Ultracap Bike

Ok, as I promised to Ypedal in the E-bike Tech thread http://tinyurl.com/5vru2k, I have fired up the digital camera and started this thread to chronicle progress (or lack thereof) for this project. Yesterday I ordered a bunch of stuff: a NineContinent kit (700c wheel, thumb throttle, Cycle Analyst, no battery) from Renaissance Bike Co., a regen-capable Kelly controller (KEB48221), lots of insulated wiring and ring terminals (8 AWG), and a big, watertight, lockable enclosure for the back of the bike. While all that stuff is on the way, I thought I could at least post pictures relating to the energy storage system: the ultracapacitors. First, let's look at a single module:

http://i400.photobucket.com/albums/pp85/jcganley/16V330F_single.jpg

Notice I have cleverly placed some heavy metals and carcinogens in the photo to show how tough I am. Just to keep the gearheads and Edisons in the crowd from laughing at the Chem E.

This thing is rated at 330 F and 16 V. It's made up of six ultracap cells in series (each 2000 F and 2.7 V). It can store about 42 kJ of energy and weighs about 4.5 kg (10 lbs). To reach a voltage closer to the kind used by the hub motor (36 V), I'll connect two modules in series:

http://i400.photobucket.com/albums/pp85/jcganley/16V330F_dual.jpg

I could add a third module (I have one) to make 48 V, but I'd be better off just using this 18-cell behemoth:

http://i400.photobucket.com/albums/pp85/jcganley/48V160F.jpg

Rated at 48 V and 140 F, it is probably more juice than I need (and weight too)... but it'll still fit in my enclosure if I get adventurous. I accomplish my charging in the lab setup by using a variac through a 20 A bridge converter:

http://i400.photobucket.com/albums/pp85/jcganley/variac_bridgerect.jpg

I slowly increase the transformer voltage to the rectifier to keep the charging current at about 5-10 A. Here is the setup before charging:

http://i400.photobucket.com/albums/pp85/jcganley/no_charge.jpg

Notice here I've connected the cell balance cables, but I won't want to do this while charging in the lab, as there's a sink of about an amp during charging. It's not recommended to use the balancer unless you're charging with at least 18 A (more than the variac can handle). Halfway charged:

http://i400.photobucket.com/albums/pp85/jcganley/charging_halfway.jpg

And full charge:

http://i400.photobucket.com/albums/pp85/jcganley/full_charge.jpg

Note that the variac is off, the caps are holding their charge. It was at about this time that I realized I had no way of safely discharging these things. I reacted instictively and started electrolyzing some water:

http://i400.photobucket.com/albums/pp85/jcganley/electrolysis.jpg

Hydrogen generated on the left (-), oxygen on the right (+). Don't try this at home, kids. I could have filled the whole room up with an explosive mix if I didn't have great ventilation in the lab.

I hope this first cut was enjoyable, I'll post more when parts begin to arrive. Certainly, I should be able to use this setup to test motor/controller/throttle coordination and possibly using the motor as a generator all while monitoring cap voltage and current. Also sorry to paste links rather than the pictures directly, this website was snipping off the right third of the photos. Thanks to all of you for your help and advice so far. I'm not done asking for it by a long shot...

JCG said:

It can store about 42 kJ of energy and weighs about 4.5 kg (10 lbs).

Click to expand...

A Joule is equal to one Watt per Second.

Normally we will figure battery power based on Watts per Hour. (Wh)

An hour is made up of 60 Seconds * 60 Minutes = 3600 Seconds

42,000 Joules / 3600 Seconds = 12 Watt Hours (Wh)

The bike I'm building will have a 634 Watt Hour (Wh) battery and will be able to go about 30 miles while at 48 volts / 25 Amps. (so I hope anyway)

Comparing the two:

634 Wh / 12 Wh = 53 times more.

Unless I've done something wrong here I think you aren't going to get very far... have I made a mistake with the math?

(it looks like your range is less than a mile per module)

Read the original thread:

http://endless-sphere.com/forums/viewtopic.php?t=7393

You'll see that range doesn't matter. It's an "experiment".

It's a cool idea and I'm all for the EESTor idea of building an ebike scaled supercapacitor.

If the secret to storage is voltage, then what's needed is higher voltage to get the extra capacity.

Storage = 1/2 * Surface Area of the Capacitor * Voltage (Squared)

...a slight increase in the voltage and you increase the storage a lot.

