After being disappointed by the slow acceleration, low top speed of a wilderness energy / aotema 24V brushed hub with SLA, I decided to go all out and put e-bikekit (Nine Continent?) motors in both the front and back wheels, along with a 48V 20AH ping battery. I'm not really going for max speed, I just want extra torque on hills and acceleration. And honestly I'm more into the building and the engineering than the riding, if you know what I mean.
But now that I've built it, the bicycle frankly scares me a bit: it's hard to control the acceleration, and the weight is just over what is reasonable for a bike. The battery was too wide to put in the triangle and I didn't think of asking Ping to split it into two halves for panniers, so it's altogether on a back rack. I welded a custom rack for it that puts it as low to the wheel and as near to the seat as possible, but it's still a high, rear weight.
While I waited for the front torque arm to come through the mail, I got impatient and tried to ride it without a front torque arm (it has torque arms on the rear). I figured as long as I didn't hot dog on the throttle too much, I wouldn't spin the dropouts. BIIIG mistake. It was actually ok for a few turns around the block. Then I stopped to check the mailbox, and as I was getting going again, I lost the front wheel. It all just sort of fell apart underneath me, like in a western movie when the horse is shot and stumbles under the rider. Luckily I wasn't going fast so I didn't go flying or anything, the bike just fell out from under me.
This got me thinking: Because dropouts open straight down, when the axle spun out and spread the dropouts open, the axle rolled the dropouts right over itself like a log roll. If bicycle dropouts were originally designed for powered hubs, they could be made safer by having the dropouts wrap around the axles like an upside down question mark. The direction of curl could be set up to be opposite the torque of the motor, so that when they spun they'd tend to bury themselves deeper instead of escaping. If I had a machine tool capable of routing a curved question mark shaped path in a steel bar, I could make these.
The torque arm actually came in the mail the next day. I hammered the dropouts back into shape, but unless I'm imagining things I think they have a tendency now to spread back out underneath the pressure of the axle nuts. In the process of putting the axle nuts back on, i discovered another disturbing thing: apparently when the wheel fell out and turned sideways, it spread the forks apart too. I had had to spread the forks to fit before, but now they were fully 2 inches too wide. Since I never trusted the thin internal washers the hub motor came with, I took this as an opportunity to move the thicker washers to the inside too. This may also have been a mistake.
When the forks spread, I don't think they spread evenly. I think one side stayed mostly straight while the other side bent. I had had to bend the forks apart anyway to fit the hub motor, because they were too narrow for hub motors. I think this weakened one side. (When I asked a bicycle shop if I could buy wider forks, they told me there's no such thing as "wide forks" and all forks are 100mm at the axle. Maybe someone on the forums knows a source for nonstandard wide forks?)
Although I couldn't tell it was lopsided by eye, I could feel that an unequal bending of the fork arms had changed both the camber and the caster angle of the wheel. It causes the bicycle to want to turn left, and on my next test ride I had to constantly push on the left handlebar to keep the wheel going straight. It's tiring and annoying.
On this second test ride I gave it more throttle. With two motors, it has some incredible acceleration. But it's hard to control. I had accidently bought a 9x7 for the front and a 6x10 for the back because I didn't know they were different motors. (This was right at the time e-bikekit was transitioning from one to the other, and I bought one on sale.) Once I realized I'd made a mistake I thought perhaps I could accelerate up to the max speed of the 6x10, and then the 9x7 would pull me slightly faster, but what actually happens is the 9x7 "wants to go faster" than the 6x10 for every throttle position. So the 9x7 ends up doing most of the work and it and its controller get hot while the 6x10 and its controller stay cool. I don't think that's the most power efficient way to divide the work. It would work better if I had two throttles, but I'll have to get some straight handlebars to do that - I had no place to put a throttle on my drop bar so I have a single throttle attached vertically on the handlebar stem.
During the test ride, once I figured out the 9x7 was doing all the work, I decided to disconnect power to the 9x7 and see what things were like with just the 6x10 in the rear. This is when the next scary thing happened: after I got going, when I let off the throttle lever, it didn't slow down. I used the brakes to slow down and it spooled back up to a faster speed on it's own! Then I reached back and flicked off the toggle switch I put on the battery, and the switch malfunctioned and didn't shut off the power! (I have the same piece of crap O'Reilly 50A switch that this forum poster had a problem with arcing with: http://endless-sphere.com/forums/viewtopic.php?f=2&t=18756&p=276095&hilit=50A#p276095. The feel of this switch is very sloppy; completely substandard construction.)
I ended up having to lay the bike on its side to get the spinning wheel off the ground. Then I pulled the Anderson powerpoles apart to physically pull disconnect the battery, which finally got the motor to stop.
