Ultracap Bike

[JCG]

Thanks guys. Good stuff... Willie, that just might be the coolest website I've ever seen. I just mapped the route I've been taking to work every day for the past couple of years (after moving out of Chinatown - no Golden Motor outlets there though ) and it's like I can finally visualize what I've been doing all this time. Thanks!

Fetcher, I would love to learn more about the way an induction motor would do its business relative to what I see with this PM brushless motor. It's easy for me to get all wrapped up in this side-project (which has been a lot more fun than I imagined it would be) but in the end, we're supposed to learn something from this and use it to make an argument for getting the old EV-1 (with its awesome 103 kW AC induction motor) running again. I have bought some books and have been bothering people in the EE department, but we have more controls people than electromechanical types. There's also the question of trying to put a 25 kW genset onboard...!

The DC converter questions (in the large vehicle's case, maybe either boost or buck-boost) and the issues with the controller concern me the most. The Kelly controller I've been using obviously does a nice buck-boost job with the huge swings in voltage that I've put it through (18-48 V) and seems to be able to regen well too. The induction motor is the long-term answer to many larger problems with e-vehicles... no ruining your magnets by going up to the Curie temperature, maybe more efficient regen, etc... I'd like to bounce some ideas off some of the experts around here where it comes to true AC motors being used with a high voltage DC bus. Shoot me a PM if you'd like to hear more.

Rodrigo, thanks for the links. I have been following work similar to Dr. Dixon's and Dr. Ortuzar's for about a year now. I'm a firm believer in the serial hybrid approach vs. other types, which makes it too bad that my bike is a wimpy parallel, through-the-road hybrid I need to do something like E=IR has been doing:

http://www.endless-sphere.com/forums/viewtopic.php?t=6429

 

[JCG]

Another series of thoughts on AC induction vs. DC brushless in vehicle applications. I did some searching and came across an old thread where they discussed electric lawn mowers... and AC regen.

http://cosmos.phy.tufts.edu/mhonarc/elec-trak/msg00253.html

One item really interested me.

Regeneration: In industry commercial inverters are used to vary AC motor speeds by intentionally changing the frequency supplied to the motor. If you overspeed the motor, the motor is returning the power back to its power supply. When you use an inverter that has a variable frequency capability, you turn the frequency down and your motor is a generator, returning its power to the controller, until your new reduced speed is reached. Or you can have a over running load, like when you drive your AC motor powered tractor down a hill, and your motor is a generator.
Here is the problem in industry: Most AC inverters cannot return the excess power to the power line. There is a diode bridge that takes AC power and charges a capacitor bank. (There are a very few specialized AC inverters that use a bi-directional bridge of transistors to charge the capacitors or return power to the AC line.) With diodes that is strictly a one way trip. The capacitor bank is connected to the motor via a bridge of transistors or gate turn off SCRs. This bridge is bi-directional. So whenever the motor is a generator, the capacitor bank is charged up and potentially overcharged. Some drives turn off to protect the controller. Some drives switch on a resistor to dump the excess energy. Here we are not really regenerative, although the motor does not know any better.
If you were to replace the capacitor bank with batteries, like some electric cars, you now have the ability to absorb a lot of energy from the motor, and you can easily be regenerative over the entire speed range of the motor, whereas to regenerate with a DC motor you need to do tricks with field windings and still may not be regenerative over the entire speed range.

It seems that if the DC bus of an AC motor's controller is accessible, you can connect it to the DC source you like, and can regen without playing the winding switching games that limits you so much with a DC motor. I am working on a way to test this using an ultracap bank by the way. I'll eventually post something somewhere else about it, as it's getting a bit off of this build-specific topic.

 

[paultrafalgar]

Any news, JCG? Eagerly awaiting developments!

 

[JCG]

Any news, JCG? Eagerly awaiting developments!

Yeah, Paul... it has been too long. Mid-term and spring break can be a rough time. I have been spending experimental time dealing with a 290 V, 27.5 F capacitor setup (6 modules in series) that is powering an AC motor and have been extracting regen from it by spinning up a cast iron pulley (flywheel) and then slowing it down. As for the bike...

The weather is finally getting close enough to "good" so that I can start trying to build a propulsion model for the bike by taking certain types of data. It mostly comes down to speed and power (or current) vs. time, like what I posted a while back. I have been riding the e-bike back and forth from work quite often lately and it's holding up well. For the propulsion model, I'll need to get a handle on three main things (I think): rolling resistance, wheel inertia, and aerodynamic drag. From that (along with F=ma) I should be able to slap together a model which tells you how the bike will behave given a driving schedule. Regen braking might be estimated as a percentage of recoverable energy while slowing down.

Let me plan the next rounds of data acquisition by solicitation... what would you like to see? I have two base options for the bike to be tested.

