How Not to build

an

Electric Bicycle

Always do the Last Part First

I had some 1/8th inch plate strapping kicking around and did some creative Bending and Welding. I think at this point the shiny little sparks from cutting steel was boring me and the big sparks from a welder had me in it's grip.

The blame for this project lay squarely with our local International Tractor Dealer. In a moment of weakness I broke-down and allowed a third party to work on my tractor. I was sick of un-charged batteries, fading head-lights when I needed them most, at night.

Regardless, they did an excellent job and the IH504 was fitted with a real pretty alternator in place of the Bosch Generator that had caused so many problems.

I asked the dealer if I could have the original generator back unless it had a Core value. He said it didn't and I could have a few if I wanted. Free generators!!! I didn't want to appear gluttonous so I grabbed the 3 largest and heaviest that I could find and left it at that.

The above image came from somewhere, I'd gladly give credit if I only knew where.

Like any great folly, a cooling-off period must precede a passionate work I waited 2 (two) years before revisiting the units and pondering their ultimate purpose. One is a Remy Delco with Field and Armature posts and frame ground. The other two are matching Bosch Units Identical to the ones found in VW Beetles from the 70's, these are the 12V Units not the 6V as I originally thought. Perhaps I could make a twin prop Electric Gyro Copter, with foam Billet pontoons to quietly fly-in to my favourite fishing holes. Or maybe a few real ugly Wind Towers with 100ft diameter props I would cast from Carbon Fibre (something I wanted to learn about) or an inverted Trike ATV with all wheel drive and Hydrogen fuel-cell power.

By chance while I was researching the Bosch Generators on the Web I saw my first Electric Bicycle. I don't know who's site it was or what it had to do with the generators, but I was hooked and had to build one. Most were ZAP bolt-ons or some form of friction drive on the rear or front wheel. I would have none of it! No off the shelf controllers, no tried and true plans or sub assembly's. Every thing from scratch, almost including the wheel.

Base Unit

This is a pawn shop special I got for $25.00, a sturdy frame with tires that hold air. As with any pioneering initiative, don't spend anymore than you absolutely have to, incase you get bored. If I do nothing else at this point, at least I ended up with a bike.

A sturdy frame supercedes the laws of physics

By placing a 180 degree twist on the back lengths any flexing in the frame was removed. I took great care to use existing holes and posts from the back brake mount. At this point things are going so well I'm thinking that I can do this in a weekend. Next I'll throw together a quick and dirty motor controller and go for a quick rip. The truth is it took 3 weekends and evenings to get to this point. Most of my time is getting up to speed on Field strength Armature speed dependancies, the long and short of it is...

Max Torque at full Field, Armature speed increases as field strength decreases.

I stole the gear off of the Rear wheel on an old bike's that the kids had out grown, machined a flange and welded the whole thing to the 15mm shaft of the Bosch Generator. I took the drive gear from the pedal assembly and bored it out to the hub diameter of my bikes rear wheel and welded it carefully into place. By fully dis-assembling the hub and removing the ball-bearings I did avoid problems that others had reported when welding to a thin free-wheeling hub. I did set my welder to the lowest current I could and still draw a strong enough spark to flow the filler.

The motor is hinged on a 12" spike that is welded directly to the frame. It pivots on it's factory flanges perfectly after much trial and error. The chain tracks the two gears at any speed I pedal the bike while it's tied-up from the ground. The chain tension is maintained by the universally useful bungee cord. This is a design element that I originally implemented on an air compressor after upgrading the motor. In that case it was just to test the pulley alignment. It worked so well it's still working today with no adjustment.

In defense of my ignorance, I did have a sense that gearing would come into play. The pair of gears I used were originally configured with the larger driving the smaller, by a pair of short 5 to 7 year-old legs. How much possible power could there be, I'd go the cautious route, reverse the drive orientation to maximize the lower end torque of the Bosch, not that it really needed any help. I honestly believed that I had thought this through.

