Very slowly, some very cool parts are coming out on the market that will make for some awesome builds. Supercapacitors are becoming a thing, and every year, the price of these high power supercaps go a little lower, and the capacity gets a little higher. It’s really only a matter of time before someone hacks some supercaps into an application that’s never been seen before. The Navy is doing it with railguns, and [David] is building an electric bike, powered by AA batteries. While [David]’s bike technically works with the most liberal interpretation of ‘technically’, it’s the journey that counts here.
This project began as an investigation into using supercapacitors in an electric bicycle. Supercaps have an energy density very much above regular capacitors, but far behind lithium cells. Like lithium cells, they need a charge balancer, but if you manage to get everything right you can trickle charge them while still being able to dump all that power in seconds. It’s the perfect application for a rail gun, or for slightly more pedestrian applications, an electric bike with a hill assist button. The idea for this build would be to charge supercaps from a bank of regular ‘ol batteries, and zoom up a hill with about fifteen seconds of assistance.
The design of the pulsed power DC supply is fairly straightforward, with a mouthful of batteries feeding the supercap array through boost regulators, and finally going out to the motor through another set of regulators. Unfortunately, this project never quite worked out. Everything worked; it’s just this isn’t the application for the current generation of supercapacitors. There’s not enough energy density in [David]’s 100F supercaps, and the charging speed from a bunch of AA batteries is slow. For fifteen minutes of charging, [David] gets about fifteen seconds of boost on his bike. That’s great if you only ever have one hill to climb, but really useless in the real world.
That doesn’t mean this project was a complete failure. [David] now has a handy, extremely resilient array of supercaps that will charge off of anything and provide a steady 24V for a surprising amount of time. Right now, he’s using this scrapped project as a backup power supply for his 3D printer. That 100 Watt heated bed slurps down the electrons, but with this repurposed supercap bank, it can survive a 20 second power outage.
It’s a great project, and even if the technology behind supercaps isn’t quite ready to be used as a boost button on an electric bike, it’s still a great example of DIY ingenuity. You can check out [David]’s demo of the supercap bank in action below.
47 thoughts on “Fail Of The Week: An Electric Bicycle, Powered By AA Batteries”
How much current can those supercaps really give? 24Vx10A for several seconds – that’s pretty low. 7 li-on 18650 batteries can give you over 26Vx30A for several minutes and will cost about $35.
You are right about the smallish (compared to load) super cap bank, but this project shows a basic lack lack of understanding power requirements.
But I like the idea and it’s definitely salvageable. Why not charge from a full size bank of li-ion batteries. Then instead of relying on the super caps to pull the entire load, use the super cap bank in parallel with the battery power for acceleration and the batteries to sustain your speed? This would work on flat ground and pulling hills.
Or better yet, see if you could use regenerative braking to charge the capacitors.
EVEN better yet eat more KFC!
I’ve searched internet for some hard data, but couldn’t find anywhere an example of high current reasonably priced supercap bank. Everyone uses normal big capacitors for high current and says that their current is “couldn’t measure, but very high” which usually means 100-150A. Nowadays you can get 18650 lions rated for 30A, so 10 of them in parallel will give you 300A.
Look up amperics on Amazon. They are selling a 12v 350A supercapacitor string for $70. I use 2 of them in parallel as starting power for my 2013 Kia Soul. My YouTube page via my website also has a few videos detailing this usage.
Or just spend the supercap money on more lithium and you’d be able to accelerate that hard all the time, plus have longer range.
Or get an ICE and some nitrous oxide.
You can get a used Tesla battery pack module off eBay for relatively cheap and have 18650s for everything. When I looked into supercaps for my bike, the cost was way higher for the performance I was looking for. Plus most eBike controllers won’t handle that kind of amperage anyway.
My bike runs at 113VDC and 45ish amps for about 20 miles at top speed (55-58 MPH), plus I’ve replaced the batteries in my electric screwdriver, cordless vacuum, etc…
The issue is, lithium batteries tend do die out of neglect if your bike doesn’t see regular use.
A set of AA batteries is cheap and has a long shelf-life, and hold a reasonable amount of energy.
But the power output isn’t that great. If he had used a bunch of C or D cells instead, he would have gotten the thing to charge up in a reasonable time.
No? Lithium cells stored between 30-80% charge should last for years without any significant wear? As long as you don’t drain the battery fully and forget about the bike for years, you should be fine.
Why would you build an electric bike if you plan to not use it for years at a time?
The boost assist is for those who don’t ride often enough to develop the leg muscles. If you ride every day, short hills are no sweat.
What I mean is, you have to take care of the batteries – disconnect them, recharge them to a suitable level, and then store them in a cool dry place instead of just leaving them on the bike out on the driveway or in the garden shed until you need it. Otherwise, since you might be weeks or months between needing the bike, it’s going to sit in the heat and the cold and slowly discharge its batteries, and then the cells will die. In any case, since you can’t leave them fully charged because lithium cells degrade fastest at full charge, so you’ll first need to recharge the battery before you set off.
