20, but this whole scenario is an excellent argument for a dead-man switch on a lanyard.
Especially on Apollo. You wouldn’t want it flying off home while Mike Collins is out doing his space-walk.
Use a effing relay… specially if its just a momentary switch. ISSUE FIXED!
That picture is a total fail on so many levels I am at a complete loss where to even begin over here! I am thinking that Adam Sandler could use it for inspiration if he ever makes a sequel to Zohan though.
AC versus DC load breaking comparison with a knife switch
Holy flurking snit! That’s my novelty cigarette lighter sorted out!
Did I just learn something and had fun doing it ??
Please be kind and respectful to help make the comments section excellent. (Comment Policy)