Powered BikeE2 Recumbent Tandem, Long Description and Video


Local time
12:30 PM
Mar 24, 2011

Why I built it

I wanted a motorized bicycle that both my wife and I could ride together. We live in south St. Louis, and 90% of our commuting is within 10 miles of the house. Our part of town also has decent bike lanes, and the city and state have reasonable regulations on motorized bikes. We go camping in the midwest as well, pulling our little Casita travel trailer behind our truck. I wanted a vehicle that could pull us both ~20 miles on campground and country roads (more with gas stops), which is enough to get from many campgrounds to recreational spots such as small picturesque towns. Also, I'm a paramotor pilot and periodically go to fly ins, where we camp and I fly. I wanted a vehicle where my wife could befriend another wife and they could take rides and maybe sight see. Finally, I'm a rotating petroleum engineer, and I work in the middle east, or wherever, for 4-6 weeks and then get 4-6 weeks off. So, unless I meet my wife in Europe or California or wherever on my days off, I'm home in south St. Louis. Building this bike was a good hobby for me.

Why I chose the BikeE2

I chose to modify the BikeE2 mainly because I already owned it. I bought it in about 2000. BikeE is out of business now, but their recumbents were known for their durability. I bought a single seat BikeE when I bought the E2 and have been putting at least 50 miles a week on it since then, with no problems at all. The BikeE2 also has good brakes and a 20" rear wheel, which helps in climbing hills.

Power choices (general)

After thinking of many power train combinations, I decided I wanted a tiny 4 stroke gas engine, coupled with an automatic transmission with deep gear reduction. I chose gas over electric because I wanted the range and I wanted to avoid the expense and problems of expensive batteries. I would have liked the lower noise levels of electric power, but I'm going to try and design some of that out. I wanted an automatic transmission so my wife could ride it more easily, and because many states require it for the vehicle to be called a motorized bicycle. And I wanted deep gear reduction because we need it to carry 2 people up and down hills, and we probably have no business riding the BikeE2 at more than 21-22 m/h anyway.


I chose the Honda GX35 because it was small, but big enough, because it was a 4 cycle engine, because it has forced air cooling, and because it is a Honda. The engine is rated at ~1.3HP at ~7000 r/m, but it has good, and increasing torque, from ~4000 r/m, when the clutch begins engagement to over 6000 r/m, and good power to its speed limit of ~8000 r/m. I wanted the emissions, fuel efficiency, lower noise level, and greater durability of a 4 cycle engine. I also wanted the forced air cooling, because I wanted to be able to climb long steep hills at low speeds, with 2 aboard. And I wanted to buy a Honda because of Honda's reputation for good design and reliability. Many of my fellow expat engineers are high mileage bikers in places like Thailand and Kazachstan, and they uniformly swear by Honda reliability.


I chose a NuVinci Developers Kit, with a NuVinci 171 CVT hub, and electronic transmission control. It is a complicated system, and I have only recently gotten it to where it is totally reliable in real world use. A mechanical CVT would have been simpler and cheaper. But I wanted the ~3.6/1 gear range that it offered, and the availability of power and economy shift modes. All of the suitable mechanical CVT's I found had about a 2.0/1 range.


I chose a Staton 18.75/1 "tri-hybrid" gear box because Staton has a good reputation, because I needed that much gear reduction for good performance, and because I could use the intermediate gear as a PTO for battery charging.


The bike can speed up to 21-22 m/h, and can coast faster. I can freewheel either the powertrain or the pedals, meaning I can use them independently. By the numbers, the engine clutch engages at ~4000 r/m, with a torque of ~1.2 ft*lb. That torque increases to ~1.3 ft*lb at 5500 r/m. That is important, because it means that any hill we can start on, we can fully engage the clutch on. Maximum power is ~1.3 hp, at 7000 r/m. Torque is increased by a factor of 18.75 in the gearbox, by another 2.08 between the gearbox and NuVinci hub sprockets, and by another 2.0/1 in the Nuvinci low gear setting. That results in rear wheel thrust at clutch engagement of over 115#, which in turn gives us the theoretical ability to climb a hill of over 18% grade. We have not yet had a hill we could not slowly climb, at least here in south St. Louis Italian Hill area, using only engine power. Acceleration on the other hand, is quite leisurely. My power/weight ratio with 2 aboard is less than 1 hp/400#, compared to about 1 hp/12# on my store bought 2 wheeler. As far as fuel consumption, I checked it riding solo, and I approached 140 m/g. I assume that will be ~100 m/g with 2 aboard.

NuVinci Detail

The transmission is a tilting ball variator, which uses what drilling engineers would call a shear thickening, non Newtonian fluid to lubricate when needed and to stick the balls to the discs when needed. It's has the same behavior, or rheology, as the stuff they ran in AMC Eagles way back when. Wikopedia actually does a good job of explaining how the CVT works. Functionally, it is a very efficient CVT with a great range, but with strict power and torque transmission limits. My power train stays under those limits, for the 171 version of the CVT, but would tear up the newer 360 version.