(but run the risk of damaging the capacitors)

Great stuff !!!!! Keep it up !

That thing looks like it could kill you in a Flash.. i love it !!

( Safe, don't invade the thread, you know exactly what i'm talkin bout here.. questions are fine.. but none of the usual conversation domination.. please.. )

safe said:

Storage = 1/2 * Surface Area of the Capacitor * Voltage (Squared)

Click to expand...

Hi Safe, that's pretty much the case. The term in that equation is actually C (capacitance in Farads) which is related to both surface area and the permittivity of the material between the plates. Maxwell gets a high capacitance by using acetonitrile as the dielectric material and high-surface area activated carbon. BaTiO3 of the purity claimed by EEStor could be expensive to produce, but we'll see.

As for the bike, the consideration here is to use acceleration bursts and soaking as much as possible back while braking. A 500 W acceleration busrt for 5 seconds only uses 2.5 kJ of the 84 kJ stored in the twin module series. The trick is getting that 2.5 kJ back by braking or pedaling against generator resistance... should be fun to try.

Ypedal: you bet... if it kills me, it will be accompanied by a flash. 32 VDC is nothing to sneeze at, especially when those packs can dump about 4300 amps with a short circuit.

Nanny here. Do you have a good indemnity policy if you kill one of your students instead?

JCG said:

As for the bike, the consideration here is to use acceleration bursts and soaking as much as possible back while braking. A 500 W acceleration busrt for 5 seconds only uses 2.5 kJ of the 84 kJ stored in the twin module series. The trick is getting that 2.5 kJ back by braking or pedaling against generator resistance... should be fun to try.

Click to expand...

I like the idea. For an ebike our overall needs are low enough that ultracaps seem a good way to go. For something like BMX where you do a lot of surging and then stopping this would be good and maybe some sort of parking lot racing could come of it.

It's a very interesting project... I hope you learn a lot and can pass that on to the rest of us...

paultrafalgar said:

Nanny here. Do you have a good indemnity policy if you kill one of your students instead?

Click to expand...

They get to use the wrenches (spanners?) and ride, not to do the wiring! :lol:

I posted the following info about this subject on the Ultra Caps? thread awhile back... http://www.endless-sphere.com/forums/viewtopic.php?f=2&t=5581&st=0&sk=t&sd=a&start=15 " http://www.afstrinity.com/xh.htm Aptera.com seems to be working on a similar powertrain using ultracapacitors for regen braking and power boosts during acceleration... "

Glad to see someone is testing it out on an electric assist bike...don't test the contacts with your tongue... :wink:

Your first instinct was to to electrolyze water as opposed to using a resistive load? LOL

Do keep this updated. I'm interesting in seeing how it goes.

Ok, time for an update since a bunch of stuff has showed up, and finals week is coming up (meaning I can relax!).

First of all, I dusted off the bike and brought it in. Here it is with new brakes and a rack added to the back to hold up the power supply and controller (and whatever else can fit):

http://i400.photobucket.com/albums/pp85/jcganley/Raleigh_C40_before.jpg

A day later, I got the NineContinent Kit from Ebikes.ca (thanks to Justin, Alison, and Zev):

http://i400.photobucket.com/albums/pp85/jcganley/9Cont_Kit_Contents.jpg

Here you can see the hub motor with 700C wheel, thumb throttle, 20 A controller (which I'll use for the initial tests) and the Cycle Analyst. I also got the regen controller from Kelly but I am having to pull some major creative moves to get the 8 pins connected to the 14 pin junction. If someone knows where I can get a real 14-pole female plug for that thing (seems to be 3/4" diameter) please let me know. Both I and Radio Shack are stumped. If I'm not too embarassed by the results I'll post them here with a photo later on.