1) Ultracap only.
2) Ultracap & DeWalt (A123) 36 V battery pack in parallel.

I've done an endurance run on ultracap only, posted earlier. I think some people wanted to see hill data, start/stop data, maybe all the data from one of my 15 or so minute commutes (round trip, I suppose). You guys let me know what you'd like to see (within reason!) and I'll take the data, post some graphs, and develop the model. Time for me to quit stalling!

 

[europa81]

Please the combination. I've enjoyed your experiment from the beginning and see a practical use of the Capacitor in efficiently buffering energy from braking to the next boost needed for acceleration. Question is how to combine battery & capacitor. There should also be a provision to feed overflow energy from Capacitor into the battery during a long descent. This is the closest no zero-loss energy-buffering and allows for charging during pedalling times if you feel like it. Great setup and thanks for your continued updates. Made me order a Cycleanalyst, too.

 

[liveforphysics]

I have been spending experimental time dealing with a 290 V, 27.5 F capacitor setup (6 modules in series) that is powering an AC motor and have been extracting regen from it by spinning up a cast iron pulley (flywheel) and then slowing it down. As for the bike...

Are you working on over-unity with that project? I love that stuff, and I've been a long time magnet motor tinkerer. Any neat photos?

Personally, I think you should buy 2 more packs of A123s, open them, and form them into a 2P13S battery. That would double your range, and enable you to make use of the full potential of the ultra-cap voltage storage.

 

[JCG]

Please the combination. I've enjoyed your experiment from the beginning and see a practical use of the Capacitor in efficiently buffering energy from braking to the next boost needed for acceleration.

I think you're right, this is probably the best setup for gathering power-speed-current data since it causes me to not be as concerned with the cap's voltage going all over the place. If I can make sure to get good items like cruising power draw, power for different acceleration rates, and I can try to fit a curve of some kind to the data points, so to speak.

 

[JCG]

Are you working on over-unity with that project? I love that stuff, and I've been a long time magnet motor tinkerer. Any neat photos?

I do indeed have some scary photos. It's a scaled down drivetrain. I'll post something tomorrow if I can get a few minutes in the lab. But, I'll post it in E-vehicles general discussion just so I don't distract too much from the bike here. I need to start a new thread... I have a bunch of questions to ask anyway!

 

[JCG]

liveforphysics, I've posted some of the photos/video you requested here:

http://www.endless-sphere.com/forums/viewtopic.php?f=7&t=9384

Enjoy...

 

[JCG]

Ok, I just finished making another strange modification to the ultracap bike. I mentioned in an earlier thread my annoyance at the fact that I was riding around with a through-the-road parallel hybrid arrangement, when all I talk to other people about is series hybrids and how great they must be. So, I converted the bike into a series hybrid.

The real inspiration for doing what I did came from a couple of places; first, fitek did a neat job here which illustrated the idea of non-traction pedaling, and E=IR had a nice thread here using an engine-driven generator. E=IR's thread introduced me to WindBlue Power, where they make permanent magnet alternators. One of their models in particular, the DC 540, is capable of generating an impressive range of unregulated voltages at low rpm. I thought that this would be perfect for a my application, since low rpms are all your legs can do, and I was OK in charging the capacitor with an unregulated DC voltage as long as I kept an eye on the total capacitor voltage. Here we go.

Here's the bike from the right side. Note the lack of a chain heading from the crank back to the rear wheel cassette. The only way to move the bike is with the motor. The 42 tooth crank now sends its chain up to a 13 tooth sprocket.

And here's the left side. The chain on this side heads from a 45 tooth sprocket to the alternator (fitted with an 11 tooth drive sprocket).

Direct connection of the crank sprocket to a small sprocket on the alternator wouldn't give high enough rpm to generate high voltages (and therefore acceptably large charging currents). So, I had to go about the dual reduction with a shaft fitted in a mounted roller bearing. I fixed it to the water bottle holder screw holes and kept it from twisting by bolting it to the lower part of the alternator.

So now, pedaling at 30 rpm should give about 400 rpm at the alternator. Here's a close up of the alternator's drive sprocket.

The back side of the alternator has two output connections from its internal rectifier. Here, I've duct taped the positive one to keep from shocking myself or creating a short to ground through the bike.

There's also a three phase AC output there which I don't use. One neat thing about this arrangement is that it gives me practical freedom from a charger, even if the cap goes all the way dead. In the old series arrangement, I couldn't send regen back into the cap unless the controller had at least 18 V to still be turned on. Today, after some time and a lot of sweat, I brought a dead cap up to 40+ volts by pedaling in place, keeping my balance by holding my hand up against the wall. It was exhausting.

