Hybrid-Electric Bicycle

7/2/2002:

I first took interest in power assisted bikes when I saw a rider go by on a bike equipped with a tiny gasoline motor. It buzzed like a weed whacker and left a faint trail of blue smoke. But the older gentleman was flying along up a mild incline, peddling with easy effort. I did some searching on the internet and found out that there are several current and defunct companies that sold bicycle motors, gas and electric. Here are some links to discussion groups for electric and power assist vehicles:

V is for voltage
Yahoo Power Assist
Yahoo Zappy

Power assist increases the utility of a bicycle greatly. Travel faster, farther and carry more. Since I can vary the amount of assistance the bike can still be good exercise. I can ride in downtown DC in 95°+ F weather to run errands etc. in regular work clothes without breaking a sweat. Parking is not an issue, and I can obey traffic signals without regretting the lost momentum due to starting and stopping. The range of my bike is ~10 miles with moderate pedaling. Costs less than a nickel of electricity per charge.

I went with electric because it’s quieter, not smelly and I can carry it up a flight of stairs. I chose the conversion kit made by Currie. Factory service is nonexistant, but EVdeals is a good, knowledgeable dealer. The motor delivers more than ½ horsepower (600W), which is plenty for a bike. Assistance is limited to about 20mph, but there is lots of midrange torque. The rear wheel is driven through an adapter which mates with the rear hub.

I had to modify the system from the start because it was not designed to work with the fat frame tubes of my aluminum Cannondale touring bike. The kit is really geared toward steel frame mountain and hybrid bikes. I had to grind notches into the drive plate that holds the motor, gears and chain. I had to modify the battery mount. I added a voltmeter that indicates half the actual voltage (to gain a significant digit). I had to mount the throttle upside-down on the drop handle bars. Finally, I made a heatsink out of a 5x7" sheet of 1/16" aluminum. The stock setup is considered prone to overheating. More details.

I designed a charger that charges the two 12V lead-acid batteries individually rather than in series like the stock charger, for better balance. Schematic

10/8/2002:

The stock throttle starts the motor with a jerk unless it's used gingerly. Also, the throttle acts like a cruise control (full power up to a set speed). I designed a simple circuit that uses a push button to ramp up power to the motor for smoother starts. More details. Schematic.

 

11/26/2002:

I purchased some NiMH batteries from BatterySpace. These are 6V 4AH packs (5 "C" cells). I made two 24V 4AH strings. The strings are connected in parallel to operate the bike and charged separately.

Pack schematic. ( fuses not shown.)
I built a light, strong box to hold them out of plywood. This battery weighs 8lbs versus 20lbs for the stock lead acid battery. More details. While others have switched to even bigger, heavier batteries like the 16AH Hawker, I prefer a lighter battery for the short-range (<10mi) riding I mostly do. The whole bike weighs about 43lb.

Update 10/04: Photos below show the inside of the old vs new NiMH packs

2/24/2003:

I have set up my Currie to run on 30V with an unmodified silver can motor. I designed a circuit to switch a seperate 6V battery in series with the main 24V batt. When the throttle is switched on the main batt voltage drops under load. When it rises [as motor speed increases] above 19V the additional 6V is switched in and latched on until the throttle is released. There are several advantages to this scheme. First, low speed motor current is not increased. Second, the effect of my smooth start throttle circuit is retained. Third, the motor's low voltage cutoff will still protect the battery. Schematic Midrange power is noticably increased, as is top speed. The batteries are in a small box under the seat.

Update 10/04: This circuit has been updated with one that maintains motor voltage 23-29V with no throttle sensing.

Update 9/05: I have retired the voltage booster for now. The new SC cell battery has much less voltage sag.

5/20/2003:

After breaking spokes and unsuccesfully trying to modify the spoke adapter I have replaced the spoke adapter with the configuration that Currie uses on their complete bikes. This is hopefully the 'final solution' to my chronic spoke breakage. I now have a hub that is threaded on the left side and the USPD drive unit screws onto the hub without involving the spokes. I bought a wide range 7 speed freewheel here. Scott McGregor has been a great help. He acquired the hub and screw on adapter from Currie and pressed a a bearing into the adapter with a hydraulic press. I had to buy a complete Currie wheel to get the hub. I had a local bike shop build a 700C size wheel on the new hub. I ride on some pretty rough city streets and the spoke adapter just wasn't holding up. This configuration is much more solid. Too bad Currie doesn't offer the screw on adapter as an option with the kit.