Anyway, with a student's help I was able to quickly get the tire and inner tube on the wheel, and snapped it into a vise:

http://i400.photobucket.com/albums/pp85/jcganley/hub_with_tire.jpg

Ready for a spinup test. I used alligator clips where necessary and put my variac-bridge rectifier combo back to work, shown here providing 24 VDC:

http://i400.photobucket.com/albums/pp85/jcganley/24VDC.jpg

Note the bolted-on inverted aluminum test tube holder acting as a heat sink for the rectifier. That thing got really hot during capacitor charging and squirting it with alcohol to keep it cool was getting old. Ok, I'm going to try and add a video, spinup test at 24 VDC:

http://www.youtube.com/watch?v=l4PIvWys9E4

We only went to half throttle to make sure that thing wouldn't twist out of the vise. I had to spin it by hand to get it going, as it appears that the NineContinent controller requires "pedal first," kind of annoying, but the Kelly shouldn't require that (I think). Here again at 48 VDC:

http://www.youtube.com/watch?v=QxCXv57DpFo

Yes, I know, boring. I also spun it by hand with the controller on and off just to check the dreaded rolling drag without load, it didn't seem too bad (we'll see more later).

Anyway, after that I was ready to start tackling the Kelly controller, which allows for regen. As others have noted on this board before, the Kelly controller can have its own independent power source and uses only a small amount of current. The voltage can be anywhere from 18 to 90 volts or something crazy like that, but I don't want to be up a creek without a controller if the capacitor is completely discharged. So, I wanted to connect 18 AA batteries (NiMH) in series just for the controller. I had a great plan to make a wooden frame:

http://i400.photobucket.com/albums/pp85/jcganley/Wood_frame_before.jpg

You can see my plans on the piece of paper. I used the mill to cut battery-sitting grooves into the wood:

http://i400.photobucket.com/albums/pp85/jcganley/Wood_frame_after.jpg

and put in some nickel foil contact strips, along with some bolted-in terminals:

http://i400.photobucket.com/albums/pp85/jcganley/contact_strips.jpg

The result: horrible, I couldn't even get the two sides joined up. There was much cursing. Then, I found some 1/2" schedule 40 PVC tube from an old experiment, which happens to have roughly the same inside diameter as a AA cell. The result:

http://i400.photobucket.com/albums/pp85/jcganley/battery_tubes.jpg

24 volts of duct-taped goodness that I can see through. I actually made some improvements on this and it looks a little better than this photo, but you get the idea. I can use this pack or recharge it with my DC power supply (voltage and current limited) as needed, but since the Kelly draws about 37 mA during operation, these 2500 mAh cells will give me several days' worth if required.

For now it's back to MacGyvering the Kelly input junction.

I have at least a thousand miles on SubC cells inserted into a PVC tube. With my project I used pennies from before 1982 as contacts on the ends and used a spring to provide tension. In order to get clean and flat contacts I used a belt sander to grind down the pennies to make them smooth.

The idea has worked well.

It's going to be interesting to see how this turns out...

If you use the tags, we would be able to see your pictures in the post without having to open each one separately......

Safe: Great! Seeing this done elsewhere gives me hope that it might be stable. Everyone is going to laugh at how big the enclosure is behind the seat, it's going to hold a ton of stuff.

michaelplogue said:

If you use the

tags, we would be able to see your pictures in the post without having to open each one separately......

Click to expand...

Yes, I need to figure out how exactly what dimension to resize my images so that they don't automatically get cropped by this site, I'll try again with my next batch of pictures.

Today's test is going to be trying the 48 V behemoth ultracap module with the NineContinent kit with the regular controller. I have to get all my stuff in the same room eventually, might as well try today. Hopefully I'll have another mini-update tonight!

JCG said:

I need to figure out how exactly what dimension to resize my images so that they don't automatically get cropped by this site, I'll try again with my next batch of pictures.

Click to expand...

If you resize them with an 800 dpi width, they will fit on most people's screens without cropping.

Ok, quick update from today's work. My camera was on its last legs (left the charger at home and didn't want to accidentally fry it with my own recharging methods here in the lab) so this won't be too intense of a read.

Step one was to get the hub motor/wheel mounted on the bike, which I promptly turned upside down to get ready for further testing. I hooked up the variac/bridge rectifier again to the NineContinent kit.



Here's a video of just spinning off of the variac power.

[youtube]rUHWtlP43NM[/youtube]

Ok, remember this guy? The beast, behemoth, big dude. 48 V and 160 F, stores 51 Wh when fully charged and can dump it all in three seconds if you let it.



It's heavy, but at least I won't need to worry about balancing multiple modules in series. The intra-module balancing circuit seems to dissipate power a lot less quickly, and therefore wastes less energy while charging, than the inter-module circuits. These things were not meant to be charged with a current lower than 18 A. I hooked up the variac/rect for charging, and connected leads from the ultracap to the motor controller's battery terminals. This is a mess, but I had to monitor the charging current to be sure I didn't kill the variac (it's the weak link).