This would prove to be one of the few critical assumptions that would prove to be fundamentally flawed and cost me in time, energy and expense.

I also was doubly pleased with my drive assemblies independence from the pedals. The motor could drive the bike and the pedals would free-wheel.

To this point, I had separated the Field Windings from the Armature and wired #10 stranded cable with spade connectors to numerous stacks of Sealed Lead Acid batteries.

Note that the Motor is bolted to the bench. I came up with this arrangement given the heft of the heavy generators, and the prospect of it or parts flying around the shop. To the right of the motor are a pair of Amp meters to monitor current draw in a no load condition. On the same panel are high current switches to energize the field and Armature individually.

The optimal seemed to be 12V on the field and 24V on the Armature. By simply cutting the Field line really quickly on and off and back on again the motor would pick-up speed to the point that the whole bike shook violently from the slightest imbalance due to such high speed.

*Current Measurements*
	Field 12V	Armature 24V
slow	8	12
accel to med	6	22
Speed peak and stop	drop to zero	pinned off scale

These observations are crude at best, but helped me to understand the Armatures tendency to increase in speed, in an effort to compensate for the diminishing field. To make any meaningful measurements I would have to get up to speed on Motor Controllers.

The motor was completely stripped down, bearings cleaned and re-packed with fresh grease. The Armature was set into the chuck of a drill press used as a lath to re-surface the commutator ring with 600 grit paper.

Over the next few weekends as friends would stop in, over a cool beverage or two I'd demo the drive assembly. Everyone would eventually yell over the loud whine of the chain and shaking frame to shut it down before it flew to pieces. Everybody wanted to know what the top speed would be.

I never did work-out the numbers beyond approximations but 90 to 120Km/h kept recurring as a guestimate. I was so caught-up with enthusiasm I started to believe those stories that we have all heard about, the 13 year old kid that replaced a 2.8l V6 in an S10 4X4 with a cordless drill motor that drives 75 miles at 75 Mph on a teaspoon of tap water.

If this was true I couldn't create this Web Page, the CIA or even worse the Big 3 may silence my dissenting voice and I would just disappear.

At this point I had improved my welding and machining skills, and learned more about motors and SLA batteries than I though possible could exist a month prior.

Quantum Leap in Power Storage

The SLA's pictured to the right are Medical Grade Hawker's that are rated at 26Ahrs. These units were generously donated by a friend I work with. One was being used to start a VW Deisel motor that was adapted to run a Generator. The other to run a trolling motor I think. Regardless these batteries are greatly appreciated.

Up until getting the Hawkers I had been using 7.5Ahr SLAs recovered from dead UPSs. Before being able to use the UPS SLAs I built the Desulphator circuit found at the HomePower web site (www.homepower.com). The results were quite impressive. The SLA's I had were bottomed out at approx 0.5v, and refused to charge beyond 7.5v with a standard charger.

The Desulphator required a fully charged battary to run an oscillator that rang a tank circuit at the resonant frequency of the sulphate deposit on the dead batteries plates. After a week to 10 days the Battaries were holding fast at 12.5 to 13V.

An important thing to realize is that 4 (four) 7Ahr cells wired in Parallel totaling 28Ahrs is nowhere close to a Hawker 26Ahr Cell. The small cells can only deliver 2 to 3Amps continuous current per cell totaling 8 to 12 Amps. I personally used the hawkers to boost tractors, and trucks in freezing temperatures with no ill effects on the cells.

1^st Generation Field Controller with non-functional Regen Brake Circuit

I looked at the Curtis Controllers, spent hours pouring over the 4QD data and schematics. I really should have just bought either one. The 4QD probably is more deserving as they have my respect for the open source atmosphere and careful documentation that is available on the InterNet. (http://www.4qd.co.uk/)

Given how much there is to know about Motor Controllers in a real world application, it's amazing that I could make this first generation brd work at all.

