Recchetta: Recumbent Electric Bacchetta Giro

I took a Bacchetta Giro 26, added a Rohloff, Kenda airless inserts, Schwalbe Marathon Plus, a brushless outrunner motor and a 24.5cc 4-stroke internal combustion engine…and in doing so, turned this bike into a viable primary vehicle for personal transportation.

Anyone with a hand drill, a hacksaw, all the necessary parts, and a bit of mechanical skills, could easily build this on a Saturday and be whizzing through town by Sunday morning. I took a little bit longer, and cheated with a CNC.

This bike is also featured on EVAlbum.

• ~15 lbs - all equipment, minus battery (including the gas motor)
• 11 lbs - weight of 24v/20ah LiFePO4 (ping brand) battery
• Assembled with roughly $450 in parts and materials (minus battery & gas motor; see details below)
• Gasoline engine's emissions are nearly half that of your typical highway legal motorcycle.

I've thrown a blog posting up about the gas setup.

In-short, I've now completed two long-distance rides (256 miles & 302 miles) over the Oregon coastal range via Hwy 30, 47 & 101…and a variety of in-town gas-assisted rides. While I'm averaging roughly 250MPG, my biggest complaint is a need for something better than the .5 liter built-in fuel tank; stopping to refill from my MSR bottles, every ~33-35miles, gets old, fast.

In general, “I need a larger chainring” was the inital thought as I turned the electric's throttle up on my first electric-assisted ride, while my legs frantically “twitched” at over 115 RPM just trying to keep up. Remember; this system was designed to facilitate a power-assist for an existing bike, not to create a lightweight motorcycle. As for the rest…well, a silly ear-to-ear grin just refuses to go away

• I rode about 20 miles, from Portland Oregon's Burnside Bridge eastward to 143rd & SE Stark, and then returned home to 20th and W Burnside. For a trip which, on a good day, normally takes 45 minutes. The power-assisted ride, with ~20MPH gusty gorge headwinds, was completed in just barely under 20 minutes!
• Earlier that day, I rode from my office near 6th & E Burnside, to TerraCycle's headquarters, and back.

The setup runs exactly how I wanted; power-assisted continuous pedaling, above 15MPH, without significantly straining myself.

This section covers the electric-assist section of this build. The goal of the e-assist, is to facilitate additional power/speed for in-town trips; less than 30 miles per day.

There are also two electrical configurations having to do with how the wires from the windings are connected to each other (not their physical shape or location). The delta configuration connects the three windings to each other (series circuits) in a triangle-like circuit, and power is applied at each of the connections. The wye (“Y”-shaped) configuration, sometimes called a star winding, connects all of the windings to a central point (parallel circuits) and power is applied to the remaining end of each winding.

Efficiency is greatly affected by the motor's construction, the wye winding is normally more efficient. In delta-connected windings, half voltage is applied across the windings adjacent to the undriven lead (compared to the winding directly between the driven leads), increasing resistive losses. In addition, windings can allow high-frequency parasitic electrical currents to circulate entirely within the motor. A wye-connected winding does not contain a closed loop in which parasitic currents can flow, preventing such losses.

From a controller standpoint, the two styles of windings are treated exactly the same, although some less expensive controllers are designed to read voltage from the common center of the wye winding.

A motor with windings in delta configuration have lower torque (by a ratio of 1.735), and higher RPM/V (ratio: 1.735), than WYE windings.

A motor with windings in WYE configuration have higher torque (by a ratio of 1.735), and lower RPM/V (ratio: 1.735), than Delta windings.

I've been experimenting with three battery packs.

All battery packs utilize a LiFePO4 battery chemistry; one of the newest mass-produced battery chemistry, and also an ideal type for EV use.

These battery packs retail for around $150, but eBay often has these for as low as $50…I picked mine up for $70 (including shipping).

The packs utilize A123 brand LiFePO4 batteries; arguably the best brand available.

These LiFePO4 batteries retail for around $17/cell, but I got in on a group buy, and grabbed them for a great price. They remain in a box, untested…testing reports due soon

PingPing, the manufacturer/seller of these batteries resides in Singapore, and while the idea of spending nearly $500 for an “unknown”, this specific seller has an excellent reputation throughout the DIY EV/eBike community. As a result, just five days after sending “Ping” $419 via PayPal,I had my very own 24v/20a LiFePO4 battery pack to play with.

I selected a Brushless Outrunner simply because:

1. The RC modeling market has been flooded with this highly efficient motor design over the past couple of years.
2. Emergency replacement parts are simply a trip to the nearby hobby store.
3. These motors efficiently pack a tremendous amount of power inside a very small package.
4. It's brushless, which means less wear & tear and a minimal amount of maintenance.
5. I wanted to use an AC three-phase motor simply because I sought to incorporate Nikola Tesla's technologies.