The 171 CVT gets 2 bad raps, neither totally deserved. People say they slip, and that you can't shift them up under load. As for slipping, it happens because they are twisted too hard, or overpowered. But if you read the specs, they are still strong enough for most motorized bicycles, as defined by most state regulations, just not for 5+ hp builds they failed with. They also shift up well under load, IF the loads are within the specs, and IF you have plenty of shifting power. You can't shift when you are transmitting 5 hp, with a twist grip shifter. You can shift up under power, when you are transmitting 1+ hp, with an power shifter. That's what I do, and it works flawlessly.

The trick of shifting under load is shift power availability, which you can really only get practically with a full charging/storage system. The Developer's Kit power shifter is actually a modified car window opener, which pulls a rod to move the shift balls. To shift up under load, you need at least 12 volts, and 5 amps (50-60 watts) of power. After several tries and fails, I ended up using a YAK54 permanent magnet alternator, a $3 Radio Shack bridge circuit, a KINTEX solar charge controller, and a cheap motorcycle battery. The alternator is spun from the 5/1 reduced intermediate shaft on my gearbox, which when sped up some with pulleys, turns the alternator at 1800-2800 r/m, proportional to engine speed. The alternator can easily produce 14+ volts and 5+ amps throughout much of its r/m range. The battery is required to shift down when coming to a stop. The gearbox is not turning then, so the battery has to provide shift down power. Of course the battery stays charged through out the ride by power diversion of the alternator power through the solar charge controller. To see why a hand shifter doesn't work, we need to realize that we need ~50 watts, for at least 4-5 seconds, to shift from low to high under power. The hand shifter is geared to do this in less than 1 second, requiring more power and torque from your wrist than we can deliver with both of our legs while riding a normal bicycle. Doesn't compute.

My shift controller has 2 tables in it of desired shift position v rear wheel rotational velocity. Each table is made up of 64 ordered pairs, from stand still to top speed. One is for an "economy" mode, keeping engine speed at 5500 r/m, the max torque speed, from ~4-15 m/h, and then speeding up in high gear. The second, "power" mode, keeps the engine speed at ~7000 r/m, the max power speed, from ~5.5 m/h to 19 m/h before speeding up in high gear. The shift tables were built by me, using a simple excel workbook tool, and then loaded into the controller. I can switch from one mode to another with a toggle on my handlebars. The job of the controller is to constantly compare my desired shift position, based on the rear wheel velocity as calculated from the rear wheel speed sensor, to the actual shift position, as measured by the shift position sensor. Then it tells the shift motor to shift up or down to reduce the difference, and sends it the power to do so. The controller itself is somewhat fragile, so I have installed the controller, the power controller, and the bridge circuit, in this high tech baby wipe box. The box is sprung by the bike shock, so it bounces less than the engine and gearbox.

Other accessories;

My mini tach, for some reason only shows half of the engine speed. That's probably because the wire is so long from the tach to the engine coil wire. I have given up on getting the right output, but it reliably show half the correct number, so I can still use it.

My front light is a rechargeable, and is more powerful than the Shimano gen hub light I originally installed. I put that on my single seat recumbent. My rear light is just a conventional blinker.

My handlebars are aftermarket, and are wider than the original BikeE2 bars. They are more comfortable, allow for more control, and allow me to install more Peewee Herman stuff on them.

My throttle controller uses a very long 10' cable, with 2 10" cable extensions, to get to the Honda engine. This necessarily added more cable friction, which I countered with this Mickey Mouse homemade throttle spring strengthening system. I have to twist somewhat harder than normal to add throttle, but it closes crisply. Any hints on improving this system would be welcome.

My basket is small, and hard to cram things in. I didn't want a front basket, to minimize turning moment of inertia. I also didn't want more weight in the back. I often bungee things like 30 packs of beer on the rear seat if I am riding solo.


Hill climbing ability
Rides low, feels safe.
Many quality components
Turns heads
St. Louis cops have seen it many times and have ignored me.


Rides rough with 1 aboard
15-20 miles between refuels, with 21 oz tank.
Need to further muffle it for the rear rider.
When I get my rear flat, I will have a major job repairing it
Heavy rear end, mainly from power train mounting hardware. Chain pull is nearly 300 lb, and the power train mounting system had to be re-engineered twice to reliably keep the power train from twisting off the bike.
Last edited by a moderator:
Also, a video of the BikeE2 on a test stand/ladder, trying to show the autoshift. This video suffers from the John Wayne western wagonwheel effect, but I think you can get the idea.
Last edited by a moderator:
Finally, a ride from my wife's helmet cam thru South St. Louis. I know I should have pointed it down a little more, but you can still see and hear us.
Last edited by a moderator: