Cuda E with RC drivetrain

Cuda E with RC Drivetrain

A project by Warren Beauchamp

I built this electric assist recumbent a couple years ago, Due to an unfortunate accident with gear and chains and a lack of knowledge of why the battery would cut out, the bike gained the nickname "finger-eater". I commuted to work on it for a couple years, and it handled great and was a good height for riding in traffic, but was a bit lethargic.

I then decided to build a new e-bike based on a hub motor and a suspended highracer recumbent, but that bike did not handle nicely, so I have been commuting on the fully suspended mountain bike I bought for parts for the last two years. That bike is still working great too, and is much better on the hills, but it's a wedgie...

I have been thinking about a new e-bike for a while now. I have decided to refit finger-eater with a high powered RC drivetrain. This system will have about the same top speed as both previous bikes (30MPH), but will get there a whole lot faster. To do this I will be using one of Matt Shumaker's RC drive units. The bike shown above is 60 lbs. With the new drivetrain and rear subframe it will be closer to 45 lbs, with 6 times the power.
My past two e-bikes used off the shelf drive systems designed for electric bikes. This system will be a bit different, and will require more custom work and electronics work. Fortunately, Matt has done most of the heavy lifting here.

Here's the first part. It's a 4000 watt (5 HP!) AstroFlight  RC motor attached to one of Matt's single stage stage reduction unit. It's pretty small and light. That's a quarter in front of it for reference. It weighs a bit more than 5 lbs. The other side of that big belt cog is threaded for a BMX freewheel. The freewheel drives a standard road 52 tooth chainring that is mounted to the rear hub through a special adaptor (shown here) that connects to the disk brake hub bolts. This system will have about twice the power of my current e-mountain bike!

Step one is to build a new rear subframe for finger-eater.  The new subframe will be lighter weight and have the proper dropouts for a rear derailleur, allowing me to lose the mid-drive.

RC Motor
Below I will explain the difference between an off the shelf hub motor and an RC system.

Hub motors have three heavy "phase" wires to power the motor, and 6 smaller "hall sensor" wires to sense the motor position.  Hub motor controllers use the information provided by the hall sensors to figure out which set of wires to power up at any given time. This allows them to start from a dead stop easily. Generally hub motors have controllers that are pre-wired to just plug into them, so they are pretty easy to set up.

Builders of RC Motors Motor drive trains have to pay attention to the wiring. RC motors have 6 power wires coming out of them, one set of wires for each two windings. These need to be arranged one of two ways, "Delta" or "Wye". Generally Wye (also called "star") is more efficient. Wye has 1.73 times less KV than Delta (slower speed per volts), but 1.73 times more torque. In general it's a good idea to stick with Wye configuration for electric bikes as e-bikes need slower motor speed, higher torque, and high efficiency. For Wye configuration, the 3 positive wires are tied together, and the 3 negative wires go to the controller. These 3 negative wires are the phase wires, just like in a hub motor.

RC Motors are AC brushless motors and are rated by their KV. KV is the relationship between voltage and motor speed (rpms per volt).  RC Motors do not have any hall sensors. The RC controller (ESC) has to figure out the correct timing synchronization for each throttle position and load condition using EMF (electro motive force) feedback. Some RC controllers are better at this than others. Castle Creations controllers are an example of a good controller. Cheaper controllers like Turnigy work fine for aircraft use, as the low RPM / High load conditions of that environment are not nearly as rigorous as those in an electric bike. For bike use, Turnigy controllers can lose the "sync" when accelerating under high load. This loss of sync sounds like the motor is screeching, and is accompanied by a loss of power. Also the Motor itself is important. To prevent the loss of sync a high quality motor like Plettenburg or AstroFlight should be used.

RC Motors controllers (ESCs) are very small and light compared to traditional e-bike controllers, they are also smarter. This is nice because they are small and lightweight, but it's bad because they do not have the ability to shed heat as easily. Because of this there are a modifications that are recommended to make the ESC more robust, like replacing two of the eight 180uf caps and adding two 1,000uf (Matt says Panasonic capacitors from Digikey part number P12393-ND) and adding heat sinking.

The Throttle for an RC motor is more difficult than a traditional e-bike system because they are designed to plug into a radio receiver. To make it work with a twist throttle, you need to use a gizmo called a servo tester, and then wire it to the resistive throttle. Matt's Instructions here


  • V (volts) - E-bike battery packs are generally 24 or 48 volts. Generally the higher the voltage that faster you go. Sort of like how fast water flows through a hose.
  • A (amps) - This is current. Generally the more current you have, the faster you get to the top speed. Sort of like how big the hose is.
  • W (watts) - This is volts times amps. Usually used to show total power used.
  • Ah (amp-hours) - This is how much current the battery can deliver and for how long. For example, A 10Ah battery can deliver 10 amps for 1 hour at it's rated voltage.
  • C - This is how much peak current the battery can deliver. A 20Ah battery rated at 2C can deliver 40A continuously without frying.
  • S and P (series and parallel) - Individual e-bike batteries are generally small cells with lower voltage and current. To get more voltage and current the batteries are hooked together in series (for more voltage) and parallel (for more current).

My past e-bikes have used battery packs only capable of 2C. These LifePo4 batteries are very safe, but not high current. Because of that, they could go fast, but not get up to top speed quickly. This new system requires high current batteries, and can draw up 80 amps continuously, and will peak higher. A 2C battery capable of that much current would weigh 40 lbs! On this bike I will use LiPo or A123 batteries to deliver high power. These batteries are rated at 30C or higher!  This means that even a tiny 5ah battery would deliver up to 150A. Because RC motor drive systems are more efficient than hub motors, you don't need as many amp-hours (ah) of battery to go the same distance. For instance, my round trip commute uses about 15Ah of my 20Ah x 48V battery on my mountain bike. Because the recumbent is way more aerodynamic and the drivetrain will be more efficient, I'm going to start with a 10AH x 48V pack.

LiPo Batteries

RC LiPo Batteries require respect. It's a lot of power in a little package. In the past treating them badly could result in flames, but the newer LiPos are a bit more forgiving. Luke on endless sphere notes: Never exceed 4.3v per cell and never discharge below 2.7v per cell. If you abide by that guideline they should behave. RC LiPo Batteries are nominally 3.7V per cell, but generally run closer to 4 volts each. When you buy RC battery packs they are rated by "S" and "C" and by the milliamp-hours (5000 milliamps is 5 A). S is the number of batteries in the pack, arranged in series. C is how fast the battery can be discharged. 6S packs are about 24V.

To build a pack that is 10Ah and 48V requires four 5000Mah of these (5ah) 6s 25C Lipo packs. This battery pack will supply 48V at 10 amps for one hour, or up to 250 amps(!) in bursts. These packs are about $70 each, or about $240 for all 4. This pack would only weigh about 7 lbs. Lipo batteries are cheaper than LiFePo4, but they cannot be recharged as many times as LiFePo4. Also because they do not have an on-board battery management system (BMS), charging them is more involved. Also when discharging there is no protection.


LVC - To ensure that the LiPo batteries are not damaged by over discharging, the system needs some type of low voltage control (LVC). There are several alternatives that monitor each cell individually, like  This LVC or a low voltage monitor like this one. Fortunately the Castle Creations controller I will be using have a low voltage control that can be programmed in, which can either shut the motor down completely or just reduce the power.

Charging/Balancing - After looking at a bunch of different LiPo charging options, this charger: Bantam e-station BC6-DC  seems to be good quality and reasonable price. I will need two, at $86 ea. Because the chargers run on 12 volts, they also require a DC power supply like this one ($70), or a MeanWell power supply (ebay). After a commute, the batteries could just be plugged into these chargers and they would be recharged overnight. This harness can be used to connect the charging leads of the sets of batteries that are in parallel.

LiPo Batteries: $240 -
Charging: $242 - Meanwell 12V 29A power supply (ebay) and two Bantam e-station BC6-DC balance chargers.
Pack specs:10Ah, 44.4V, 25C (250A Peak)
Construction difficulty - low

A123 Batteries for comparison
A123 Batteries are LiFeP04 formulation, will last longer than LiPo and are very safe. They also have a high 30C rating, which means each 2.3Ah cell can deliver 70A!.  To buy a 10Ah x 48V A123 pack with BMS and charger is about $1300. To build a 48V x 10ah battery pack with A123 cells would require 5 cells in parallel (11.5Ah) and 15 cells in series (49.5V).  That's 75 cells!

The cheapest I have seen these is $6 (plus shipping) each for bare cells ($450), but most people buy Dewalt 36V electric tool packs, which contain 10 cells each, for about $110. These cells have tabs on them, which (potentially) makes them easier to connect. To charge and balance them you could use 15 of these 2 amp Chargers ($10 each).

Batteries: $550 - $800
Charging: $150
BMS: ~$100
Pack specs: 11.5Ah, 49.5V, 30C (311A Peak)
Construction difficulty - high

Because the A123 solution is over twice the price of LiPo, I will go with LiPo.


Still with me? Here's the rear disk brake hub and the motor drive adaptor with a 52T chainring.

I measured the dropouts on the old e-bike rear subframe at 5". I need 5.5" for the MTB hub. So, I really do need to build a new rear subframe.

I built up the rear wheel and mounted the motor drive. Fortunately I was able to repurpose an existing wheel and spokes by just replacing the hub. The original hub had smaller flanges, but I lucked out and was just able to re-spoke the  2 cross wheel as 3 cross to compensate. Plus now the wheel will be stronger.

This is the current rear dropout. "Opus the Poet" suggests that I just spread the dropouts a 1/2 inch rather  than building a whole new rear subframe. Good idea. If that works I'll cut out the rear facing dropouts and braze in a Breezer style dropout.

I removed the old rear wheel and 450 watt drive unit  from the rear sub-frame. I was able to wedge the new rear wheel in, so I won't have to expand the dropouts too far. The new drive unit also fits in and lines up well.

In this picture the drive drive unit is sitting on a pedal to provide the proper clearance for the drive pulley. I need to make an adaptor bracket to hold the drive unit in place. It will allow the drive unit to bolt to the existing mounting plate in the subframe.

The Breezer dropouts should arrive soon.

I used a 1/2" threaded rod to crank the dropouts apart.  Stretching it out to 6" wide was enough to get the extra 1/2" in width I needed after if sprang back.

The Astro 3210 is rated at 275 RPM per volt. I'm planning on running it at about 48V, so that means a potential top speed of 13,200 RPM, but Matt says it's more like 7,000 RPM under load. It is attached to Matt's step-down drive unit which is 3.5 to 1. This reduces the RPM to 2,000 RPM.

The output from the drive unit is a 16 tooth BMX freewheel, which drives the 52 tooth chainring. The gear inch calculator shows that with an 18" diameter tire and this gearing will give a speed of 32MPH.

I fabbed up the mounting bracket yesterday, but only got it about half brazed before my torch ran out of gas. Fortunately, it's tacked together enough to test mount it.

Here's the right hand side of the drivetrain.

Here's the drivetrain mounted in the new bracket, with the chain in place and the wheel in the right position. This shows me where I need to mount to Breezer dropouts. The bracket allows the drive unit to slide fore and aft about 3/4".  This (theoretically) is enough to adjust it so the chain is tight and no idler is needed. I may need to add a chain guide on the power side of the big chainring to ensure that the chain does not pop off under load.

The drive unit is a snug fit in the frame, but it fits! The chain has a couple clearance issues. On the top side it hits the canti-brake mount. On the bottom it hits the frame bracket. I can solve this by using a slightly smaller chainring, OR I can file chain clearance on the bottom, hack off the canti brakes, and try to fit a disk brake.

A disk brake requires about 3/4" clearance on the back side. Currently I only have 1/2". It looks like there may be room though for a 1/4" spacer and there may still be clearance for the brake rotor. Hmm...


To solve the chain clearance problem, Matt has sourced a 13T freewheel to replace the 16T freewheel shown above. I have been working on the human power side of the drivetrain. I added a 60T chainring up front, and replaced the mid drive cassette with a nice TerraCycle idler. Once I get the new freewheel installed on the drive unit I'll know where to locate the rear wheel, and will be able to braze in the breezer dropouts. It should all come together pretty quickly.
Friday I picked up the new 13T freewheel, controller and throttle assembly from Matt. Matt showed by how to program the Castle Creations 160 amp controller, and I got to see the his controllers data logs from the monster dual motor trike that went 70MPH. Wow! The scary thing is that it only took a few seconds to shoot from 30MPH to 70MPH.

I also now have the batteries and the charger (still waiting on the power supply China). I am amazed how small the Li-Po batteries are.

I attempted to make a 1/4" aluminum spacer to allow me to mount a rear disk brake outboard of the chain ring adapter, but I must not have been awake because I used the wrong hole saw on it. Arg. That was my last scrap of 1/4" aluminum plate, that I saved from building a suspension fork brake bridge 10 years ago. I ended up investing a few hours making 2 spacers from 1/8" aluminum plate instead. After all that I mounted a disk brake rotor with the spacers and tried to slide a disk brake caliper between the chainring adapter and the brake rotor, but it looked like I needed another 1/16" or so. When I put the wheel on the bike the brake rotor would rub the frame. That's not going to work, I guess I'm stuck with the caliper brakes.

The only issue I am concerned about with the LiPo batteries is the remote chance that something may go wrong and one of the batteries may "flame on" during charging. I have been reassured that this only happens when the batteries are damaged or when they are improperly charged, but hey, sh1+ happens!.  I will order this small fire resistant Security box to put the battery pack into while charging. That should be plenty safe.

I received the spokes last week, and got the wheel built up, so now I have matching wheels with quick releases!

Because I'm now using a 13T freewheel, I need to change the chainring to a 48T instead of the 52T I was using. Arg, something else to find. This will also give me a bit more clearance on the caliper brake boss.

still needed - 48T road chain ring, connectors.


Here's the stuff that I need to put into a nice neat package that will fit in my seat back bag. Pictured are the four 5Ah 22V LiPo battery packs, the Castle Creations Phoenix ICE HV 160 RC motor controller (ESC) , fuse, mongo Anderson connectors, and a switch.

This is the Bantam e-station BC6-DC battery charger. It's tiny.  I'm hooking up the batteries into two parallel packs, so I'll be able to charge two batteries at once. Before I do do that though, I need to charge each pack separately so that each cell is the same voltage (balanced). That way when I attach them together the voltage won't try try quickly flow from one pack into another.

Of course I can't use it at all until I can hook it up to 12 volt power supply. I ordered a Meanwell 12V power supply from EBay a couple weeks ago and apparently it is on the slow boat from China.

In the meantime I thought I would use this computer power supply, which said it supplied 14A at 12V, and . It had two boards in it, one of which had two leads supplying the other board. I was hoping that would be 12V as the other board has an 80 pin connector and no clear point to tap into the 12V, but when I hooked up my Fluke it said 395 volts!

I will wait patiently for the Meanwell 12V power supply...

Here's the drivetrain after fitting the 13T freewheel and the 48T chainring. I now have the proper clearance around the brake mounting tabs. Now I can mark the location of the dropouts, cut and file the existing dropouts, and braze in the new droputs

The  aluminum box I ordered from Mouser arrived today. In this picture I just jammed everything in the box to see if the lid will screw down. Everything fits, barely. I'll fill in any gaps with Coroplast.

Matt says that the ESC does not get hot. I'll screw it to the inside of the box so if it does heat up it will have a nice big heat sink to dissipate heat into.

The Meanwell supply arrived today (the silver box under the Bantam e-station). I had to use the manual to figure out how to use it but it was pretty straightforward. I am now balance charging each battery before I start connecting them in parallel.

All cells are now either 4.20 or 4.19 volts. I had to perform a couple balance charge / discharge cycle to get the individual cells to within .1 volts of each other.

It took me a while to figure out how to set up the throttle/brake/gear shifter on the right handlebar. The Magura throttle is huge! I finally ended up using a Shimano 8 speed trigger shifter and a Forte (Performance brand) MTB brake lever. I had to surgically remove the end of the trigger shifter gear down-shift lever so it would clear the Magura throttle.

Another challenge I had was how to mount the power side idler so the chain would clear the rear stays. Originally I had mounted this idler directly to the BB cup shown here. It was nice and sturdy, but the chain rubbed the bottom of chain stay.

I decided that instead of brazing a bracket on, I would make one that bolts on using the BB cup. Because I wanted a real challenge, I decided to whittle the bracket from 1/8" thick steel, using only a swiss army knife.  The bracket is picture to the left of the idler.

Here's the bracket installed on the bike, and held in place by the BB cup.

Here's the power side idler in place. Everything clears now. I'm not done with the human power drivetrain yet, The return chain also needs an idler.

Now that all the batteries are charged, I have started to combine the battery packs  I think I have now figured out how NOT to solder together three large wires...

To properly solder 10 gauge wires together, first press them together and do not twist the strands. Then take a long piece of solid copper wire, like phone wire, and wrap it tightly around the wires. This holds it all together and makes it much easier to solder.

Here's the battery box after adding Anderson 75 amp connectors to the wires. I think if I did this again I would not use these as they are HUGE. I would use the connectors from Hobby King that match the connectors on the batteries.

One nice thing about the Anderson connectors is that they offer panel mount options.  I used that ability to mount them to one end of the box. This allows detachable output for the motor wires (left side). On the right is the fuse. Disconnecting this Anderson connector disables the battery pack, so effectively this is the on/off switch. 

In the middle are the throttle wires. Because this is an RC system, you can't just plug in a twist throttle. You need more parts (of course). In particular you need a "servo tester" and a "battery elimination circuit" (BEC). The BEC converts the 48V from the battery to 5V for the throttle controller.  The throttle controller, well... controls the throttle.

Here's Matt's tutorial on building an RC throttle system. 

Here's the box with the lid screwed on. It all fits!

Here's a schematic of the batteries and RC drive system. Not shown are the battery balance charge wires

Parts shown:

There are two methods to charge LiPo batteries, one is the balance charger. For that I am using the bantam e-station mentioned earlier. The other method is "bulk charging". Bulk charging allows you to just charge the whole pack at once, which can be done without taking it apart.

The method shown here is a basic bulk charging configuration. It uses a MW charge controller to control the current put into the battery by the Meanwell 48V power supply. When the voltage of the battery pack reaches 50 volts, the charge controller stops the charging. This prevents the pack from being over charged. This method requires that the LiPo packs be well balanced, so the packs should be balanced charged every 5 to 10 charges, depending on how well balanced they stay. The MW charge controller comes as a kit. Here's a picture of the assembled MW charge controller

A more complex and better method involves the use of cell level monitoring to prevent high or low voltage conditions.

I ordered a Cycle Analyst,  the e-bikers dashboard. It monitors battery voltage, battery power, amount of power used, etc. Oh, and it's a bike speedometer too. Very cool. The version that works with the RC components has a small box that is mounted between the battery and the controller, so it's pretty important to fit that in the box.

The Cycle Analyst will be very handy when I start testing tires and things. I will be able to change tires and/or aerodynamic accoutrements and see how many watts each configuration takes to travel at some specific speed. The ESC (controller) also performs logging, but I need to get a special serial to USB convertor to download the data.

Here's how the pack is balance charged. Two paralleled 22.4V LiPo packs are attached directly to the balance charger. Each of the two 22.4V packs is charged separately.

Here's a video of the drive unit testing on my basement floor.

I received the Cycle Analyst yesterday, and the shunt fits in with everything else, if I shoehorn it all in.

Because I may have to open this box up every week or two to balance the batteries, I decided to mount the ESC on the outside of the box. It's not quite as clean but still one self contained monolithic block and much easier to fit all the wires in the box.

The connectors from left to right are throttle, cycle analyst, bulk charging/accessories, motor phase wires (black) and power disconnect (red).

I found a nice old laptop bag that this battery pack fits in nicely. One of the bags pockets slides right onto the shelf. This means I'll just be able to use the shoulder strap to hold the battery on the bike, making it easy to take with me for charging.

I had a bit of a scare after wiring everything up with the CA and testing it in the basement. As soon as I turned it on the CA said I was using 8885 watts with no throttle. At full thottle no change. I got out the manual and found how to zero it. The CA then said 0 watts all the time. Hmm. I send Jason at GrinTech a note to see if he had any bright idea, and went to bed. The next day he replied that that's what happens when you splice the wires and don't hook up the blue one or white one. Ah. On my second temporary splice that I did not solder as I needed to figure out how long to make the wire on my recumbent, there was indeed a bad connection. Works fine, case closed. All is right in the world.

Today I made the bracket for the return idler and confirmed that I don't really need to braze in new rear dropouts. I'm using the bolt-on derailleur hanger from an ancient 10 speed instead. Works fine. I have a bit more tweaking to go on the electric and human powered drive-trains, but the end is in sight!

 I finished the human power drivetrain (at least enough to pedal), and mounted the CA. I wanted to see what the top speed was while the road was dry before the next snowstorm blew through tonight. I pedaled it up to 5 MPH before rolling on the volts. Wow, this thing really pulls hard up to speed. I blasted up to 27MPH in about 2 houses. That was at 25 degree F, with low pressure in the tires, in the dark. I'm pretty confident it will go a couple MPH faster. Fun.

The CA said the battery was at 49.1V, and top current was 80A.

 just weighed the bike. 52 lbs, Whoa. The battery weights in at 13 lbs, so 65 lbs total. I don't think this bike is any lighter than in it's previous incarnation, but it's got a bit more power.

Here's a spiffy video of testing the RC drive unit on the recumbent bike in the ice cold garage:

This bike will need lights. Bright lights. Because I want to power them with the main battery, I will need to use a DC to DC converter to reduce the 50V from the main battery to the 8.4V that the MagicShine lights require. This can be done with the Castle Creations BEC. has good prices for Magicshine head and tail lights with no battery. So, about $150 for a bright solution.

While the Castle Creations products above would have been fine, I ended up getting an adjustable DC to DC convertor from Lyen. It was less that $20 and does not require USB for programming.

That makes this super lighting system under $100.




A while ago I tried to mount a disk brake on the chainring adaptor, but it didn't have enough back side clearance. Apparently  the trick is to use a 203mm disk brake, as that clears the adapter. I finally found one that didn't break the bank. It's an Avid BB7.

I bolted it on with a single 5mm bolt to check clearance. The stock bolts were not long enough to make through the chainring adapter, 1/8" spacer, and the disk brake.

As seen here, the huge brake entirely hides the chainring. The brake caliper is just resting there. I'll need to make a special bracket and braze it onto the rear dropouts.

Here's the top view showing the tight tolerances. The brake rotor definitely clears the frame, but may rub on the extra wide BMX chain. I will probably space the chainring in with some thin chainring shims to buy the extra clearance there.

I tried four separate bike bags to mount this battery. Some were too big, some were too small, and some were odd shaped.

So I spent a whole bunch of hours making this spiffy aluminum cage that bolts to my battery shelf. The battery slides in and out from the side.

Unfortunately, it's really ugly.

Next I decided that if I mounted some feet to the bottom of the case that would keep it from shifting around on the rack, then I could just strap it down with some Velcro.

Except then I would have to drill more holes in the case and the screw heads may abrade the battery packs.

My final thought was to use a chunk of self adhesive Zote foam on the bottom of the box, and cut grooves in it to fit the cross tubes on the rack. That holds it nice and firm when strapped down with velcro, plus it will be easy to remove.

Here it is. I still need to hide the wires a bit. I thinks I will build a black Coroplast cover for the ESC and wires that will let the air flow through.

I soldered up the Lyen DC to DC converter with aAnderson connectors on one end and two connectors for the MagicShine lights. I used the screw adjustment on the little blue potentiometer you can see in this picture to set the output voltage to 8.4 volts because that's the voltage these lights like to see.

I did some testing up and down the street and on the bike trail, and shot some video. The bike needed to have the belt tightened, as it was skipping at low speeds, but above about 10 MPH the motor pulls pretty hard.
I had a nice ride today to test the range of the bike and just enjoy some time cruising. I went about 50 miles but pedaled for over half of it so it wasn't a great test of the range, but it does tell me it will make it for the 15 mile round trip to work and back.

The Cycle Analyst says the peak wattage is around 3500 watts if I crank the throttle wide open at 10MPH. Flatland cruising at 20MPH takes a bit over 200 watts, 30MPH takes around 700 watts.

The bike makes it up the big hill on the way home at around 25MPH. Plenty fast. I left barely enough power in the battery to get up the hill. It made it up but then I had to pedal the rest of the way home. Matt set the Castle Creations ESC LVC setting to shut down at 39.5 volts to protect the battery and it did that very well. When the CA said 40V, the ESC said no more E-power for the motor.

I'm charging the packs now. The cells were not too well balanced yet (ranging between 3.2 and 3.5 volts) when I started the charge but I'm hearing it takes several cycles for them to start staying in balance. This is only charge #3 and the first time I have depleted the batteries to LVC shutoff. After about half an hour they were all happily at 3.77v.
Apparently I should be making sure I stop using the pack when it gets down to 44V. This will ensure the cells stay balanced and they last a long time. Seems goofy to me as that is the "nominal" voltage of the pack, but whatever, I should change the LVC on the controller to do this.

It's been pretty cold this spring but I have gotten some riding in. I have tested this new bike and it's electric motor system several times and it seems to be working well with no thrown chains or other unexpected problems.

The system now uses a 13T BMX freewheel to drive the 43T chainring that's mounted to the disk brake mount on the rear wheel. That gives me the 30 MPH max speed I was looking for. Unfortunately the freewheel locked up on my last test ride, which didn't cause a problem other than the wheel was driving the motor when the motor wasn't engaged. That added less drag than I would have imagined. I took the freewheel apart and it looked like one of the ball bearing somehow got into the area with the pals. This crimped the pal retainer ring and broke it, and also allowed the ball bearing to jam up under the pal. I reassembled it with what was left of the retainer ring and it seems fine for now but I suspect that retainer ring will creep around and the one of the pals will no longer be spring loaded. Matt is getting me new a new freewheel to replace it.

Also I ordered a new e-bike helmet, the Casco E.Motion helmet from Germany. Yes it was too expensive, but it's got spiffy ear flaps that I'm hoping will quiet some of the wind noise that I get buzzing down the road at high speeds, plus some interesting venting and air control features which are a bit hard to see in this photo.

I haven't ridden the new e-recumbent to work yet, but I will as soon as we get a decent day to ride.

I have been riding this bike to work a couple times per week now and it's working fine.

26 wh/mi average with 30 MPH cruising, jackrabbit starts, hills, etc.
Over the past week, rode 38 miles, used 22 Ah, so I can make it to work and back on 8.2 Ah. Maybe I don't need to charge at work but the system is a lot more peppy on a full charge. Bulk charging using the modified Meanwell and the BMSBattery chargers is working well, and the pack seems to staying in balance.

The Casco helmet is working great on the recumbent after modding the ear flaps with some leather to cover the holes. I still hear the wind but it's not painful.

Max wattage seems to be less than 3000 watts. You have to pedal up to 5 MPH before throttling up or else the motor/drivetrain has issues. I'm going to visit Matt to have him make me a torque clutch as that seems to really help at low speeds. It really pulls from about 10MPH to 25, then gradually makes it up to 30.

I finally bought a Castle Link USB connector so that I can change the RC controller (ESC) settings and view the logs. Here are the HV160 settings for use with Astroflight 3210 and 3220 RC motors. Here's about half of my commute (11.66 minutes):

This log shows wattage over about 12 minutes of my commute. It shows the frequent stops and starts of my journey. Jack rabbit starts make the system peak out at about 4000 watts, but only for a couple seconds. 

Here's an  HD 10 minute video of my commute to work and back on the e-recumbent.

Maximize it for best quality.

I had been experiencing some oddness while the AstroFlight 8510 motor was running at low RPM for some time, and the drivetrain was starting to make some strange clunking and other noises, so I decided it was time to take it apart. The reduction unit seemed fine and it's bearings were in good shape, but the shaft of the motor had a lot of play in it and the motor made clunking noises when spinning it by hand. Ewww. Not good.

I removed the motor from the drive unit. I suppose the jingling when I shake it is not good?  The bearing on the drive end was trashed. It was missing many bearings. That little flake on the stator is probably what is left of one of them. The inner bearing cover plate was a crumpled little wad. After wiping down the stator there were a couple small gouges but it looks very good, considering the hard use it has been through.

The inside of the motor looks good too. The bearing on the cap end had a lot of play in it but had not lost any bearings yet. It needs a little cleaning and some new bearings but after that it should be good as new.

What worries me is that the drivetrain only had about 250 miles on it. 250 hard miles. Hopefully the new bearings will last longer than the original ones.

Also Matt will be making a clutch unit for the drivetrain soon, which will cut down on the low RPM stress and allow the belt to be a little less tight.

 I think the reason that bearings failed so fast was because I was running the belt tension too high. 

I installed the new bearings from Grainger to replace the bad ones in the motor and Matt made the clutch modifications to the large pulley. The motor runs as good as new.

This August we have had a lot of nice days and I have been commuting to work many times per week on this bike. Fun!

I upgraded the motor to a AstroFlight 3215. It's just a bit higher power and faster. I also changed from using the BEC & Servo tester to the new beta version RC Cycle Analyst, which has the BEC and servo controller built in. Pictured here is my entire new RC e-bike system.

It took a little while and I thought it broke my ESC, but eventually I figured out how to make the beta version of the ebikes RC CA work with my Castle Creations Phoenix HV ESC. It didn't break it but apparently the RC CA sent signals to the ESC that caused it to flash error codes, so I had to do a full reset of the ESC using the CC management software to clear the errors.

Here are the RC CA settings that worked for me
Here are the CC HV160 ESC settings that worked for me

Here's the new motor mounted, I had to cut a notch in the frame to make it fit. 

I sent the ESC (controller) back to Castle Creations to be upgraded to a Phoenix ICE2 as part of a service advisory. It's taken a long time for them to upgrade it, but it looks like they have finally shipped it. I am looking forward to getting this bike put back together and tested!

Gearing is 3.5 to 1 on the Schumaker drive, and I am using a 14T freewheel and a 39T chainring on the rear wheel, for an additional 2.7 to 1 reduction. Using the RC drive calculator with the above parameters, voltage set to 48, the motor KV set at 203 and the % of speed set to 80% parameters gives 47 MPH which is just how fast this bike goes.:

The controller arrived and I soldered the connectors back on. The bike is now basically ready for a test ride. Hopefully we will get a nice dry and not too cold day for me to take it for a test ride soon!

