GGoodrum
1 MW
I'm amassing enough pieces now, that I thought it was time to start a build thread. The bike I'm starting with, shown below, is a Dahon Curve D3 16" folding bike. This was actually the first folding bike I converted, originally with a front-mounted Crystalyte 408. I eventually decided that I wanted to go to an x5 configuration, but since they wouldn't fit on a 16" wheel, I ended up getting a couple of 20" models. The x5s are beastly on a folding bike, so I converted one of them, a 20" Dahon Mariner, to use a Cyclone 1000 setup. it is the one in my Avatar.
I'm actually quite happy with the performance of the Cyclone setup, which runs off a 16s6p 48V/13.8Ah a123-based pack. The Cyclone motor, which has a kV around 70, drives the front crank, which allows use of a Nexus 3-speed hub in the rear. The only fault I find in this configuration is that they use a ridiculously small #40 (1/2") 6T motor sprocket, so it is noisy as hell. Apparently they also use a cheap freewheel in the crank, although it still appears to working okay for me. Eventually, this can be replaced with a higher quality unit, like the White Industries ENO model, but I'm not going to worry about it until I need to. Anyway, with a 55A controller limit, this setup peaks out at about 2600W, but because of the gearing benefit of going through the 3-speed hub, this combo actually outperforms the 24s4p 72V 5303-based setup with a 4110-modded controller set to about 65A.
This surprising performance "boost" by taking advantage of the bike's gearing has started me on this quest now, to convert all my bikes to similar weight-saving configurations. I'm also now able to go back and standardise my setups to all use 16s/48V packs. Going to 72V was a pain because a lot of my "accessories", like a WattsUp, or an EagleTreeSystems eLogger wouldn't work with pack voltages that high. It has also allowed me to come full-circle in regards to power options. Like Matt, and others here, I come from a longtime background in electric RC models, mainly electric helicopters in my case. My TPpacks website was originally started five years ago as a way to help Charlie Wang sell his first generation Thunder Power LiPo packs. Back several years ago there wern't any larger scale electric helicopters, so I ended up doing a conversion kit for one of the more popular glow-powered models. Now, I'm happy to say, electrics have pretty much taken over. I also got into a123 cells with electric helis as well. First we got into using the eMoli cells in the Milwaukee V28 tool packs, but they had heavy steel cans so these setups took a pretty good performance hit, when compared to LiPos. When the lighter a123s first became available, about a year later, we could get close to the same performance, without too much of a weight penalty. In particular, they were great with the airborne photography setups a lot of us were doing because we could get decent durations and we cold recharge them in about 20-25 minutes. With two sets of packs, I could go pretty much continuous.
Impressed with the power we were pulling out of these RC motors (4-5kW...), I then decided to try and do an electric bike conversion using one. At the time, I had two ebikes. The first was a Giant eSuede, which used a 36V/9Ah NiMN battery. Power was pretty anemic. I eventually replaced the 30 NiMn D-cells with a 10s3p eMoli pack, which works a bit better. The second ebike I had was a mountain bike that had a Bionix 36V kit installed. For that, I kept the 10s6p Konion/Sony pack, and added a 10s2p eMoli "booster" pack, for extra range. Again, I really wasn't happy with the performance, which then caused my epiphany about using one of the larger RC motors, and make use of the bike's gearing.
My first attempt at this was to use a Hacker A60-18L, with 6.3:1 initial gear reduction, using a couple of the glass-impregnated hard plastic gears used on one of my larger helicopter setups. The output shaft of this reduction had a 7T #40 sprocket that then drove the larger of the chainrings on a Townie 21 cruiser-style bike. Here's what the motor setup looked like:
The controller is a Kontronik PowerJazz, which is good for 63V and 200A. I used an Astroflight servo tester which I modified to remove the pot and was instead wired directly to a 0-5k Magura resistive-type throttle. The ESC and servo tester was powered by a 5V UBEC that was connected to half the pack voltage. I "hid" the servo tester and UBEC inside the "box", which was mounted to the frame downtube.