8/20/2003:

Built a headlight with white LEDs. I used 18 of the 8000mcd white LEDs from lsdiodes.com. Three strings of 6 LEDs. Each string has an LM317L to regulate current to approx. 15mA. Reaches full brightness at approx. 22V. It's always lit when the bike is on: DRL. A sheet of frosted plastic (vellum) makes a nice diffuser.

9/4/2003:

I've been reconditioning my NiMH batteries every 3 months. For the 6V boost battery I built a "powered zener diode" circuit to. It draws current from the batteries until the voltage drops to 5V (1V per cell). The current tapers off to nil at about 4.8V. Consists of a power darlington transistor, a zener diode and a couple of resistors. Five cells are discharged to 1V per cell without cell reversal. I'll get around to posting a schematic soon. For the 24V pack I have a bench supply set to 23V and a big ol' 50 ohm 250 watt variable resistor across its output. The resistor loads down the pack and the power supply keeps the pack from dropping below 23V. I drain the pack at 5A or less until it's current drops to ~0.5A, when the resistor is dialed to max. The pack goes no lower than about 24.5V under operation, so 23V is low enough to wipe out any sign of voltage depression 'memory effect'. Performance is noticably improved after a deep discharge.

3/17/2004:

The motor died 3 days ago. Testing revealed the something shorted in the speed control line circuit. Resistance from the speed control line to ground used to be 600K ohm, then it dropped to 50 ohms. I'm guessing a shorted input transistor. Unfortunately the controller is buried in a potting compound, so I had to buy a new $140 motor to replace a $1 transistor. Thanks to Scott McGregor of evdeals.com for getting a new motor to me in two days with UPS ground. The motor has a black case with cooling fins instead of the silver can, so I'll probably not need my homemade heat sink any more. I've added a 1K resistor in series with the speed control line on my smooth start throttle, to prevent the possibility of any current surges. I will update the schematic to reflect the change. Interestingly, the new motor came with a 1K resistor added to the throttle line and not potted.

10/2004:

My second 400W motor died after 6 months. It growled and blew the battery fuses. I now have a 600W motor in its place. The new motor has a cutout threshold of 21V instead of 16V so I had to design a new 6V booster circuit. It features a power MOSFET instead of a relay. It's set to add the 6V pack in series when the main pack drops to 23V. Schematic.

I've built new battery cases for the main and boost batteries. The main pack uses two of the Batteryspace 24V 4000mAh packs. This makes the wiring simpler than the eight 6V 4Ah packs. I found that the Batteryspace 4500mAh packs have too much series resistance to work well. Had to return them.

Update 9/05: I have retired the voltage booster for now. The new SC cell battery has much less voltage sag.

I'm using hose clamps to mount the packs to the bike. I drill a mounting hole in the strap. This arrangement is much cleaner-looking than my previous arrangements with U-clamps, tie wraps, etc. I moved the boost pack down to the side opposite the motor, clamped to the rear rack. This lowers center of gravity slightly and partially counters the one-sided weight of the motor. I'm using perforated board on part of the batt cases instead of completely enclosing them, for cooling. I have a range extender pack to mount on the rear rack. This will consist of a 24V 8Ah pack and a 6V 4Ah pack.

I've designed new charger circuits for the batteries. Negative delta V detection by itself is too aggressive for charging large NiMH packs, in my opinion. Temperature sensing is the key.

My smooth start throttle now uses a big red button securely fastened to the handlebar by a hose clamp.