The rear wheel pulley is a major facet behind the secret to my overly-simple drive design. Put simply, the pulley is manufactured and sold by a US company which only markets kits for adding gasoline motors to a regular bike. The key point is, most of these tiny little gasoline motors turn at upwards of 7,000RPM…and so could an electric motor

Originaly designed for use with gas motors that'll spin at roughly 7,000 RPM, I've re-purposed the wheel-mounted pulley to instead spin under the power of an electric motor. With the belt tensioner I've designed, there's zero slippage.

The Golden Eagles drive ring was designed for small ICEs, so I don't know how it'd handle the torque of really hard drag-race style acceleration from an electric engine, but…there is a way to find out

As for me, I just wanted to move around town (friends, clients, office, social) within “reasonable” timeframes, while still cycling, and possibly keeping up with automobile traffic - not competing with it, and not being regularly startled by yet another overhanging pickup/delivery mirror whizzing overhead. From a variety of discussions at a recent OHPV meeting, a surprising number of others sought similar goals, their concerns were a hub motor's weight, presumed inefficiency and general lack of friction-free freewheeling, and the sticker-shock, weight and presumed complexity of EcoSpeed's well-engineered mid-drive product. One particular person wanted it all, just like me; a Rohloff AND a sub-$1,000 rear power-assist with a brain-dead simple drivetrain exceeding 90% efficiency.

Using half of a TerraCycle EasyReacher rack, I've been able to very easily secure the front-half of my motor mount to the bike. During the manufacturing process, they drill & tap a few anchor holes for their CNCing process. After manufacturing, these holes are, in nearly all cases, simply ignored; left totally unused.

After a short discussion with the geniuses behind this rack, I was assured nothing terrible would happen if I ended up utilizing these otherwise benign holes. So-far, they're assertions have been dead-on…and this rack now not only holds the eletric setup, but there's a ~10lb gas engine hanging off the electric's initial mount.

Overall, since my RANS Rocket also had one of these wonderful racks…the choice to purchase another one was a no-brainer.

Currently I'm using the 13-tooth motor pulley, with a clutch bearing press-fit into its bore.

By sheer luck, I managed to catch the folks at TerraCycle during a slow moment, and given my lack of access to a suitable drill press, they were happy to help bore out the pulley, and press-fit the bearing.

I've designed and manufactured a high-torque sealed sprauge clutch based freewheel assembly, combined with a 12-tooth pulley. It's working great, and photos are due soon.

Now using a high-torque common-shaft freewheeling assembly; combining both gas and electric motor outputs to a single shaft.

While not to dismiss my efforts, to give credit where due; the idea of a cheap yet very strong motor mount, that uses nothing but a simple square tube (an incredibly obvious idea in hindsight), was most likely first put into practice by Matt Shumaker. If you have access to a CNC, and time on your hands, his "E Box" is one slick setup.

For high power; drag racing, or simply breakneck bicycle acceleration, Matt is exactly who you'd want to watch. Last I heard, his Actionbent Midracer easily covers 0-30MPH in well under 5 seconds (and it's capable of faster, he's just limited it to that rate). His design uses an RC style brushless/sensorless outrunner, LiPoly batteries, a Castle Phoenix HV110 w/ ~$10 in upgraded capacitors, and a belt drive reduction off the motor's shaft, to a BMX freewheel that spins a normal chain sprocket.

So, ignoring my mother's champion drag racing blood, I'd get my hands on a power-measurement device such as the CycleAnalyst or eLogger. After you have a good idea what the power requirements are for your desired rate of acceleration, then its simply a matter of going online and cherry-picking your equipment from thousands of RC model based options.

Though, to help narrow your choices, I'd suggest two things:

1. Stick with Scorpion motors, they're apparently the only brand which uses N50 NdFeB magnets (200C/392F temperature rating), while other brands come close, the highest I believe they go is 180C… As you well know, if your motor gets too hot, you'll not only burn out the windings, you'll demagnetize its magents. FWIW, my motor was not even warm to the touch, after ~7miles of nearly all green lights, and a continuous ~28-29MPH on a mostly-slight uphill climb. Here's a comparison chart of the major brands.
2. Consider sticking with Castle Phoenix electronic speed controllers. While there's plenty of less expensive options out there, Castle Phoenix has a great 1yr warranty, and an unbeatable out-of-warranty offering: http://www.castlecreations.com/support.html For the really power-hungry, Castle Creations Special Projects has speed controllers rated for 85A@90V for $500, and 200A@90V (continuous 18KW=24HP!!) for $750. Castle is a proven “high-end” brand in the RC market, and with fully-programable ESCs coupled with a great warranty, you'd be hard-pressed to go wrong with them.

Not “flaws”…areas for improvement

There's three main options on the market at the moment

1. Eagle Tree eLogger
2. Cycle Analyst (aka: Braindrain)
3. Watt's Up Meter

I grabbed an Eagle Tree eLogger; most features, smallest package, and a really neat display setup.