Here's what I think happened: The unpowered controller was generating voltage on the throttle signal wire, which was feeding into the powered controller and making it think I was pressing the throttle down. But, where does an unpowered controller get electricity to generate a throttle signal? Why, from the motor of course! Every motor acts as a generator too. It gets even better: a motor controller has a bridge of six protection diodes that happen to also be the same wiring pattern as a three phase rectifier! When the motor turns, they rectify the current and dump it directly into the DC bus capacitors. This generator effect would be why it only happened after I got moving, and why it had a feedback effect (the faster the bike went, the more voltage was generated by the spinning motor).
The other part of the puzzle was that I had connected the second controller to the throttle using only the throttle signal wire and NOT the throttle ground wire. The recommendation I've seen on these forums saying to connect the throttle signal and the throttle ground wire is not quite right (I speak from having a degree in industrial electronics): you actually want only one ground path per device, going to a star ground at the battery. Connecting the throttle ground across two controllers is potentially bad - it creates a ground loop. Motor current could go from one controller, across the ground loop created by the throttle ground connection, and exit through the other controller's ground, putting high current across a signal ground that isn't designed for that. So UNDER NORMAL CIRCUMSTANCES, you actually do NOT want the throttle grounds connected; you want just the throttle wire going across and you should rely ONLY on the thick power ground wires to complete the circuit ground. What I'd done wrong here, which made things not a normal circumstance, was I disconnected the power ground wire of the controller. This meant that rather than a single ground point, I had no ground point at all for the controller. (Just because I have a degree in industrial electronics, does not mean I never do stupid things.
) With a floating ground, and electricity generated in the controller by the wheel, it managed to raise the voltage on the throttle wire until the other controller interpreted it as full throttle. In this particular case, if i had wired it with that ground loop between the throttle cables then it would not have been a floating ground and would probably not have done this, BUT I still maintain that you shouldn't do that; my real mistake was disconnecting the power ground to the controller. I should have either disconnected only the red power wire while leaving the ground connected, or I should have disconnected everything including the motor too.
After I figured out what was wrong, I wanted to pedal home manually, but I then discovered that in laying down the bike I'd thrown the chain off the rear cassette and gotten it caught between the smallest gear and the frame. I couldn't free the chain without tools, so I had to reconnect everything and ride it home entirely on electrical power. I had to trust that I'd correctly reasoned about what was going wrong electrically, and that it wouldn't take off again on its own again once everything was reconnected.
So that was a long story, but what I want to ask you all is, how can I make this beast a bit safer to ride? Can I order extra strong, extra wide forks over the internet? Would you move the battery from a heavy rear rack to a ponderous front position? (Will that make it hard to steer?) I'm also thinking about switching to a recumbent. I've never ridden a recumbent, but the longer wheelbase and shorter height of a recumbent looks like it would make it easier to manage the extra weight of these two motors and battery. I might be able to mount the battery behind me closer to the ground. And with a three wheeled bike, I could add yet another motor!
But now that I've built it, the bicycle frankly scares me a bit: it's hard to control the acceleration, and the weight is just over what is reasonable for a bike. The battery was too wide to put in the triangle and I didn't think of asking Ping to split it into two halves for panniers, so it's altogether on a back rack. I welded a custom rack for it that puts it as low to the wheel and as near to the seat as possible, but it's still a high, rear weight.
While I waited for the front torque arm to come through the mail, I got impatient and tried to ride it without a front torque arm (it has torque arms on the rear). I figured as long as I didn't hot dog on the throttle too much, I wouldn't spin the dropouts. BIIIG mistake. It was actually ok for a few turns around the block. Then I stopped to check the mailbox, and as I was getting going again, I lost the front wheel. It all just sort of fell apart underneath me, like in a western movie when the horse is shot and stumbles under the rider. Luckily I wasn't going fast so I didn't go flying or anything, the bike just fell out from under me.
This got me thinking: Because dropouts open straight down, when the axle spun out and spread the dropouts open, the axle rolled the dropouts right over itself like a log roll. If bicycle dropouts were originally designed for powered hubs, they could be made safer by having the dropouts wrap around the axles like an upside down question mark. The direction of curl could be set up to be opposite the torque of the motor, so that when they spun they'd tend to bury themselves deeper instead of escaping. If I had a machine tool capable of routing a curved question mark shaped path in a steel bar, I could make these.
The torque arm actually came in the mail the next day. I hammered the dropouts back into shape, but unless I'm imagining things I think they have a tendency now to spread back out underneath the pressure of the axle nuts. In the process of putting the axle nuts back on, i discovered another disturbing thing: apparently when the wheel fell out and turned sideways, it spread the forks apart too. I had had to spread the forks to fit before, but now they were fully 2 inches too wide. Since I never trusted the thin internal washers the hub motor came with, I took this as an opportunity to move the thicker washers to the inside too. This may also have been a mistake.
When the forks spread, I don't think they spread evenly. I think one side stayed mostly straight while the other side bent. I had had to bend the forks apart anyway to fit the hub motor, because they were too narrow for hub motors. I think this weakened one side. (When I asked a bicycle shop if I could buy wider forks, they told me there's no such thing as "wide forks" and all forks are 100mm at the axle. Maybe someone on the forums knows a source for nonstandard wide forks?)
