A different kind of E-bike

[JCG]

The model Golden Motor sells from their site is probably the worst, but there's another you can get from e-bikekit.com that I'm going to try out soon that seems to spin more freely. An X5 would be serious overkill for a capacitor powered bike and probably be hard to pedal, so I wouldn't go with one of them.

I have had my eye on the E-bikeKit 36 V, 500 W simply becuase they claimed the ability to do regen on their website. The tech on the phone told me that the regen engages with the pull of the hand brake... I don't mind buying a bunch of pieces separately, but it would be nice to get everything in one kit. If regen is decent with the kit's controller, I may be able to have all I need to do initial testing with that kit (keeping in mind that I can go get a Kelly controller later). Please, let me know what you think of that kit - both regen and rolling drag!

If the motor drag is really a killer, I'll consider first getting a 4XX motor, Kelly controller, and assorted accessories to round things out. Changing things up with an external motor and clutch is the last resort. Time is kind of working against me now that the end of the semester is looming and I need as much "plug and play" as I can get with these students.

 

[johnb]

But this reminds me of another problem with a capacitor's discharge curve: A motor will not want to spin faster than the voltage with which it is fed, meaning that the speed you can pedal at (powered or not) is directly proportional to your remaining capacity. If, say, the motor would top out at 25mph at the capacitor's full voltage, it's going to try to recharge the caps if you try to pedal faster than about 13mph. Only way around this is a big diode preventing power from flowing backwards into the battery, but that will kill regen, too, so you'd also have to have a relay to provide a pathway around that diode that's activated when regen is.

I don't see that being a problem. When the controller is not activated, the MOSFETs will not be conducting so will not allow current to flow to the cap. That's assuming the the throttle is switched off once you run out of motor push.

 

[JCG]

A quick note about these ultracaps, since there has been some concern over their change in terminal potential as they discharge... these things have huge capacitance which will limit the voltage swing as long as they're used for short periods of time only. For example:

The capacitor bank is constructed of two 16 V, 330 F modules to make a series unit of 32 V, 165 F. If, during acceleration, I pull 35 A for 8 seconds, then:

i = C*dV/dt
and
dV = i/C * dt = (35 A)(8 s) / 165 F = 1.7 V

I could expect only a 1.7 V drop, most of which (perhaps 1.2 V) I could hope to reclaim when I brake at the next stop light. I shouldn't be seeing huge voltage swings during the brief motoring bursts; but rather just over the long term of a full trip (if I can't pedal it back up, but that's going to be part of a fun experiment).

 

[swbluto]

Maybe one could get an accurate value by driving it with another motor, or weights and string.....

"I find that I can pedal with my 409 and hardly notice any drag." What figure do you put on that?

Anyway, it's the input power needed that we're concerned with, in this case.......

The 40x motors have a cogging torque of about 0.5 to 1 N-m which translates to about 10-20 watts at 200 rpm. The 409 has the lowest cogging torque.

That's about .5 n.m. at 0 rpm and about 1 N.M. at 400 RPM(With the graphs). According to Justin, the cogging torque should be irrelevant to the winding count and there's quite a bit of variability from motor to motor. Of the four hub motors I saw, there was a range of ~1 N.M. at a given RPM with no clear pattern(As winding count increased, the cogging torque just jumped around.).

 

[Mathurin]

The resistance of a bottle generator is sufficiently objectionable to make it worthwhile to disengage it every time you don't need the lights. They'll suck ~8w at 30km/h.

A while back on Velectris forums one of the members determined resistance from motors by wheel component weights, spindown time and doing science. I don't recall the numbers he got and can't seem to find that topic again, but he'd found the Phantom wasted about 1/2 of what an 408 did.