9/2005:

I've replaced the two parallel strings of C cells with three strings of SC cells. 9.9AH total. They fit in the same diamond frame battery box, but I had to remove the power switch and charger jack. SC cells are popular for radio controlled cars and robotics, which demand a high drain current. They are more widely available than large NiMH cells designed for high drain. BatterySpace SC cells have 2X lower internal resistance than their C, D and F cells. I measured a pack resistance of 120 milliohm, 1A / 1Sec load. What this means is much less voltage sag to the motor/controller and less cell heating under heavy load. The 24V pack also comes with a built in 10K thermistor. Blocks of 4 Anderson Powerpole connectors function as power switch, parallel strap and charge connector.

I call my charging method TCCV. Temperature Compensated Constant voltage. The algorithm is to charge the pack at a constant voltage of ~1.4V per cell, temperature compensated until the current falls to C30-C40. Then maintain a constant current at that level. There is minimal heating using this technique. Charging stops if the pack reaches 45C. For over a year I had good results charging short (the 6V booster) strings in parallel with this method. The constant voltage keeps the cells from reaching the region where current “hogging” occurs. For the longer strings of cells each string needs to be charged seperately. Implementing the basic charger is simple: just run a voltage regulator in parallel with a current source. The current source can be approximated by a resistor. Here is the schematic of my 24V, 5A, 3 channel charger and a graph of the charge profile.

4/2006

I had to replace the drive plate that carries the motor and gears. The material I ground out and the hole I drilled in it struggling with the spoke adapter weakened it. Combined with the heavy 600W motor the stress was too much. Cracks started forming. On one side I had a reinforcing strap brazed on but that cracked too. Now I have a washer on the axle to shift the assembly and with the threaded hub adapter I hope it will be OK if the drive plate presses lightly against the frame tubes. I've got bits of an old inner tube as a go-between.

4/27/2006

I bought two 400W Kollmorgen motors. The price was right. They were originally intended for Currie scooters, so they used to spin in the “wrong” direction. The supplier rewired them for reversed operation, for a small fee. On the plus side they have replaceable controllers, are a little quieter than the 600W BMC, have a smoother throttle response and a better reputation for reliability. The biggest downside is they have a lower top speed for a given voltage. I tried advancing the timing, which raised the free run speed slightly, but reduced low speed torque. However they work over a wider voltage range. The low volt cutoff is 18V. I bench tested one up to 35.5V. I went no further because I suspect the internal caps are rated 35V. I resurrected the battery box for my old 6V booster and installed a 4.8V pack composed of 12 NiMH cells 4s3p. I can mount that on the back rack in the same place as the range extender pack. With a few Anderson Powerpole connectors I can run the motor on 24V or 28.8V. For now the extender pack is still 24V. Lots of smooth power at 28.8V. I had to double the capacitor value on my smooth start throttle to make up for the higher (~15V) throttle supply voltage. These motors have a slightly different mounting characteristic than the BMC motors. They need long (70-75mm) bolts and on my bike at least I had to use washers to keep the planetary gear from rubbing against the motor face. Since the controller is on the back side of the motor the flat cooling fin is no longer useful. It's a cleaner appearance, really.

12/12/2006

I made a new headlight with S-Flux LEDs. Two strings of 10 in series. I went with the amber 10,000mcd type instead of white for a more distinctive look as well as being a little cheaper. These LEDs have a wider beam pattern so no need for a diffuser. They're bright enough to see a dark road by and the whole array uses ~1 watt (2x18mA).

5/31/2007

I haven't made an entry for quite awhile, simply because the electric bits have been working flawlessly. I now call my bike a hybrid-electric, cashing in on some of the positive vibes of the “H” word. A couple months ago I replaced my chain rings, freewheel and chain. Even though I changed the chain when it showed moderate “stretch” (1/16” per foot) cogs don't last forever. I also have new Continental City Contact tires on order.

My NiMH setup is still going strong after more than a year of daily use, but the dual cases are kind of clunky. So, I'm cogitating my next generation battery system. I'm sticking with NiMH SC cells at 28.8V. The Kollmorgen is fine at this level. I will use maybe these or these 14.4V packs and design a multi-channel 14.4V TCCV charger using switching regulators for less heat, but a good ol' 60HZ step down transformer up front. Easier to design and debug than line-op switchers.

Ride On Josephine, Ride On!