With the gas/electric combined setup, my initial idea of a handlebar-mounted potentiometer is no longer viable. As the gas motor has a cable-driven throttle, and the electric throttle needs to be integrated into this.

So, instead of mounting a fully sealed cermet potentiometer to the handlebars and running the wires into the PWM signaling device (aka: servo tester). I've re-purposed my front derailer's gear shifting lever, run the cable over a machined “disk” that's mounted on the potentiometer. From there, another cable runs to a spring (for automated throttle returns), and a final cable runs to the gas motor's carburetor. The combined result, is a single throttle control, where, without any mechanical tweaking, will simultaneously “give” power to either the gas or electric motor.

Allied Electronics carries the potentiometer I needed; 4.7K ohm, 3W, sealed cermet “PE30”. Mfg part # PE30L0FL472MAB (click for detailed specs). The disk and cable/spring assembly is undergoing a slight design revision - so no photos yet

I had a bunch of belt tensioner ideas, but even the one I've currently engineered onto the bike is no optimal. Of-course, one facet of long-distance bike rides is plenty of time to think. I've re-engineered by belt tensioner again, enabling me to easily (and safely!) engage/disengage while rolling. The catch; it simply integrates a piece of old bike brake cable. Yes, I'll post pics the moment its done.

In this design, there was one significant weak spot; the one-way/clutch needle bearing:

1. The needle bearing setup was press-fit into the pulley, and while the TerraCycle folks used loctite designed for press-fits, nobody (including myself) expects this to hold for longer than a few hundred miles. The setup was largely intended as a proof-of-concept, before I spent even more money on this project; if this bearing setup fails, I'll be out a whole ~$15 in parts. Under the circumstances, it's an expense worth tolerating for the information I've already gained.
2. The bearing itself is only rated for 1.5Nm of torque. From a dead stop you're asking for a lot more than 1.5Nm from it. The saving grace for me, is that my design is only intended to provide a speed-boost when I'm already moving at ~15+ MPH, at that speed, with this motor/battery setup, I'm only asking in the neighborhood of 0.6Nm from the bearing.

The image to the right is a 5mm HTD pulley mated with a 12mm sprag clutch-bearing. Yes, I finally got access to a small lathe (bought it).

I'm still considering whether to turn part of a 12mm hardened rod to 6mm, for press fitting into the motor… The alternative sounds like an easier short-term setup though.

These folks supplied my metal:

Metric roller clutch spec sheet

VXB's one way bearings list http://www.vxb.com/page/bearings/PROD/kit949 https://sdp-si.com/eStore/Direct.asp?Exp1=41799&GroupID=227 http://store.todayshightech.com/e7070197.html http://gbgearworks.com/ http://www.torquetrans.com/pulleys/timing/index.htm http://www.gears-reducers.com/timing-belt-pulleys.html http://www.onviollc.com/dotnetnuke/Pulleys/tabid/73/Default.aspx

There's one major flaw/issue with centrifugal clutches; many (including the one on my EHO25 lack any freewheeling provisions. Meaning, if I'm cresting a mild climb at ~20mph, and wanted to disengage the engine for the downhill side, I actually have to brake to ~15mph before the centrifugal forces lighten enough to disengage the clutch…even if I've turned off the motor already. I hate braking.

The fix is easy, and will be integrated into the next generation of the drive-train; the a freewheeling assembly

I cant find a CVT transmission small-enough, so I'm putting my CNC to good use. The primary purpose, is that while cruising I'd like to keep the motor RPMs down, and when hitting really steep climbs, the motor is not spinning fast enough and therefore begins lugging…the power is there, if I could only get the RPMs up…I want just a bit more speed. Reports, pics and writeups to follow soon.

While I still continue to play with the idea of combining the drivetrain into the existing “human side”…I'm not totally sold on the idea yet; seemingly too much complexity, drag, costs, and other sillyness. Remember; one of my original design goals was to build something using as much off the shelf parts as possible.

• For the main battery, I'm thinking a deans-plug style kill switch.
• A coworker suggested a watersports equipment killswitch lanyard.

These are photos of an older design, that's no longer in use.

I'm using a 24.5cc EHO25-R/S Robin/Subaru 4-cycle gas motor, and an 11-tooth pulley. Let's just say…that's plenty of power to push a recumbent cyclist, with an ExtraWheel trailer, uphill at an ear-to-ear grinning pace (yes, while also pedaling too). One example; I now climb one steep mountain pass at ~10mph, when (on a good day) I'd human-power it at only ~2-3mph.

Compliant with EPA Phase 2 (Class IV) and CARB tier II emission regulations.

In translation, this means that the EOH25 engine is EPA-permitted to exhaust less than 50 grams of HC+NOx/CO per 1kW/hr of generated power.