The point of using primary cells in an electric boost is that they have practically no self-discharge so they can sit for 10 years without degrading. You don’t have to baby them, you just turn the switch and go.
You’ll still have to take care of the batteries even with D cell alkalines. The difference is now you have to change batteries out frequently when their charge is depleted. A typical alkaline D-cell has a capacity between 18-27 watt hours. My ebike consumes about 20 Wh per km of travel. With my commute to work of 20km round trip, that’s about 400 Wh consumed, which would drain 14.8 high capacity D-cells.
Or I could just plug my bike into a charger and only replace my lithium battery after several years of daily use.
That’s why it’s a booster, not a continuous power source.
That’s why it’s a “fail of the week”. If you go through the effort of making an electric assist / drive system for your bike, you might as well pair it with a decent power source so you can get some use out of it.
None of the batteries I’ve used have displayed short shelf-life. And then there’s LiFePO4, the most durable battery chemistry I’ve ever seen–period.
Are you thinking of lead acid?
I like the Lithium INR’s also… they are impressive… though the LiFePO4 are great also… especially since they’re closer to 1.5V x2 AA alkaline = 3.0V ~= 3.4V x1 LiFePO4
It’d probably work better with a little onboard generator instead of the AA batteries.
If you want fast recharging, you’re going to put a major drag on the pedals.
Considering the small generators are hardly efficient, to get the 4,000 Joules necessary to run the booster for 20 seconds, assuming 50% efficiency, in some reasonable time like 200 seconds of cycling, you need to pull 40 Watts out of the pedals, and that’s a significant load for a casual cyclist.
For a point of comparison, the average guy can maintain about 200 Watts for an hour. Note that because the combined human mechanical and metabolic efficiency is about 25% that’s already getting you hot and sweaty, and generally pissed off about the whole business. A casual rider may be putting out only half of that or 100 Watts.
Perhaps [David] should have checked the definition of a Farad before he started. 1 Farad is enough stored energy to produce 1 amp of current at 1 volt for 1 second. 100 Farad is 24V for 4.17 seconds. You’d need a lot of 100F capacitors to get any meaningful boost.
If it even was that good. Each capacitor has a max voltage of 2.5V or so, so he needs multiple in series and we all know that when you put capacitors in series, the capacitance goes inversely down so his 100F becomes 10F.
You need way more caps. Capacitors in series drops the capacity, need to do series and parallel to get back that capacity. I got my shipment of caps in 2 weeks ago and tested a few, its amazing how much energy they still have in them. I think in a hybrid setup of lipo/super cap is where they will really shine. Use the battery to supply the cap and the cap can dump power when needed and make life easier for the lipo!
Well that is a tricky topic Mike. You need more information in order for that fact to be true.
If the supply voltage is the same, adding an additional capacitor in series will decrease the capacity of the string.
However, adding an additional capacitor allows greater max voltage. So by increasing the supply voltage along with the additional capacitor in series will keep capacity somewhat steady (there’s a balancing parameter that affects total capacity). Doing this will increase max power ratings as well.
Yes, Regen braking is the right thing to do here. With option to trickle charge while pedaling as well. Caps should efficiently charge on the fly. No need to condition. Get rid of batteries.
Yeah, makes sense (to me) to charge the caps slowly with pedal power. Would cause the rider to expend slightly more energy during normal situations but then give them a boost on hills. I could see that being super useful, especially on a commute, to keep from having to pay for (or recharge) new batteries and keep from getting sweaty from a few hills or overpasses.
Regen bikes have been tried, but they always fall on the same point: too slow to charge up, or too hard to pedal.
The AA batteries in this case are giving out about 5 Watts. A regular casual cyclist is putting out about 100 Watts. If you want to squeeze the charging time down from 15 minutes to 1.5 minutes, you’ll need to put a 50 Watt load on the pedals and that’s going to be hard on the cyclist.
On my back burner is a electric medical scooter, to be converted into a childs electric “off road” buggy.
Time to get crackin’, grandson is waiting.
And no, it will not use regenerative braking, (Bosch (r) )
Regenerative breaking is not something invented by Bosh. It is as old as electric cars, from when speed was controlled by switching different series parallel combinations of batteries in the 1800′. Most 3 phase motor controller chips have regeneration incorporated, with DC motors it’s a little more complicated, but have been done for 30 years and off the shelf controllers have this. Remember it is an extra braking system and it saves wear on front suspension and brakes, so definitely worthwhile.
All true. Modern systems are all A/C power that is provided from the on board D/C to A/C inverter.
(I’m sure you knew that)
While tinkering with D/C power, high amps are not your friend.
My early experience with D/C was 36 volt gas turbine starter motors.
Amps were quite high, and the system locked out after most start failures due to temp.
I wish I had one of the motors for a gocart..
A lot of early EV designs used recycled series wound starter motors. They provided plenty of torque but being starter motors, they were not rated for continuous load and would quickly over heat. The problem is they have no provisions for cooling.
Regenerative braking wasn’t so much invented as discovered, because it’s the natural consequence of a brushed DC motor speeding past the battery voltage.
Another interesting point about DC motors is that increasing the field current makes the motor turn slower and with more torque. You can design motors that have an adjustable “k” and use a smaller current from the batteries to control the speed.
“with a mouthful of batteries”
Such a tingly sensation
This is the fun kind of project I tend to enjoy. Just going for it on a Sunday afternoon fun and mad science. Spend the following week reading up on what I managed to do wrong by folks with more knowledge. That’s definitely a way to learn. I think he can iron out the kinks in this for something fun and useful :)
Interesting idea, even if it didn’t work out.
One of the things I’ve wondered about is using a combination of supercapacitors and batteries for both rapid acceleration and rapid regenerative braking. Basically, if you know you are going to want to tromp on the accelerator pedal and really go fast (like “launch assist” on the Nissan GT-R), it would charge up the supercapacitors, which, when triggered, would dump a higher amount of power into the motor than the battery pack could normally deliver on it’s own. Basically the same thing in reverse for regenerative braking – dump it first into a capacitor, then “trickle” charge the batteries from that. I imagine really stomping on the regenerative braking would generate too much power too quickly to dump directly into the batteries, so having the capacitor act as a buffer between it and the battery pack would allow it to handle larger spikes in regeneration.
Then again, I don’t know much about this stuff. Somebody feel free to correct me.
Correction not needed..
The facts are, electric motors supply maximum torque at zero RPM.
That “Fact” is why all diesel electric freight trains are built the way they are.
So, “Dumping” voltage into caps may or may not provide additional torque.
To be determined..
>”The facts are, electric motors supply maximum torque at zero RPM.”
Not true for all types of motors. Induction motors specifically don’t, as you have no induction at 0 Hz, and the torque decreases as the “slip” of the motor increases beyond a point.
PMDC motors do, with the caveat that the efficiency is abysmal and the motor overheats after a little while.
Also not true for parallel wound DC motors and squirrel cage AC motors, the type that run most fans and loads under 1 HP. The only motors that provide high torque at low RPM are series wound motors.
The reason why diesel-electric trains are built the way they are is mainly because a single wheel on a train has very little traction and the force has to be spread over multiple wheels.
Connecting all those wheels to the diesel motor and keeping the torque evenly spread despite the mechanical tolerances would be practically impossible. With electric motors, each pair of wheels has its own motor, and the powered wheels can be put under the carriages as well, so the tractor doesn’t need to pull the whole train alone.
So, if I follow this thought, a diesel motor connected to each axle would have the same torque at low RPM?
I think not..
A diesel motor connected mechanically to each of the driving wheels could have any arbitrary number of reduction gears and produce just as much torque – the point is that the mechanical complexity of connecting all the wheels to the motor is too great, hence the electric motors.
I was told that diesel locomotives dump the current from their electric motors into a shunt for braking.
The person who said this was told the handrail around the locomotive is part of the shunt.
As far as re-gen goes, would it still make sense to have the regenerative braking charge a capacitor bank first, then use that to charge the batteries? I guess I’m under the assumption that there’s a limit at how quickly you can dump power into a battery to charge it. Any regenerative braking that happens that generates more power than that limit would have to be dumped somewhere – either as heat, or temporarily stored.
You’d think the super capacitor fast/high charge rate design would be ideal for braking along with a fast/high charge rate battery that seems with relays, or some sort of switching, triggered when the stage of fastest charging device is charged you could charge control what charges the fasted first so to later use the energy in the same order though to discharge the stored energy to accelerate. Especially if you have smaller engines that aren’t as efficient on acceleration. I’ve complained about this for years… I want to say in the late 90’s though for sure when super and ultra capacitors were coming out more in the early 2000’s up at Tech since then the battery technology was still more in the lab and not main stream mass produced yet.
I’m not aware of any main stream implements yet that are consumer stock. Race cars I want to say were the first I observed and now there is more potential for consumer grade though I haven’t seen any kits yet for the hybrids or all electrics other than to replace the starter battery.
There is definitely potential not only in battery advancement in density… super capacitors as well. Then using the capability of like say saving energy in using less high friction mechanical operations… using less loss electro-mechanical or electronic operations to conserve more energy and utilize more effectively for longest term lifecycle lowest maintenance operation of components and the system in general.
I have 500f super caps in my house power supply.
So that when I change over form 120v-12v to batteries there is no drop in the voltage.
Not getting ‘Fail’ as setup appears to be proof of concept and attempting to run a scale EV on a bank of 10 AA NIMH batteries reaching absurdity. Steam punk points for analog gauges. Minus points for lack of : schematic, graphs, and had to zoom and pause to get EBM-PAPST DV6424 off grainy vid for motor spec. Ha.
Please be kind and respectful to help make the comments section excellent. (Comment Policy)