 

[dnmun]

The capacitor bank is constructed of two 16 V, 330 F modules to make a series unit of 32 V, 165 F. If, during acceleration, I pull 35 A for 8 seconds, then:

i = C*dV/dt
and
dV = i/C * dt = (35 A)(8 s) / 165 F = 1.7 V

what happens when dt is on the order of milliseconds when the capacitor shorts out? have you calculated the amount of energy that is released inside the car if the capacitor bank fails, as in an accident?
it would be neat to see pictures of them, but it is depressing to read about an EV1 still in existence about to be converted.
why not consider just replacing the battery packs with lifepo4 and kelley controller and maybe figure out how to supplement the battery pack with your ultra capacitor.
what if you could buy some cheap lifepo4, say 12 48V20Ah ping packs, and in a gesture to science, he sells them to you for $6k delivered.
kelly controller $600-900, 200A at 48V(56 charged) so 11.2 kilowatt or 15 hp. total of 12kWh, use 100wh/km, so 120 km max.
i assume the project started with the capacitors, i never think they will be capable of overcoming the true risk of catastrophic failure and explosion, you don't really need to take circuits, this stuff is not like analog circuits.

see if you can start a project to support your ultra capacitor conversion by creating a network of charging spots to stop at during your commute or at each end. then make them so electric bike riders can use them also, that would be a great use of grant work too. BOL, dennis

 

[JCG]

Dennis,

I know how you feel about the EV-1. Long ago we did the calculations to show that just replacing the old lead acid batteries with Li-ion would give a 300 mile range. It has been depressing to see it gather dust in our parking garage while we wait for our College to deliver a place for students to work on it and bring it back to life... but the plans are in place.

The car is much closer to running condition than many might think. The batteries are gone, but the wiring, motor, inverter, and other power electronics are still all in place. The kicker: GM disabled the motor's controller, or more precisely removed the logic board inside it. One of the lightest, most efficient 3-phase induction motors ever made and all of the huge IGBTs are still under the hood, but there's nothing to tell them what to do. There is a guy in Ohio who has made controllers for EV-1s, and did one each for Ohio State and BYU. We'll eventually plead with him to visit when the time is right.

The capacitors are not as hard to deal with as you might think. In fact I'd rather deal with them than the full battery packs of the old days, since at least when I'm installing them (and working on the car frame around them), they're discharged and couldn't hurt a fly. Get a small, efficient diesel genset installed in the trunk with a high power bridge rectifier, and you can charge those things up in a hurry. The 100 kW motor in the EV-1 should pull around 300 A peak during acceleration (like it did with the old lead acid batteries), and around 50 most of the time (cruising). Initial calculations show the series hybrid EV-1 getting 80 mpg on the highway, and 120+ mpg in the city thanks to the fast-charging caps. I don't see the cost of high-energy batteries coming down enough, soon enough. I think the Chevy Volt is going to die a silent death because the batteries will make it cost $45k. Ultracaps have a better shot in a series hybrid... Maxwell has already reduced cost an order of magnitude and they're not even mass producing yet. Best of all, the caps are designed to outlive the rest of the car (1 million charge-discharge cycles translates to hundreds of years of practical use). We have chosen a complete capacitor bank of 486 V, 16.5 F. Give me a high power booster (which we'll likely build from scratch), and this thing will go 700 miles on a tank of biodiesel.

And give me a 300 A charging station to share with the EV-1... my bike'll be charged in less than 18 seconds.

Keep the faith.

 

[dnmun]

as you likely know, when honda was developing the insight hybrid Integrated Motor Assist series type hybrid drive, they originally considered using large capacitors for power storage, but ended up using batteries.

i would always be wary of being in an enclosed space with a fully charged ultra capacitor. batteries cannot release that much energy in such a short time period if they are shorted, because of internal resistance and the resulting drop in the cell potential, and therefore current when shorted, but this is not the case with large capacitors. you need to do that calculation, it will be a large safety issue with the DOT.

we had large rf capacitors in the rf generators we used to drive nitride deposition equipment, and they were nothing close to this in size, yet they blew up one of my technician's screwdriver when he shorted them out while performing safety setup in advance of working on the rf generator.

i don't think zenn will ever get over this hurdle either if they ever can actually develop their own design of an ultra capacitor.