Put into perspective, comparing the average EPA emissions requirements of the bike VS common road vehicles:

EPA Phase II handheld engine classes are as follows:

Until now, because of the non-road engines’ relatively low overall contribution to air pollution, emission control for these engines has not been a major design consideration. Consequently, these engines are not as clean as highway vehicles, which have been subject to regulatory controls for more than 20 years. Emissions from nonroad engines contribute as much as 15 to 20 percent of pollution in cities across the United States. Those emissions, described above, include hydrocarbons, nitrogen oxides, carbon monoxide, and carbon dioxide.

Hydrocarbons (HC) are unburned or partially burned fuel molecules that react in the atmosphere to form ground-level ozone, a major component of smog. Some hydrocarbons are toxic and may cause cancer or other health problems. Hydrocarbon pollution from nonroad engines also occurs as fuel evaporation when gasoline vapors are forced out of the fuel tank (for instance, during refueling) or when gasoline spills and evaporates.

Nitrogen oxides (NOx) result from subjecting nitrogen and oxygen in the air to the high temperature and high pressure conditions in an internal combustion engine. Nitrogen oxides react with hydrocarbons in the atmosphere to form ground-level ozone. They also contribute to acid rain.

Carbon monoxide (CO) is a colorless, odorless, poisonous gas that results from incomplete fuel combustion.

Carbon dioxide (CO2) is the ultimate product from burning carbon-based fuel including gasoline. Carbon dioxide does not directly impair human health, but it is a “greenhouse gas” that contributes to the potential for global warming. As engine fuel economy declines, carbon dioxide emissions increase.

http://www.ebikes.ca/drainbrain.shtml http://www.electricvehiclesusa.com/product_p/th-bnl-103.htm http://www.eagletreesystems.com/MicroPower/micro.htm

Measurements from http://www.sheldonbrown.com/cyclecomputer-calibration.html

Typical “resting” cadence is around 75 RPM

Using the online calculator.

For a 26” wheel, with the reduction's output shaft at 240RPM & the Rohloff in high (ratio: 1.47) gear:

For a 26” wheel, with the reduction's output shaft at 240RPM & the Rohloff in direct (ratio: 1) gear:

For a 26” wheel, with the reduction's output shaft at 240RPM & the Rohloff in low (ratio: 0.28) gear:

One should never reverse the polarity on their speed controller's battery terminals. Especially after making modifications that they knew would void warranties…

Though, after that little lesson, I humbly fired-off an email to Castle Creations, with the hopes of a positive response - especially knowing they purported a very customer-friendly out-of-warranty plan. Their response was straightforward enough:

On Nov 25, 2008, at 10:19:
Gregg,

Yep, send it in. You could take advantage of the non-warranty option even if you
smashed it with a hammer, lit it on fire, and then sent in the charred remains. ;)

Joe Ford
Product Specialist
Castle Creations

I still struggle with the idea of using a gas engine, though the reality is, anything else is wholly-impractical if I'm looking for a power-assisted speed boost; hauling enough batteries to take me along my most favored route (Portland Oregon to the Oregon coast).

My cyclist “resume” can boast:

• Being an Oregon Randonneur; having successfully completed up to 300km long organized rides, and a 500km solo ride through Oregon's high desert.
• Commuting for just-over 20 miles/day, for roughly a year, in-between downtown Portland Oregon and outer-SE Portland (near Gresham).

With that variety and volume of on-road experience, one thing is now absolutely-certain; I have no doubt that I'm safer with the power assist:

• Passing vehicles give more room.
• Passing vehicles are generally more cautious & courteous overall.
• I was often “smoked out” by some redneck-driven diesel pickup…that has never happened (yet) while running power-assisted. I think it's because I've confused the poor guys; normal cyclist only move that fast on downhills, I move that fast all the time.
• In downtown Portland, and on many major roads (Burnside, Glisan, etc…), I can often travel with traffic, instead of drivers being forced around me (and my speeds are reasonable at ~20MPH on the flats).
• I've had more pickup drivers show a genuine “how do I do that” interest in my bike (at traffic and food/fuel stops), than your typical “eco” crowd of hybrid or small-car drivers.

Before my power-assist days, I'd crashed the bike a few times. Given how unpleasant road rash is, I had some cyclist shorts custom-tailored. They're simply the inverse of most motorcycle chaps; where there's normally no leather in the seat/butt area, I've had it added, and where there was normally leather, I've instead had spandex put in place. They work great, and are comfortable for upwards of 200-300 miles/day, in ~70-80F temperatures.

Given my propensity towards always donning high-visibility clothes and my always-running yet blindingly-bright Dinotte headlight and taillights, I've been accused of looking like a ”…Christmas tree on crack…” Oh well; as one friend put it - it's cheaper than an ambulance ride.