I charged it to ~38 V, turned the variac off, and ran a quick test.

[youtube]SJfRZJpmUrM[/youtube]

And then another at full voltage, full throttle. No variac, capacitors only. The voltage didn't swing much at all, as expected, but again 1.5 A is not much. It will be more fun to try under load.

[youtube]ei8NQJzUwaM[/youtube]

Next up, mount the huge enclosure on the rear rack and try to take a test drive. It'll be a charge and go operation, around the block a couple of times (I hope).

Michael: thanks for the width number for the images - I think it worked!

JCG said:

Next up, mount the huge enclosure on the rear rack and try to take a test drive. It'll be a charge and go operation, around the block a couple of times (I hope).

Click to expand...

I'm going to bet that you get about 2-3 miles out of it. The idea is great to watch as you get it developed. Since charging can be nearly instantaneous you could go 2-3 miles, charge it, then go again and again and again.

Near infinite battery life.

Cool. 8)

safe said:

Near infinite battery life.

Click to expand...

don't count on it.

Toorbough ULL-Zeveigh said:

don't count on it.

Click to expand...

http://www.maxwell.com/ultracapacitors/index.asp

With over 1 million duty cycles on the large cells (the type in my modules), the caps will easily outlast the motor and everything else on the bike, the bike itself, and the next one, etc. Li ion cells are lucky to get 1000 duty cycles before they lose most of their capacity.

Nice work !!! keep it comming !

About that rear rack.. and the weight of this magic blue box.. i speak from experience, and as most here will confirm, at some point the bike Will Fall Over, I hate it when that happens but it happens.

The pack seems very well built and should handle a tumble but what kind of uhm.. looking for the right word here.. " event " ? can one expect from a fully loaded charge taking a fall and opening up .. ?

I've dropped an 8ah nicad pack that put a nasty dent in one cell, nothing bad happend and amazingly it still works.. have not yet poked a big hole in a lithium cell.. but i have accidentally shorted a 72v pack for a millisecond ( long enough to nearly shit my pants )

If this thing can discharge in 3 seconds " safely " i can't imagine a dead short..

JCG said:

Toorbough ULL-Zeveigh said:

don't count on it.

Click to expand...

http://www.maxwell.com/ultracapacitors/index.asp

With over 1 million duty cycles on the large cells (the type in my modules), the caps will easily outlast the motor and everything else on the bike, the bike itself, and the next one, etc. Li ion cells are lucky to get 1000 duty cycles before they lose most of their capacity.

Click to expand...

ah yes, if only spec sheets were real world.
it also sez 10 year lifespan.
that's only 2-3 times the lifespan of Li-ion/NiMH (or equal to A123).
a little shy of inifinite wouldn't you say?

Supercaps do fail, just like anything else.
I've replaced enuf of them over the years, albeit in the 1-50F range.
very intolerant to the slightest overvoltage, worse than tantalums.
not saying they ain't great, but no point in overstating its capabilities leaving people with an unrealisitic impression.

Ypedal said:

About that rear rack.. and the weight of this magic blue box.. i speak from experience, and as most here will confirm, at some point the bike Will Fall Over, I hate it when that happens but it happens.

The pack seems very well built and should handle a tumble but what kind of uhm.. looking for the right word here.. " event " ? can one expect from a fully loaded charge taking a fall and opening up .. ?

Click to expand...

Ypedal, this is a very good observation, and it has been on my mind for a while. In fact, this weekend I decided to rethink my plan for designing an enclosure for this thing. Rather than a plastic electrical enclosure with a lid, I am now looking at making a sheet metal sleeve (waterproofed by me) that will fit over the cap/controller combo and to use the structrue of the cap module to do most of the bracing required to keep it attached tightly to the rear rack. The design is being made up on paper, and I'm scouting around for spare material to cut up and rivet together. This might take a while unfortunately...

Great project. I'm imagining pedalin' like a madman all the way right up to a red light, to charge up the pack for when the light turns green..

I doubt that he's going to be able to recapture much. But I like the idea...

As a sort of stop-start sprint bike this could be awesome.

You could make races where there was a lot of starting and stopping and bursts of speed (sort of like a BMX track) and this would make sense.

Everyone doesn't want to grind along at 100 mile long rides.

So there might be uses for the idea outside of the lab.