The bike getting some testing before the installation of the chain
Before coming to any conclusions, let’s review the cost of this project and its variations.
Cost (in money & time): Achieving a balance of expenses in currency and time is important, but it depends on your situation. Originally, I had more time than money. Now the balance is tipped the other way. Still, some activities are just not worth the effort unless you are truly poor.
Bicycle Prototype – Long Wheel Base
- 1 x Cheap Mountain Bike – free, from college move out
- 6 feet 1/8″ wall square tubing – $15
- 1 x 20″ wheel – free (found in a scrap pile)
- Exercise bicycle handlebars – free, found on trash day
- Office chair – free (it was headed for the curb side trash pickup anyway)
Total cost: $15
Build time: about 16 hours
Add electric motor, etc, NO batteries (w/ shipping)
- Kollmorgen 300W with integrated controller – $45 (electricscooterparts.com)
- 80 tooth #25 sprocket – $15 (tnc scooters)
- 15 tooth #25 sprocket – $12 (electricscooterparts.com)
- Sprocket to bicycle freewheel adapter – $10 (http://tncscooters.com/product.php?sku=103210)
- 120 links of #25 chain, heavy duty – $25 (electric scooter parts)
- 150 links of #25 chain, heavy duty – $30 (electric scooter parts)
- Magura twist throttle – $55 (electric scooter parts)
- Nuts and bolts – $5 (Ace hardware)
- Batteries (TBD)
Total cost: $197*
Build time: 13 hours. Mostly mounting the sprocket on the wheel and making the motor mounting plate. Note, I was not aware of the time that I could purchase an adapter from the sprocket to a bicycle rear wheel for roughly $10. With this, I would have saved about 5 hours of time. The motor mounting sucked up another three hours. Once you figure it out though, it should be easier to repeat.
Motor Controller
- 6x MOSFETs – $4 each – $24
- Atmel microcontroller – $3
- AVRISP MkII – $38
- 3x h-bridge drivers – $3 each – $9
- other assorted pieces – $10
Total Cost: $46 + $38 (one time)
All parts from Digikey.
Oxy-Acetylene Welding Rig
- Two regulators, torch, cutting attachment, #2 tip, acetylene tank, small cart – $60 (used)
- Striker – $10
- Tip cleaners – $5
- Welding mask & gloves – $5 (garage sale)
- 3 c-clamps – $5 (garage sale)
- oxygen tank – $100 (used)
- Cost to fill tanks – $70
- 3 pounds mild steel welding rod – $12
Total cost: $267
All parts bought on Craigslist or Airgas supply
Other stuff
- 250W geared DC motor -$45
- Kollmorgen 300W with blown controller – $10
- BMC 750W motor – $160
- Other h-bridge driver chips – $12
- SOIC to DIP converters – $10 and about 1 hour
- Parallel to USB – $25
- Two parallel AVR programmers – $20 and about 4 hours
Total cost: $282
Cost of this project so far: $833
Other Options
Just for comparison, the other option I considered at the beginning was to use a cheap hub motor. In that case, I would have saved hours of motor mounting time (this time is actually greater as I spent five of those hours working with Ben and we ended up having to use his mill– not a common garage tool by any means). Instead of paying $185 for the motor and accessories, I would have spent roughly $450 and gotten a kit. The questions I have to ask myself when evaluating that option are: Is ~$300 worth nine hours of time? Does the setup I have provide something that the hub motor cannot (weight, performance, gearing)?
As you can see, unless you have a fairly well stocked garage, you may be better off going with the hub motor from a cost/time perspective, unless your needs cannot be addressed by the hub motor (they can be quite heavy, for example).
Lessons Learned
3″x1.5″ square 1/8″ tubing makes for a very strong boom– and a very heavy one. The easiest way to save weight on the bike would have been to use 1/16″ wall tubing instead. It would have probably done the job just fine. When choosing tubing, you can use thick walled by small diameter tubing or large diameter tubing with a thin wall. You get different performance from each type. Larger diameter with a thinner wall is the way to go. It ends up being strong enough, plus, you can stuff electronics and other bits inside. That will keep people from walking off with them!
Tiller steering. It stinks. Since the bike has such a long wheel base, it’s not too bad, since the bike is often too long to make abrupt maneuvers. However, it causes more problems than ackward slow speed steering. The handlebars interfere with the riders legs once they get turned far enough. Cables headed to the back of the bike can’t be located on the handlebars. The routing doesn’t work. Running the cable to the head tube and then back would require an 180 degree turn that the cables just don’t like. That meant welding a bar to the boom to mount the shifters, and having no rear brake.
Each weld costs 30 minutes. Between the time jig the work pieces, get gloves and a mask on, grab some rod, and then clean up the weld, it really does seem to take about that long. Minimizing the number of welds is key to actually building a ridable bike (rather than one that sits half finished in the corner).
No.25 chain is finicky. Even a misalignment along the plane of the sprocket by one chain width caused problems. The motor mounting thus needs a side to side adjustment more than it needs lots of tensioning adjustment. Bolting the mounting at the bottom with two bolts and then having one bolt on top with a lock washer for tensioning worked really well.
Brakes are a pain. Calipers are OK– mostly, they just need the cable to be adjusted correctly. V-Brakes need to aligned in two different axes and then the cable has to be right too.
18 degrees head tube angle (from vertical). I’ve tried going to the extremes in both directions, and bad things happen. In either direction, the steering gets very sensitive and twitchy. With a greater angle, you can fix it by adding an offset to the wheel, but the amount can easily get a little ridiculous.
Cable steering is too complicated for a bicycle. And when it fails at speed, it’s not fun.
Conclusions
It hasn’t quite sunk in that I finished this bike. I thought about it for several months before I even took the project on; and then over the next few years worked on it sporadically. A major part of the problem was a lack of tools, so I spent a lot of time and money fixing that problem. Of course as I worked on that, the bike wasn’t getting worked on, so it felt like a never ending project. Once I had what I needed most (cordless drill, dremel, sawzall and oxyacetylene) all it took to finish was a reasonable deadline. I had to work hard on the bike the last few days, but I finally got most of it finished. It only took a few more hours of mucking with the chain alignment and readjusting the brakes to make it ridable.
Let’s see, I started sometime around January ’03 and finished the second week of December ’07. Well, only about 5 years 
Mind you, I did ride the recumbent bicycle a good deal without electric power!
One other consideration: When I started this project I had very little money. I like to avoid waste too. This worked out well with the speaker project I did. Let’s see how it stacks up for the electric bicycle. I have seen lots of builds on the internet that require chopping up multiple bikes and using a limited amount of pieces from each. I really do not like this approach. If we’re making one bike, we should only have to chop up one bike. Any piece that can be reused without too great an effort should be reused. Some people throw things out left and right just because they are easy to replace– I don’t.
Number of donor bikes: 2. This was only because we messed up the front end when welding and rewelding to get the right head tube angle. Remove the bearings before applying heat!
Brakes: fine to reuse the pads. They get a little stiff but work.
Cables: toss the old ones. You can buy a box of them for $7 at Walmart. When the cable rusts up, it’s difficult to get it back in the sheath even after cleaning it.
Chain: unless its really rusty, this isn’t too difficult to revive. I’ve even revived a totally frozen chain by soaking it in WD40 and then working the links with two pliers during my college days.
Seat: a chair destined for the dumpster. Works well!
Handlebars: Also destined for the land fill. One more thing that didn’t get thrown away…
Well, not bad. Time for the next bike!