The design is based on the a 555 timer configuration that has a near 10% to 90% range in pulse width. I have to give credit to the source of the 555 Timer circuit, http://hometown.aol.com/RonOuul/index.html . This an excellent set-up and survived through 4 (four) generations of designs and has never failed, and will certainly incorporated into Gen 5. The MosFETs are cheap and readily availablet.

Above is the second Generation design, it drops the regenerative braking in favour of a flywheel diode and filtering cap to help smooth the wicked spike that results by a back EMF from the inductive loads collapsing field.

Test Jig hung from ceiling

This is a cleaned-up version of the test jig, once all the wiring is in place it is a bit hard to make things out clearly in a picture.

In this configuration the motor's field is controlled by the PWM and a fine degree of speed control is possible. The bike and framing is suspended from the rafters of the shop about a foot off of the floor.

For over 2 weekends I would run the bike continuously at varying speeds for extended periods. The howl of the chain screaming at high speeds would force me to slow it down after 10 minutes. I never left the bike running unattended, as I had seen flames scorch previous boards by improper wiring or other failures at high current.

The not knowing what to expect each time it was powered up was a real buzz but very nerve racking after having parts burn such large areas of boards that a new one would have to be etched and re-built to go further.

Each minor revision would be bench tested with an oscilloscope and initially no load and smaller motors before returning to the Test Jig.

Fiberglass and molding the Battery Compartments

The basic drive system and controller seemed to well under way (they weren't) so I could make a real nice job on a couple of fancy saddle-bags to hold the 36 lbs the hawkers weighed together.

I have always disliked fiberglass as it is messy and time consuming. Any automotive body work I'd done was either with 20 gauge sheet metal or just bondo on small areas.

I was back on the InterNet for hours on end reading as much as I could about positive and negative molding and making plugs out of foam. One site gave basic instructions on making a "Hot Wire Foam Cutter". Not knowing that a sharp utility knife works just as well for the work I was going to do, I ripped apart an old heater that had a noisy motor and used a section of the nichrome element as the cutter. The handle is from a dead skill-saw wired to a 6V SLA with a switch by your thumb. The wire is held by ceramic insulators and kept tight by springs top and bottom.

This setup would slide effortlessly through 2" foam board and leave a smooth melted finish on the cut edge. I did try a 12 V cell and caused a small fire as the foam instantly ignited when the bright orange element touched it and liquefied foam down the handle.

I wouldn't waste time making one of these unless you intend to do large plugs for say a motorcycle fairing or car body panel. But now that I have one I will use it again.

Two 2" forms were glued together to form a 4" wide plug, The edges are quickly beveled off on a 45 degree angle in hopes that the glass will have a nice custom look. I made two identical plugs as the plugs are destroyed by the glassing process.

I considered the numerous methods of coating the plugs so that the epoxy wouldn't attack the plug as it was being applied. There are a number of tricks including releasing agents that keep the epoxy from adhering to the plug and blow holes that are molded into the finished product to allow compressed air to break the seal.

Eventually I settled on carefully duct taping the entire plug and coating it with a light skin of grease. This worked great, and will be my preferred method for future projects.

Battery Boxes (1 w/Duct Tape, 1 without)

Save yourself a pile of time and don't try raw glass on an untreated plug. I had to prove to myself that the tape and grease was really necessary, it is. The epoxy slowly began to soften the core. By the second layer of glass I should have known that things were going wrong.

The secondary problem was getting the plug cleaned out of the saddle-bag. The epoxy and foam had partially melted together creating a near 100% contact surface that was stronger than any natural portion of the foam block.

The following picture shows two bags, the pink one ended up being scrap. This cost me a weekend including re-cutting, glassing, sanding and clean-up.

Angle grind off all High spots and edges

As each layer would set-up and harden, I took an angle grinder over the entire surface to break off any bumps and remove any air bubbles that may have been left behind.