i am a believer in universal and standardized distributed charging stations, with the ebike as the personal electric vehicle, and a social networking structure to allow sharing of the charging hubs, which could be located anywhere and everywhere to allow the commuter to carry the smallest battery possible, and be able to get a boost 'fast' charge of 1C or so to continue on when the batteries run low. people can carry an integrated charger built into the controller so all they need to do is stop off at the house of someone on the EV charging list and plug in to the 240V AC outlet after entering their ID into the monitor for the charging hub. manage security and authentication through the facebook like social networking structure, and bill the cost for the charge to their account.

i think the idea of carrying your energy source with you will become passe and a relic of times past of high energy density fuels derived from crude, or now bio crude(biodiesel and bioethanol), which in itself is a crime against humanity, with over a billion starving now. imo.

if you are a chemist now, it's not too hard to make the leap over to physical chemistry, properties of materials, solid state, alloy chemistry, then you can be in battery development too. anode structure developments and so on. BOL, dm

 

[etard]

Welcome D.C.,

I bet there are some local bike shops that sell electric bikes, surely Georgetown or Alexandria would have at least an izip that you could ride. There is also a manufacturer somewhere there in NoVa that has a bike with a dual hub drive system that gets up to 30 mph. Go ride one, you will see why we are all after lighter weight. Better yet, go strap about 60 pounds to your bike and ride it to school, then you will see.
I would suggest to make it more enjoyable for your students, get a tandem. I bet one of the parents and or faculty have one they NEVER ride that takes up TOO much room in the garage. It might be possible to have one person peddling into a regen setup while sitting at the light.
Use a 24 volt hub for lighter weight. If you push 30 amps out of the motor, isnt that 30 amps of force dragging against you? I think it is noble to test the limits of technology, and I applaud your involving possibly the next Dr. Tesla in your quest.

 

[Link]

I don't see that being a problem. When the controller is not activated, the MOSFETs will not be conducting so will not allow current to flow to the cap. That's assuming the the throttle is switched off once you run out of motor push.

The body diodes of the FETs let it happen. :?

I could expect only a 1.7 V drop, most of which (perhaps 1.2 V) I could hope to reclaim when I brake at the next stop light. I shouldn't be seeing huge voltage swings during the brief motoring bursts; but rather just over the long term of a full trip (if I can't pedal it back up, but that's going to be part of a fun experiment).

FWIW, the most I've ever seen anyone get out of regen on an ebike is 7%, and they were staying within the batteries' charging capabilities and using lithium (something like 98% coulometric charging efficiency). And given that you're going to go through more electrons (and therefore voltage) each time you accelerate because the decreased voltage equates to decreased power...

 

[johnb]

I don't see that being a problem. When the controller is not activated, the MOSFETs will not be conducting so will not allow current to flow to the cap. That's assuming the the throttle is switched off once you run out of motor push.

The body diodes of the FETs let it happen. :?

You are right, power MOSFETs do have that feature.

I could expect only a 1.7 V drop, most of which (perhaps 1.2 V) I could hope to reclaim when I brake at the next stop light. I shouldn't be seeing huge voltage swings during the brief motoring bursts; but rather just over the long term of a full trip (if I can't pedal it back up, but that's going to be part of a fun experiment).

So you are expecting to recover 70% of the acceleration energy when braking. I don't see any way of getting more than 50% efficiency during acceleration and regen, giving an overall recovery of 25%.

Have you been using http://www.ebikes.ca/simulator/ for your design?

 

[JCG]

FWIW, the most I've ever seen anyone get out of regen on an ebike is 7%, and they were staying within the batteries' charging capabilities and using lithium (something like 98% coulometric charging efficiency).

So you are expecting to recover 70% of the acceleration energy when braking. I don't see any way of getting more than 50% efficiency during acceleration and regen, giving an overall recovery of 25%.