Although I couldn't tell it was lopsided by eye, I could feel that an unequal bending of the fork arms had changed both the camber and the caster angle of the wheel. It causes the bicycle to want to turn left, and on my next test ride I had to constantly push on the left handlebar to keep the wheel going straight. It's tiring and annoying.
On this second test ride I gave it more throttle. With two motors, it has some incredible acceleration. But it's hard to control. I had accidently bought a 9x7 for the front and a 6x10 for the back because I didn't know they were different motors. (This was right at the time e-bikekit was transitioning from one to the other, and I bought one on sale.) Once I realized I'd made a mistake I thought perhaps I could accelerate up to the max speed of the 6x10, and then the 9x7 would pull me slightly faster, but what actually happens is the 9x7 "wants to go faster" than the 6x10 for every throttle position. So the 9x7 ends up doing most of the work and it and its controller get hot while the 6x10 and its controller stay cool. I don't think that's the most power efficient way to divide the work. It would work better if I had two throttles, but I'll have to get some straight handlebars to do that - I had no place to put a throttle on my drop bar so I have a single throttle attached vertically on the handlebar stem.
During the test ride, once I figured out the 9x7 was doing all the work, I decided to disconnect power to the 9x7 and see what things were like with just the 6x10 in the rear. This is when the next scary thing happened: after I got going, when I let off the throttle lever, it didn't slow down. I used the brakes to slow down and it spooled back up to a faster speed on it's own! Then I reached back and flicked off the toggle switch I put on the battery, and the switch malfunctioned and didn't shut off the power! (I have the same piece of crap O'Reilly 50A switch that this forum poster had a problem with arcing with: http://endless-sphere.com/forums/viewtopic.php?f=2&t=18756&p=276095&hilit=50A#p276095. The feel of this switch is very sloppy; completely substandard construction.)
I ended up having to lay the bike on its side to get the spinning wheel off the ground. Then I pulled the Anderson powerpoles apart to physically pull disconnect the battery, which finally got the motor to stop.
Here's what I think happened: The unpowered controller was generating voltage on the throttle signal wire, which was feeding into the powered controller and making it think I was pressing the throttle down. But, where does an unpowered controller get electricity to generate a throttle signal? Why, from the motor of course! Every motor acts as a generator too. It gets even better: a motor controller has a bridge of six protection diodes that happen to also be the same wiring pattern as a three phase rectifier! When the motor turns, they rectify the current and dump it directly into the DC bus capacitors. This generator effect would be why it only happened after I got moving, and why it had a feedback effect (the faster the bike went, the more voltage was generated by the spinning motor).
The other part of the puzzle was that I had connected the second controller to the throttle using only the throttle signal wire and NOT the throttle ground wire. The recommendation I've seen on these forums saying to connect the throttle signal and the throttle ground wire is not quite right (I speak from having a degree in industrial electronics): you actually want only one ground path per device, going to a star ground at the battery. Connecting the throttle ground across two controllers is potentially bad - it creates a ground loop. Motor current could go from one controller, across the ground loop created by the throttle ground connection, and exit through the other controller's ground, putting high current across a signal ground that isn't designed for that. So UNDER NORMAL CIRCUMSTANCES, you actually do NOT want the throttle grounds connected; you want just the throttle wire going across and you should rely ONLY on the thick power ground wires to complete the circuit ground. What I'd done wrong here, which made things not a normal circumstance, was I disconnected the power ground wire of the controller. This meant that rather than a single ground point, I had no ground point at all for the controller. (Just because I have a degree in industrial electronics, does not mean I never do stupid things.
) With a floating ground, and electricity generated in the controller by the wheel, it managed to raise the voltage on the throttle wire until the other controller interpreted it as full throttle. In this particular case, if i had wired it with that ground loop between the throttle cables then it would not have been a floating ground and would probably not have done this, BUT I still maintain that you shouldn't do that; my real mistake was disconnecting the power ground to the controller. I should have either disconnected only the red power wire while leaving the ground connected, or I should have disconnected everything including the motor too.After I figured out what was wrong, I wanted to pedal home manually, but I then discovered that in laying down the bike I'd thrown the chain off the rear cassette and gotten it caught between the smallest gear and the frame. I couldn't free the chain without tools, so I had to reconnect everything and ride it home entirely on electrical power. I had to trust that I'd correctly reasoned about what was going wrong electrically, and that it wouldn't take off again on its own again once everything was reconnected.
So that was a long story, but what I want to ask you all is, how can I make this beast a bit safer to ride? Can I order extra strong, extra wide forks over the internet? Would you move the battery from a heavy rear rack to a ponderous front position? (Will that make it hard to steer?) I'm also thinking about switching to a recumbent. I've never ridden a recumbent, but the longer wheelbase and shorter height of a recumbent looks like it would make it easier to manage the extra weight of these two motors and battery. I might be able to mount the battery behind me closer to the ground. And with a three wheeled bike, I could add yet another motor!