At this point the bags look like like a real mess, I almost scrapped them a couple of times but figured that I should see it through if for no other reason than to determine if the Duct tape and grease was really required.

Please note that I'm wearing a breathing mask, a Respiratory Therapist I work with once mentioned the high number of auto-body men that she treats from years of unprotected inhalation of paint and fiber particles during the sanding process. As obvious as this may seem I never thought of it until I heard this.

Sand between layers to break glaze (4 Heavy Layers)

Beyond the course work with the angle grinder, there was a significant amount of time spent sanding minor lumps and bumps. The secondary purpose of the sanding was to help the next layer adhere by breaking any glaze that may have formed from the epoxy.

At this point this is becoming too time consuming, but I've committed to seeing it through. In retrospect I doubt that breaking the glaze is required but I did get a nice finish as the glass was building.

In total I laid 4 (four) layers of glass and resin, which yielded all the strength I figured I'd need. As It turns-out the completed bike has been dumped and fallen over on it's own a few times and the boxes have stood up as well as can be expected.

Hack Saw Lid along tape line (get a bandsaw…)

There is no particular reason beyond impatience why that I decided to cut the lids on the boxes now. I didn't have a good reason to run a line of tape along the cut line either, I may have thought that it would help me cut a straight edge, it didn't.

The lids do mate nicely to the boxes but are far from being straight cuts. Now I definitely have to get a band saw, so many tools needed, so little money...

When I build my next Fiber Glass project I will leave the cutting to the end as now I have 4 distinct pieces to finish as opposed to 2.

Box with Core Stuck inside

Remember I did the boxes 2 different ways, one with duct tape and the other without.

The box with the duct tape was easy to clean-out. I used a large butcher knife to cut the foam into sections and simply pulled them out. The tape came off ok in most spots. I would use more grease in the future and be sure the plug is fully coated. The Resin vapors did attack the plug a bit and caused a few spots to be hardened more than main body.

The box that was untreated foam is a completely different story. The majority of the foam reacted with the resin, and was caramelized or crystallized into a solid pink chuck of plastic. I couldn't cut into it with the knife, and didn't want to bust the box to extract the plug.

Ultimately I poured gasoline into the box to eat away the majority of the foam. This was by far the worst part of the job. Luckily I had the sense to do this outside by our burning pit (unlit while working with the gas).

I should note that gasoline is a toxic substance and contact with skin should be avoided, apparently the Benzene will absorb into your skin and potentially will settle into your internal organs. As Benzene is a chemical compound that accumulates in the body and won't flush out, repeated exposure over the years will result in health problems.

The above being said, as I tried to scoop out the pink goop with rubber gloves on, the area where the resin had reacted with the foam created razor sharp bubbles that cut both gloves and hands. I would never recommend doing this for extracting a plug.

Applying 2^nd of 3 Metallized Body fill Coats (Sand between each)

The boxes didn't quite have as smooth a finish with minor lumps and bumps and even some cross hatching pattern of the fiber mat showing. Next I applied 3 (three) layers of metallized body fill, sanding each layer until I had built-up most of the low spots and was ready to move onto a glazing coat.

1^st of 2 Epoxy Skin

Keep in mind that I have no experience with fiber glassing or fabricating from scratch. Since I still had a quarter jug of resin left, I thought I would try and achieve an extra layer of durability and seal the body fill coat with 2 (two) layers of resin.

As it turns out this wasn't such a bad idea as I got a nice smooth finish. Unfortunately the high sheen made any imperfections jump-out under good lighting. At this point I figured I learned as much as I was going to and wanted to move onto other things.

A Quick Coat of Flat Black Paint

This picture is a bit confusing, the boxes and welded framing is suspended from a basketball hoop to be panted outside.

The boxes were drilled out and bolted to 3 points of channel I welded to the frame. I cut 1" squares of 1/8" rubber to keep the channel from wearing away at the Fiber boxes.

Pot Mount for Basic Throttle

For now I made this crude mount for the pot control. It is positioned so that I have quick and easy access to it with my right thumb.

The bracket is 20 gauge sheet metal (auto-body steel) that was cut into a T shape and drilled for the pot. The top of the T is bent around to form a loop that is secured with a screw clamp to the handle bar. The adjusting knob is from an electric heater (the one I scavenged the nichrome wire from for the foam cutter).

I did spend a few hours trying to rig a spring to return the pot to zero should I dump the bike, but it proved too finicky and a better arrangement would present itself at a later date.

Basic Wiring

It's spring now and there's ton's of jobs around the farm that need attention, but I'm obsessed with this damn bike!

I've got a rat's nest of wires connecting the Field, Armature, Controller and Battaries. Up until this point I'd always had a small table and extra room to assemble and test the assembly. After a couple of unexpected sparks I decide to calm down and take my time.

Connection by connection I solder and crimp wires and apply labels, double checking my work as I go.

2^nd Generation Controller Burnt-up

Here in Canada we honour the fond memory of Queen Victoria on the 3rd Monday of May with a Statutory holiday called Victoria Day.

I spent the entire 3 days to end-up with this. After painting assembling and wiring everything together, one flick of the switch undid all my efforts.

This board is the 2nd Generation controller. Any body that knows anything about Controllers would have laughed at the very idea that this would have worked in a real Load application. The flaws are too numerous to go into here.

Two Months Later... June 2001

Since May of 2001, I scraped the 2nd Gen Controller as well as the 3rd Generation. Swapped the Bosch Generator for a Briggs & Stratton Starter motor. and added a secondary gear reduction unit to lower the top speed and increase the torque.

The placement of parts has become quite a problem as I don't want to cut into the framing or re-cast the battery boxes.

With the current arrangement, there are now two freewheeling drives. The original that allows the motor to drive and the pedals to be stationary, and now the motor drive assembly has a free wheel incorporated to allow pedaling and not having to drive motor as well.

4th Generation Controller

This unit has held-up well and survived the second motor. The metal framing is from an old Computer power-supply. The MOSFets now have Major Heat Sinking with active cooling via the electric fans that blow constantly.

This unit can handle a theoretical Peak current of 180Amps at 24VDC, and close to 70Amps continuous.

The main problem with the 3 previous versions of controller had to do with how I was firing the MOSFets on and off. For the Electronics savvy, the GATE of the MOSFet was being driven by a cheap transistor that acted as a current source from the square wave pulse on a 555 circuit. The transistor had such a poor slew rate that the clean square wave was being ramped on and off. The effect of this was that MOSFet spent a considerable time in a transition state, neither on nor off. During the transition state the MOSFet has to dissipate the full load being applied as heat.

In a NO-Load situation the previous circuits performed ok, though there seemed to be a bit more heat than I would have liked.

In a full Load condition or test, the MOSFet's just weren't up to the task and would burn-up.

Gear Reduced 9:1 Drive with Free wheeling Fixed Hub Drive

The Gear Reduction Unit is a 10 speed axel with wheel and spokes removed. Again, the hub is dis-assembled and the bearings removed before welding a Second Drive Side sprocket.

Again using the lowest possible current setting on the welder that would draw a spark, there appears to be no warping of the hub as it runs very smoothly when re-assembled.

This combination yields a barely passable proof of concept design.

Yes the motor will propel the bike. Yes I finally succeeded in designing and building my own Motor Controller. The Fiber Glass experience will certainly come in handy. Lastly I feel that I'm 3 times the welder that I was 6 months ago.

Is the Bike Usable for Day in Day out use? No!

This design is too complex, too top heavy, and lacking in adequate safety features to take it out on a busy road.

Would I do it again? Yes!

I felt to keep the family life in balance, I would cool it for a while, and document what I had done both right & wrong. This Web page was the last thing I wanted to do before starting a new set-up.