I never really got a chance to fully test this setup, because there was just too much power going through for the gears to hold up. I stripped three sets of gears before giving up. The other problem I had, which I hadn't thought through, was that the pedals turned with the motor drive. That sent me back to the Internet, in search of front-mounted freewheels. That's when I found Endless-Sphere. After discovering the more powerful setups people were doing with hubmotors, I decided to switch horses, and the RC-based setup went to the back burner. Now, being much wiser ( ), I'm ready to come full-circle, and get back to doing this right.
During the brief amount of testing with the Hacker that I did manage to get in, it was clear that it can handle the power of the bike environment without getting too hot. After I crunched the first set of gears, I took it easy starting off, and was able to go through a couple of shifts and ride around the neighborhood a bit. The motor barely got warm. I never worried about the PowerJazz, as it wasn't really being taxed, and it has a built-in fan mounted on top. I also didn't have the sort of issues that Matt has seen with his HV110 ESCs that he kept blowing before adding the caps, but I may yet have an issue. We'll see.
Here's the specs on the motor and the controller:
Hacker A60-18L
Kontronik Power Jazz
The German-built Hackers have always been great performers, but not inexpensive. At the time, there weren't too many choices out there, so we were stuck with the high prices. Eventually, the Chinese started doing cheap knockoffs, that were a fraction of the cost, but they didn't perform anywhere near as well, mainly due to the use of cheap magnets, and unpure steel in the stator, which causes heat-generating, power-robbing, eddy currents. The quality eventually improved, however, and it forced Hacker, and others, to also move production to China. In Hacker's case, however, they did a better job on quality control and used the best magents, high-quality steel and good bearings. For awhile, I was importing motors from Taiwan which were better quality than the mainland knockoffs, and almost as good as the Hackers, but even these suffered from quality issues, and I lost a ton of money on replacements.
I also used to like the Korean-built Hyperion motors, as the quality was better than the Chinese variants. I see they now have one with very similar specs as the A60-18L, the Z5045-18. I'm still technically a dealer for these, so I think I will get one to campare with the Hacker. The specs nearly are identical, and it is about $100 less expensive.
The highest-quality motors I've ever used come from Steve Neu, down in San Diego (Neumotors.com). He has cases made in China, but everything else comes from here, well sorta. He uses very thin, coated stator laminations, from the highest quality steel, but they are machined in India. The magnets come from somewhere in Pennsylvania and he has the stators hand-wound, just across the border in Mexico. Final assembly and test is done at his facility in San Diego. Anyway, these motors are about as efficient as you can get, and yet still quite reasonably priced. What I really like is that they are not outrunners, with spinning cans, which I think is not the best answer for use on a bike, but they do have large diameters so they have high torque at lower rpms, like an outrunner. We actually coined a nickname for the first of these,"ORK", which stood for OutRunner Killer.
At the time I was doing the Hacker setup, I talked to Steve about using one of his 2200 series"BAMs" (Big-Ass Motor... ), but he didn't really have one with a low enough kV. What we ended up doing was using the 2215-3Y (kV: 480...) with a massive P62 6.75:1 planetary gearbox. I still have this motor/GB combo, shown below, and eventually plan to use it with Matt's eDrive to repower my Townie. It is a little hard to get a size reference, but the whole assembly is about the size of a Coke can. The GB output shaft is 12mm. This gearbox is uber-expensive, however, so not really a great solution. Steve now also has a longer core version, the 2230, and it is now easier for him to do lower kV versions, without losing efficiencies, so one of my conversions will probably use one of these, coupled with a single-stage eDrive.
Getting back to the project at hand, I've got lots of bits and pieces ordered, with some starting to show up. Originally, my plan was to use the Dimension "Big Cheese" 110 BCD adapter and drill it out, as shown in one of the threads here, in order to install a 22T White Industries ENO high quality freewheel with the Echo Trials cranks, which screw into the freewheel. I've ordered the 22T ENO and the Echo cranks, but the 22T vrsion was out-of-stock at Web Cyclery until late last week, so these items just shipped. At somebody's suggestion in the Freeweel Crank Resources thread, I looked into getting an IPS crankset (Independent Pedaling System...), which are used in some tandem bikes so that the "stoker" doesn't have to always pedal. Tandem purists will tell you that this is cheating ( ), so these IPS drives aren't so easy to find. I ordered one here: IPS crankset. The freewheel is implemented simply by including a standard 16T BMX freewheel inside a "sandwich" that makes up the spider arms, which have the same standard 110mm BCD as the Big Cheese adapter, but no drilling is required. I have no idea about the quality of the freewheel used, but since it is a standard unit, I can replace it later with a 16T ENO. It did come packed with grease, at least, so we'll see how long it lasts. Here is what it looks like with the two Sugino MTB chainrings I'm planning to use:
In order to get a good balance between high low-end torque, and adequete top speed, using the SRAM 3-speed hub, I figure I need a reduction to the crank of around 26.5:1. I want to keep the pedaling part of the system close to what is now, mainly to have a "get home" capability, should somethng fail. Right now there is a 44T chainring on the front, and a 16T cog on the hub. 2nd gear on the hub is 1:1, with 1st and 3rd being 0.75:1 and 1.33:1, respectively. It is roughly equivalent to having a 3-speed cassette with cogs of 22T, 16T and 12T. My initial plan is to modify my existing Hacker "box" setup to replace the gear drive with the same 5mm HTD belt drive that Matt is going to use with the eDrives. It turns out this box setup will be easier to mount on the Curve D3, in the area behind the seatpost, than the current edrive mount, so this will get me something play with until I can figure out how best to get Matt's mount to work. Initially, this conversion will be a "test mule", of sorts, so I can try some different motor and controller combinations. Eventually, once I figure out what works, and what doesn't, this will get a prettied up, blinged out eDrive setup.
The motor has a kV of 149, which is perfect, I think, for a 48V setup and a single-stage beltdrive reduction. With a 12T pulley on the motor, and a 72T pulley on the output shaft, you can get a reduction as high as 6:1. The current box setup has a 7T sprocket to drive the crank, but in its place I'm going to use a freewheel adapter I also ordered from BikePartsUSA.com. This adapter basically reduces the 1.375" threaded hole in a MX freewheel, or a track cog, down to 15mm. I'm then using a bore reducer to get it down to the 10mm I need for my current output shaft. Matt is having special freewheel adapters made that will go right on his 1/2" output shaft. Anyway, I plan to use this adapter with a 12T track cog. It will drive a 53T chaining on the IPS crankset, so the secondary reduction is 4.42:1.
To get to my desired total reduction to the crank of 26.5:1, I need the 1st stage to be 6:1, so the 12T/72T belt drive option seems perfect. On my existing box setup, however, it will be hard to add an idler pulley, and I'm worried that not enough teeth on the 12T pulley will be engaged, so initially I'm going to use a 17T motor pulley. This also keeps me from having to drill out the 12T pulley to 8mm, which is the size of the Hacker shaft, as anything below 17T has a 6mm bore. The 17T has an 8mm bore already. Anyway, this decision messes up my gearing, so it was back to the spreadsheet. What I found is that I could get close to the same gearing by changing the front "pedal" chainring to 40T (from 44T...), and go to a slightly larger 18T cog (from the existing 16T...) on the rear hub, without changing the pedal-only ratios too much. I went too far on my Cyclone setup, changing out the supplied 36T chainring for a 24T "granny" gear, and then going from 18T to 23T on the rear cog. It actually has too much low-end torque now, and it has made the pedals pretty useless. Even in 3rd it is like using the lowest granny gear on a mountain bike, with a big 1st gear sprocket in back. That's why this time I'm more sensitive to making changes to the existing gearing for pedaling. Still, I think this is a worthwhile compromise and a good starting point.
Today I'm going to work on converting the box setup to the HTD belt drive. I will post some pics when this is finished. Next, after that, I will work on getting the box assembly mounted to the frame. Finally, I will install the new crankset, change out the rear cog, and get the motor/box-to-crank chain installed and adjusted. Finishing touches will be to refit the rest of the electronics, install the throttle, add an ET eLogger to the setup, with its remote display and then finish up the 16s4p a123 pack/BMS combo I've been working on for this bike.
-- Gary
I'm actually quite happy with the performance of the Cyclone setup, which runs off a 16s6p 48V/13.8Ah a123-based pack. The Cyclone motor, which has a kV around 70, drives the front crank, which allows use of a Nexus 3-speed hub in the rear. The only fault I find in this configuration is that they use a ridiculously small #40 (1/2") 6T motor sprocket, so it is noisy as hell. Apparently they also use a cheap freewheel in the crank, although it still appears to working okay for me. Eventually, this can be replaced with a higher quality unit, like the White Industries ENO model, but I'm not going to worry about it until I need to. Anyway, with a 55A controller limit, this setup peaks out at about 2600W, but because of the gearing benefit of going through the 3-speed hub, this combo actually outperforms the 24s4p 72V 5303-based setup with a 4110-modded controller set to about 65A.
This surprising performance "boost" by taking advantage of the bike's gearing has started me on this quest now, to convert all my bikes to similar weight-saving configurations. I'm also now able to go back and standardise my setups to all use 16s/48V packs. Going to 72V was a pain because a lot of my "accessories", like a WattsUp, or an EagleTreeSystems eLogger wouldn't work with pack voltages that high. It has also allowed me to come full-circle in regards to power options. Like Matt, and others here, I come from a longtime background in electric RC models, mainly electric helicopters in my case. My TPpacks website was originally started five years ago as a way to help Charlie Wang sell his first generation Thunder Power LiPo packs. Back several years ago there wern't any larger scale electric helicopters, so I ended up doing a conversion kit for one of the more popular glow-powered models. Now, I'm happy to say, electrics have pretty much taken over. I also got into a123 cells with electric helis as well. First we got into using the eMoli cells in the Milwaukee V28 tool packs, but they had heavy steel cans so these setups took a pretty good performance hit, when compared to LiPos. When the lighter a123s first became available, about a year later, we could get close to the same performance, without too much of a weight penalty. In particular, they were great with the airborne photography setups a lot of us were doing because we could get decent durations and we cold recharge them in about 20-25 minutes. With two sets of packs, I could go pretty much continuous.
Impressed with the power we were pulling out of these RC motors (4-5kW...), I then decided to try and do an electric bike conversion using one. At the time, I had two ebikes. The first was a Giant eSuede, which used a 36V/9Ah NiMN battery. Power was pretty anemic. I eventually replaced the 30 NiMn D-cells with a 10s3p eMoli pack, which works a bit better. The second ebike I had was a mountain bike that had a Bionix 36V kit installed. For that, I kept the 10s6p Konion/Sony pack, and added a 10s2p eMoli "booster" pack, for extra range. Again, I really wasn't happy with the performance, which then caused my epiphany about using one of the larger RC motors, and make use of the bike's gearing.
My first attempt at this was to use a Hacker A60-18L, with 6.3:1 initial gear reduction, using a couple of the glass-impregnated hard plastic gears used on one of my larger helicopter setups. The output shaft of this reduction had a 7T #40 sprocket that then drove the larger of the chainrings on a Townie 21 cruiser-style bike. Here's what the motor setup looked like:
The controller is a Kontronik PowerJazz, which is good for 63V and 200A. I used an Astroflight servo tester which I modified to remove the pot and was instead wired directly to a 0-5k Magura resistive-type throttle. The ESC and servo tester was powered by a 5V UBEC that was connected to half the pack voltage. I "hid" the servo tester and UBEC inside the "box", which was mounted to the frame downtube.
I never really got a chance to fully test this setup, because there was just too much power going through for the gears to hold up. I stripped three sets of gears before giving up. The other problem I had, which I hadn't thought through, was that the pedals turned with the motor drive. That sent me back to the Internet, in search of front-mounted freewheels. That's when I found Endless-Sphere. After discovering the more powerful setups people were doing with hubmotors, I decided to switch horses, and the RC-based setup went to the back burner. Now, being much wiser ( ), I'm ready to come full-circle, and get back to doing this right.
During the brief amount of testing with the Hacker that I did manage to get in, it was clear that it can handle the power of the bike environment without getting too hot. After I crunched the first set of gears, I took it easy starting off, and was able to go through a couple of shifts and ride around the neighborhood a bit. The motor barely got warm. I never worried about the PowerJazz, as it wasn't really being taxed, and it has a built-in fan mounted on top. I also didn't have the sort of issues that Matt has seen with his HV110 ESCs that he kept blowing before adding the caps, but I may yet have an issue. We'll see.
Here's the specs on the motor and the controller:
Hacker A60-18L
Kontronik Power Jazz
The German-built Hackers have always been great performers, but not inexpensive. At the time, there weren't too many choices out there, so we were stuck with the high prices. Eventually, the Chinese started doing cheap knockoffs, that were a fraction of the cost, but they didn't perform anywhere near as well, mainly due to the use of cheap magnets, and unpure steel in the stator, which causes heat-generating, power-robbing, eddy currents. The quality eventually improved, however, and it forced Hacker, and others, to also move production to China. In Hacker's case, however, they did a better job on quality control and used the best magents, high-quality steel and good bearings. For awhile, I was importing motors from Taiwan which were better quality than the mainland knockoffs, and almost as good as the Hackers, but even these suffered from quality issues, and I lost a ton of money on replacements.
I also used to like the Korean-built Hyperion motors, as the quality was better than the Chinese variants. I see they now have one with very similar specs as the A60-18L, the Z5045-18. I'm still technically a dealer for these, so I think I will get one to campare with the Hacker. The specs nearly are identical, and it is about $100 less expensive.
The highest-quality motors I've ever used come from Steve Neu, down in San Diego (Neumotors.com). He has cases made in China, but everything else comes from here, well sorta. He uses very thin, coated stator laminations, from the highest quality steel, but they are machined in India. The magnets come from somewhere in Pennsylvania and he has the stators hand-wound, just across the border in Mexico. Final assembly and test is done at his facility in San Diego. Anyway, these motors are about as efficient as you can get, and yet still quite reasonably priced. What I really like is that they are not outrunners, with spinning cans, which I think is not the best answer for use on a bike, but they do have large diameters so they have high torque at lower rpms, like an outrunner. We actually coined a nickname for the first of these,"ORK", which stood for OutRunner Killer.
At the time I was doing the Hacker setup, I talked to Steve about using one of his 2200 series"BAMs" (Big-Ass Motor... ), but he didn't really have one with a low enough kV. What we ended up doing was using the 2215-3Y (kV: 480...) with a massive P62 6.75:1 planetary gearbox. I still have this motor/GB combo, shown below, and eventually plan to use it with Matt's eDrive to repower my Townie. It is a little hard to get a size reference, but the whole assembly is about the size of a Coke can. The GB output shaft is 12mm. This gearbox is uber-expensive, however, so not really a great solution. Steve now also has a longer core version, the 2230, and it is now easier for him to do lower kV versions, without losing efficiencies, so one of my conversions will probably use one of these, coupled with a single-stage eDrive.
Getting back to the project at hand, I've got lots of bits and pieces ordered, with some starting to show up. Originally, my plan was to use the Dimension "Big Cheese" 110 BCD adapter and drill it out, as shown in one of the threads here, in order to install a 22T White Industries ENO high quality freewheel with the Echo Trials cranks, which screw into the freewheel. I've ordered the 22T ENO and the Echo cranks, but the 22T vrsion was out-of-stock at Web Cyclery until late last week, so these items just shipped. At somebody's suggestion in the Freeweel Crank Resources thread, I looked into getting an IPS crankset (Independent Pedaling System...), which are used in some tandem bikes so that the "stoker" doesn't have to always pedal. Tandem purists will tell you that this is cheating ( ), so these IPS drives aren't so easy to find. I ordered one here: IPS crankset. The freewheel is implemented simply by including a standard 16T BMX freewheel inside a "sandwich" that makes up the spider arms, which have the same standard 110mm BCD as the Big Cheese adapter, but no drilling is required. I have no idea about the quality of the freewheel used, but since it is a standard unit, I can replace it later with a 16T ENO. It did come packed with grease, at least, so we'll see how long it lasts. Here is what it looks like with the two Sugino MTB chainrings I'm planning to use:
In order to get a good balance between high low-end torque, and adequete top speed, using the SRAM 3-speed hub, I figure I need a reduction to the crank of around 26.5:1. I want to keep the pedaling part of the system close to what is now, mainly to have a "get home" capability, should somethng fail. Right now there is a 44T chainring on the front, and a 16T cog on the hub. 2nd gear on the hub is 1:1, with 1st and 3rd being 0.75:1 and 1.33:1, respectively. It is roughly equivalent to having a 3-speed cassette with cogs of 22T, 16T and 12T. My initial plan is to modify my existing Hacker "box" setup to replace the gear drive with the same 5mm HTD belt drive that Matt is going to use with the eDrives. It turns out this box setup will be easier to mount on the Curve D3, in the area behind the seatpost, than the current edrive mount, so this will get me something play with until I can figure out how best to get Matt's mount to work. Initially, this conversion will be a "test mule", of sorts, so I can try some different motor and controller combinations. Eventually, once I figure out what works, and what doesn't, this will get a prettied up, blinged out eDrive setup.
The motor has a kV of 149, which is perfect, I think, for a 48V setup and a single-stage beltdrive reduction. With a 12T pulley on the motor, and a 72T pulley on the output shaft, you can get a reduction as high as 6:1. The current box setup has a 7T sprocket to drive the crank, but in its place I'm going to use a freewheel adapter I also ordered from BikePartsUSA.com. This adapter basically reduces the 1.375" threaded hole in a MX freewheel, or a track cog, down to 15mm. I'm then using a bore reducer to get it down to the 10mm I need for my current output shaft. Matt is having special freewheel adapters made that will go right on his 1/2" output shaft. Anyway, I plan to use this adapter with a 12T track cog. It will drive a 53T chaining on the IPS crankset, so the secondary reduction is 4.42:1.
To get to my desired total reduction to the crank of 26.5:1, I need the 1st stage to be 6:1, so the 12T/72T belt drive option seems perfect. On my existing box setup, however, it will be hard to add an idler pulley, and I'm worried that not enough teeth on the 12T pulley will be engaged, so initially I'm going to use a 17T motor pulley. This also keeps me from having to drill out the 12T pulley to 8mm, which is the size of the Hacker shaft, as anything below 17T has a 6mm bore. The 17T has an 8mm bore already. Anyway, this decision messes up my gearing, so it was back to the spreadsheet. What I found is that I could get close to the same gearing by changing the front "pedal" chainring to 40T (from 44T...), and go to a slightly larger 18T cog (from the existing 16T...) on the rear hub, without changing the pedal-only ratios too much. I went too far on my Cyclone setup, changing out the supplied 36T chainring for a 24T "granny" gear, and then going from 18T to 23T on the rear cog. It actually has too much low-end torque now, and it has made the pedals pretty useless. Even in 3rd it is like using the lowest granny gear on a mountain bike, with a big 1st gear sprocket in back. That's why this time I'm more sensitive to making changes to the existing gearing for pedaling. Still, I think this is a worthwhile compromise and a good starting point.
Today I'm going to work on converting the box setup to the HTD belt drive. I will post some pics when this is finished. Next, after that, I will work on getting the box assembly mounted to the frame. Finally, I will install the new crankset, change out the rear cog, and get the motor/box-to-crank chain installed and adjusted. Finishing touches will be to refit the rest of the electronics, install the throttle, add an ET eLogger to the setup, with its remote display and then finish up the 16s4p a123 pack/BMS combo I've been working on for this bike.
-- Gary