At low capacitor voltages, the current draw from pedaling is around 8 amps and you feel significant resistance, but nothing overwhelming. As the voltage climbs, it's easier to pedal and you're forced to go faster to keep a 2-3 A charging current.

The alternator and gears don't stick out as much as you might think. The way I mounted it, I was able to keep my legs (and pant legs) clear.

I took the bike for a test drive this afternoon and it was a weird experience. It would take a lot of getting used to to pedal a bike without the expectation that it will help you move, and that messed me up a few times at low speeds. I was able to cruise at a moderate speed without observing a capacitor voltage drop, so that would mean that I was able to provide a the required electric power by pedaling the generator. Regen from the hub motor still worked as usual.

Overall, it was interesting, but it was pretty clear that this wasn't the best bike to try a series hybrid arrangement. With a two-wheel bike, if you're stopped at a light, your foot needs to be on the ground - not running the pedals to charge things up. This kind of thing might work best in a trike, I believe. You can just sit in one of those with the brakes clamped and charge things up while waiting somewhere. I'm not in the mood to install training wheels, so I'll go back to the parallel arrangement once I get some more time to work on the bike.

 

[paultrafalgar]

Excellent work, JCG! More ground-breaking stuff. Am I right in thinking you concluded that series hybrids are not worth further investigation, or is it worth trying on a trike? What would be the comparative efficiencies of the two systems, given that parallel hybrids have to use a drivetrain and series hybrids don't?
Any news on Ammonia-drive?

 

[JCG]

Hey Paul,

No, both hybrid architectures are fine with their associated pros and cons - it just turns out that this conversion, in removing physical traction from a 2-wheeled bike, causes worse problems than it solves.

I liked:

A) The ability to make the cap "come back from the dead" without relying on the controller and regen
B) The IDEA of charging things up without riding around in circles endlessly
C) The higher charging efficiency that the alternator would give you vs. playing phase games with the BLDC motor

I didn't like:

1) Holding on to something while I pedaled to charge at a stop
2) always using the motor to move around, especially at low speed/high torque where DC motors have terrible efficiency
3) the fact that I added some weight and complexity to the operation

In a trike, you can at least get rid of (1), and (3) might not be so hard to swallow. Eliminate the concerns over those, and (A) and (C) can really be nice bonuses. I wish I had a trike!

The ammonia genset will be an end-of-the-summer thing. Annoyingly, I'm still waiting on a quote for the engine. But, the motor controller on the EV1 gets replaced at the end of June. I can't wait!

 

[Link]

1) Holding on to something while I pedaled to charge at a stop

I've always wanted to fit a large gyroscope onto the bike somehow so I can just spin it up with my braking energy and just sit there at stops to screw with people. Try it for me? :wink:

 

[JCG]

I've always wanted to fit a large gyroscope onto the bike somehow so I can just spin it up with my braking energy and just sit there at stops to screw with people. Try it for me? :wink:

A practical way to screw with people? Yes... I'm on it!

 

[paultrafalgar]

1) Holding on to something while I pedaled to charge at a stop

I've always wanted to fit a large gyroscope onto the bike somehow so I can just spin it up with my braking energy and just sit there at stops to screw with people. Try it for me? :wink:

Link, I'm surprised at you! A man with an IQ of 167 hasn't noticed that the kinetic energy of a gyro is dependent on MASS, just what you don't want on an ebike!

 

[JCG]

Link, I'm surprised at you! A man with an IQ of 167 hasn't noticed that the kinetic energy of a gyro is dependent on MASS, just what you don't want on an ebike!

No problem, we'll just take a light one and spin it up to a bajillion rpm.

 

[paultrafalgar]

I don't mind having my leg pulled, but I'm damned if I'll have my scrotum minced by an escaping turbine rotor going a jazillion rpm!

 

[Link]

Link, I'm surprised at you! A man with an IQ of 167 hasn't noticed that the kinetic energy of a gyro is dependent on MASS

...which only partly determines its moment of inertia when combined with its effective distance from the center of rotation. Either use a small one spinning at...uh..."a bajillion" RPM or have most of the rotor's mass at its edges and spin it at a more reasonable RPM.

LOL @ worrying about mass on an ebike.

Also, I've got some spare 2x4s lying around that could POTENTIALLY be used as the frame for a large, lulzy box that JUST MIGHT be able to stand on its own corner (or perhaps keep a bike propped up on nothing) for some magic reason.

I'll have it ready by Tuesday?

 

[Link]

I JUST NOW GOT THIS. DERP.

 

[JCG]

I JUST NOW GOT THIS. DERP.

MS Paint strikes again! :lol:

 

[317537]

Interesting thread. Read some of it and will read the rest.

I was thinking of installing cheap electrolytic caps. I know its not like the beasts this esu has but they can store energy from regen for accelerator bursts.