There are a lot of records on the web of designs to blend ultracaps into next-gen hybrids due to their ability to soak large regenerative currents - pure electric storage, low internal resistances of the newest caps. The regen efficiency estimate of up to 80% using ultracaps is discussed by people ranging from Honda (used it in their FCX vehicle) to EERE with the US Dept. of Energy. Most of the technical journal papers I have are copyrighted so I can't post them here, but a nice public-domain white paper is here from BGSU:

http://www.bgsu.edu/downloads/tech/file55534.pdf

Regarding a series hybrid bus, designed for the stop-and-go urban environment. Note the 150 kW regenerative power he has measured. That would destroy a lead-acid (and even moreso, a Li-ion) battery pack with a similar voltage rating to the capacitor banks. Jeff has designed, built, and tested these buses. He also happens to be the "guy in Ohio" I mentioned earlier who I hope to enlist in the EV-1's resurrection. ISE Corporation in California has also had success incorporating ultracaps on their hybrid buses.

Can (to scale) such high currents and high recovery powers be obtained using a hub motor and a suitable e-bike (maybe kit-based, maybe Kelly) controller, using a human switching system? I don't know, but I'd like to know. Given that I almost have what I need on hand, let's see what happens. I'm under no illusion though, Murphy's Law is suspended for no one. The way I see it, the efficiency of regen in the bike application will either be good enough or not, it will either be a limitation of the motor, the controller, or the human (or all of the above). Time to find out.

In the spirit of that, I think I'm about ready to pull the trigger on an order. Any objections to (keeping in mind all previous bits of advice):

http://www.goldenmotor.com/e-Bike-DIY/DIY-ebikekit.htm
700c rim (to fit my Rayleigh C-40), front wheel, 36V/500W, regen controller, twist throttle ($427)

Comments welcome as always! There are several other motor options there, by the way. If this controller can't do the high current regen, I'll get a Kelly next.

Thanks all, for so much help so far.

 

[paultrafalgar]

JCG, if it proves true that the ultracaps+regen are not a success, I hope you won't ditch it but will add a small LiFePO4 pack and try the combination. Could be interesting. Good luck.

 

[dnmun]

goldenmotors has a bad reputation, never responding, go search here for the numerous complaints.

the optibike seems to have the best reputation for regen, but even they eat up the battery packs.

 

[JCG]

goldenmotors has a bad reputation, never responding, go search here for the numerous complaints.
the optibike seems to have the best reputation for regen, but even they eat up the battery packs.

Thanks D. Does Optibike sell components or kits? I seem to have found only complete units on their site.

 

[johnb]

FWIW, the most I've ever seen anyone get out of regen on an ebike is 7%, and they were staying within the batteries' charging capabilities and using lithium (something like 98% coulometric charging efficiency).

So you are expecting to recover 70% of the acceleration energy when braking. I don't see any way of getting more than 50% efficiency during acceleration and regen, giving an overall recovery of 25%.

There are a lot of records on the web of designs to blend ultracaps into next-gen hybrids due to their ability to soak large regenerative currents - pure electric storage, low internal resistances of the newest caps. The regen efficiency estimate of up to 80% using ultracaps is discussed by people ranging from Honda (used it in their FCX vehicle) to EERE with the US Dept. of Energy. Most of the technical journal papers I have are copyrighted so I can't post them here, but a nice public-domain white paper is here from BGSU:

http://www.bgsu.edu/downloads/tech/file55534.pdf

That 80% efficiency figure quoted in the reference is for the capacitor only. My 25% figure was calculated for the motor/generator only. Adding the capacitor figure would give an overall efficiency of 20%. This is for a bike with typical direct drive motor. I can believe that a HEV with a more sophisticated control system may improve on this.

 

[dnmun]

jcg, optibike is hand built, about 24 bikes a year i think. designer/builder is jim turner at optibikes, north baseline road, boulder, co. google him, i used to drive by that address to visit an old girlfriend, never thought the best ebike in the world would come from that end of boulder.

you can go to the tidalforce forum on googleforums too and ask around there. a lotta people with comments about regen.

anyway, he may be interested in your project, and can give you some help, or connections to people with familiarity with the real problems using regen. you might need to talk to stephen at kelly controllers to see which one would work best for regen too.