I played with some low voltage utra caps and found the more saturated the DC field is the better they charge.

I have no idea if the charge input has circuitry installed for ripple rejection.


I found my lithium phone battery chargers to not be pure enough DC to get near enough charge into any ultracap. Vot sure about other chargers though as chemical battery chargers do not require the same pure DC field ultra caps do.

High current can create ripples. Charges and discharges that chemical batteries have no issue with.

How fast can the swich mode be? Fastest switch mode you can get. You can best charge the caps on an SLA battery. overvolt the caps by a few volts if possible and cut off before the terminal voltage reaches max.


Capacitors charge infinitely, even though you might reach 98% capacity in a 20 mins or however it takes to charge, the other 2% will take an eternity.

I wish I had your toys. I can only speak as if I owned this ultra cap bank myself and what Id try with it.

Forgive me for my suggestions.

 

[JCG]

Leslie, an ultracap pack might not be as expensive as you think. Depends on what you're seeking. Once you know what total voltage & capacity you'd like for your application, there are a lot of options that you can customize yourself.

The biggest caps sold by Maxwell are 3000 F, 2.7 V cells. Let's say that your bike/rider combo is about 115 kg and you want to try and store regen braked energy from 30 kph speeds. KE = 4000 J, and let's guess that at best your DC motor controller can recover 50% of that into capacitor current (2000 J). The required capacitor energy would be E = 1/2 C V^2, so your CV^2 product needs to be about 4000 J. If you happen to run a 36 V battery, then your required capacitance is 3 F.

Maxwell's HC series, retailed by Tecate Industries, are rated at 2.7 V each so you could get 14 of them to match the required voltage. 14 capacitors in series to produce 3 F total capacitance need to be at least 42 F each. Here are some nice 50 F units for 9 USD each.

 

[317537]

I am collecting 2.5 55f caps. I have a few of them.

Not enough just yet. My bike is a 48v 24ah sla ugly beast. Caps with SLA's for regen would be better for the batts.

I am worried about the caps soaking up to much regen and dumping it into the batts.

I did this experiment with a 6 amp reg and solar panel. I put some caps over the panel out and hooked it up to the regulator, when I removed the source, the 5X470F caps dumped it load straight into the batts. SLA do not like sharp current feeds. To some degree the current draw should be limited by the battery technology however you can fry SLA with current.


I think more than one cap in parallel on SLA's may give too wider path for current. Even wiring the 6 caps in series with three 1 ohm 10 watt parallel resistors would allow me to reap enough benefits for a project.

My SLA's cant take more than 5 amps, so problems are going to pop up when the caps aggressively soak up and dump the load.

 

[Reid Welch]

My SLA's cant take more than 5 amps, so problems are going to pop up when the caps aggressively soak up and dump the load.

Just woke up. Six pages behind. I know how basic caps work and understand the ripple issue and delicacy of high farad, low voltage caps.

My IQ is better termed "QI" (for strange intelligence)


Stupid posit: resistance-limit the regen current going to the ultra caps from a DC motor. LCR filter too?
The losses would be huge, but hell, it's regen charging, and what's a hot resistor between two legs, anyway?
a choke, if chosen to be used, would need to be huge and full of heavy, heavy, copper and iron.

Radio Ranch Reido, certified future graduate certificate-ee of the IEEEEK:
[youtube]oNc3a8A11wA[/youtube]
What does this have to do with Gene Autry?
Well, consider yourselves as the animals, the singer-thread-starter as god, and myself, as
a grad-ju-ate student, rarin' to go to work in 1936!

 

[317537]

I planned on using 3 resistors for each 12vbattery, the leads and terminals to and from the resistor are thin and produce much heat. 4X 10 watts at 1.8 ohms (parallel) might provide less resistance over the leads.

Wouldnt the energy dump from regen be shared with the battery? The voltage differences between sag and peak is only minimal 2 volts at best on each SLA, at .3 ohm id only be heating up the resistor to .2 watts, hot resistors, ouch?. The rest would either would be charging the battery or dumping into the controller. Total regen at 5 amps 48v 240 watts. I want 2 to 3 amps to charge caps. Yes sure those resistor would almost melt charging the caps from 0 to 14v but when does this happen after the first initial charge from 0 volts.

It takes me say around 60 seconds to charge from 0 a 25f cap with 500 ma at 5v. Holding the leads will burn your fingers and using a PC PSU will freeze the PC as you draw more than an amp from a 500ma supply. You hear all the cap plates inside the PSU flex as soon as you connect a 0v cap to the charging power. Once its charged there is very little heat created to recharge caps, this is why the resistors will hold up in a regen situation.
14v
+ ---| |---| |---| |---| |---| |---| |---WW--|ii

T