 

[velias]

There was an article a year ago in circuit cellar ink, August 2007
they had schematics and everything of a battery and ultracap powered hub motor bike
I guess the ultracaps helped with the acceleration from a dead stop, it seems thats all they could do.
I get the impression that the article was just a student project done by some more poor car-less canadians and wasnt fully completed or tested.
Here is the abstract and link to it:

Electronic Bicycle Design
by Roxanne Ahmed, Michael Buchanan, Andrew Clarke, Chris Elder, & Warren Shrieves
This electronic bicycle will work only as hard as you do. Motor power is controlled by the amount of torque you apply to the pedals: the harder you ride, the more torque is added to the rear wheel. Lithium-ion batteries and ultracapacitors power the system. p. 24

http://www.circuitcellar.com/magazine/205.html

 

[Mathurin]

Circuit cellar in question: http://rs101.rapidshare.com/files/44090186/205.zip
And for programing the pic: ftp://ftp.circuitcellar.com/pub/Circuit_Cellar/2007/205/Buchanan-205.zip

I absolutely LOVE their torque sensor. DIY bionx style, just about as good as an ebike accelerator could get? Gonna have to dig into it when I'm less tired. On the surface it looks like a a solution to the need for those retarded motorcycle-like twist grip throttles.


Mr. Léonce Rudelle actually has a patent for his regen thing. The video that made him famous is rather Randy Draper-esque, tho:
http://www.wideo.fr/video/iLyROoaftIIS.html

And here's a link to his patent, not sure how it works:
http://www.google.com/patents?id=rc4ZAAAAEBAJ&printsec=abstract&zoom=4&dq=rudelle#PPA1,M1

 

[JCG]

Thanks all for these updates. Your help here and older threads elsewhere are really going to make this whole project go smoother than it would have gone, that's for sure.

Dennis - your point about efficiency is well taken. I foresee the need to do a good job of data logging during operation to get really good numbers to report - voltage and current history, avg. and peak power, etc. Suggestions on the best system monitor (that won't cost a million dollars)?

Paul and Velias - using a battery-ultracap combo, where the cap is there to take the huge current loads off of the battery, is a really popular topic among Maxwell's top brass. I think it makes a great fall back option for this bike in case caps alone for bursts fails for one reason or the other.

Most of all thanks for those links. I'll spend some time during lunch today checking those out... should be worthwhile.

I've got a new angle on the items - will be buying an incomplete kit (motor, wheel, throttle, rear rack) from EBikeKit.com and a separate controller from Kelly (likely the one I mentioned in a previous post). Purchase should occur next week when my department Chair is back (with his purchasing card).

 

[dnmun]

i repeat the warning about goldenmotors, go see if you can use the nine continents hub motor kit from justin at ebikes.ca if you don't want to use a C'lyte direct drive hub motor. i know you will get better service, and justin is about the smartest cookie in the forum. he may even have ideas about how to do your regen, and he is designing a new controller which may have application for you.

 

[JCG]

i repeat the warning about goldenmotors, go see if you can use the nine continents hub motor kit from justin at ebikes.ca if you don't want to use a C'lyte direct drive hub motor. i know you will get better service, and justin is about the smartest cookie in the forum. he may even have ideas about how to do your regen, and he is designing a new controller which may have application for you.

Ok, I'll give him a call today or send and email instead.

 

[JCG]

Renaissance's Nine Continent kit comes with the Cycle Analyst... I like. This plus a Kelly controller:

http://tinyurl.com/5qveat

and I'm set for the first experiments.

 

[swbluto]

Interesting. I have my doubts the 9-continents motor will draw(or generate) 100 amps, though, so you may be buying more controller than you need.

I might be wrong. Letme see using my program... It seems to draw about 60 amps with no current limit at 33 volts, but I don't know what the EMF voltage would be during regen. It doesn't seem like I know how to determine the EMF voltage... :/

 

[JCG]

Interesting. I have my doubts the 9-continents motor will draw(or generate) 100 amps, though, so you may be buying more controller than you need.

Yeah - it must be kind of obvious that I'm spending departmental money instead of my own. I like the wide input range for this controller, its programmability, and the high current handling (if needed, but the system probably won't even come close).

I also have no idea of what kind of generated emf to expect, other than I hope that the controller can work with it! I may need to engage regen with a potentiometer or something. I'll need to play with the controller to know, I suppose.