So, the real world is finally getting warmer and I was able to take it for a ride around the block (actually several laps around the block). The controller works well, though it seems like minimum throttle cuts in a little harder than with the old servo tester / BEC system. Performance at very low speeds seems to be comparable to the old servo tester / BEC system, so I still have to pedal up to 5 MPH or so before throttling up. The new motor has a bit more grunt and longer legs. I had the CA set to cut off at 99A and 35 MPH, and the power definitely shut off at 35MPH and did not all more throttle until I got below 35. With the 3210 it maxed out at 30 MPH. When I looked a the IMax (amps) later it showed a max of 99 amps. 80 amps was max with the old motor..

Next I need to remove the speed limitation to see how fast it goes, and see if the watts / mile is significantly worse than the 26 wh/mi average with 30 MPH cruising that I was seeing last year. I really don't need to go faster than 30 MPH on a bicycle so I will probably be changing the final gear to a bigger chainring to reduce the max speed.

I finally did the cutting and brazing required to mount the disk brake as shown in the 1/20/2011 update above. Here it is still red hot right after finishing the brazing

Tolerances are very tight but it appears to be working properly and is not rubbing. Now I have great stopping power!


Here's the rear of the bike all put back together and painted with the new rear disk brake. Maybe I should repaint the brake caliper bracket too. I had to do some grinding on it to make the brake fit properly.

I have been commuting on this bike and with the new motor the bike tops out at a speed of about 47 MPH. I still try to keep the speed down to 30MPH on the commute as that is the speed I feel comfortable riding at with just a bike helmet. Also the motor overheats quickly at 47 MPH. It's nice to have the ability to go that fast though, in case of an emergency, so I'm not going to change the gearing. The bike seems to behave the same on the low speed side. I still need to get it up to 5 -10 MPH by pedaling before applying any significant e-power.

I bought a sensor kit and new sensored controller and installed it on the Astro 3210t motor to see if that will help the low speed performance. It's all put together but I have not yet had a chance to try it on the bike.

I added a sun visor to my Casco e-bike helmet. It's is made from a large PET plastic pretzel jar. It had the right curve to it. I cut it out with scissors and spray painted it black, then taped it to my ebike helmet's built in tiny visor.

So far so good. It has stayed attached up to 40 MPH, and does not add any perceptible drag. Plus now I can ride to work with both hands on the handlebars. Yay! I can see that the electrical tape will probably fail soon so I will need to decide if I want to glue it or screw it on permanently. I have not had good experiences with self adhesive velcro.

Today I did maintenance on the Saso (MEKS) 20" "carbon" suspension fork. It had gotten to the point where when I hit a large bump it would just stay stuck in the "compressed" position. It was pretty easy to partially disassemble. I removed the knobs on top of the stanchions, and the large nuts to expose the "guts", and carefully pried the rubber suspension tube seals out. This allowed be to squirt some bike oil into the inside and outside of the suspension tubes. I adjusted the "rebound" screw a bit. The fork seemed to slide fine and looked in good condition so I just put it back together. It seems to bounce without sticking now.

I am now thinking about new batteries. My RC batteries are approaching 3 years old and are still working well, but the balance after many bulk changes is not as good as it used to be. I am now using cell checkers to check the the cell voltages before and after charging to ensure that none of the cells are getting over or under changed.  Following the guidelines I have outlined above should keep me safe, but I have heard too many horror stories of people's garages or houses being burned down because of LiPo batteries bursting into flames. Battery technology moves slowly and only recently have I been made aware of a Lithium battery formulation called NMC. This formulation has the high current capabilities of LiPo, plus the safety of LiFePo4. The electric car makers (Chevy Volt) are now using this formulation. I am in the investigation phase but will soon be posting information about what the best way is to build or buy this type of battery pack.

I haven't taken a picture of this bike in a while. Here's a picture of the bike from this spring. I haven't ridden it at all this summer. I haven't been commuting.

I have ordered the parts to build a new 44V 10Ah battery pack using Samsung 24R 18650 cells. This should be a safe and powerful pack.

Also the silver paint had seen better days so I have since painted the bike black.

Here's the bike painted black, and with a new 18650 battery pack. The first shakedown cruise went well but I need to find a way to attach the battery pack more firmly. Here's a closeup of the drivetrain

I also sewed up a seat base cover and added more padding, and finally made a steel mesh cover  for the motor gears to keep my fingers out.

I haven't been riding to work this year so this bike mostly sits there with it's new battery pack waiting to be ridden. Yesterday I got the bike out, topped off the battery and pumped up the tires to take it for a ride around town. After my first fast take off the power cut off. Crap. After taking the pack apart and poking at it, the fuse holder wire just fell off of the battery switch. Massive corrosion of the wires had caused it to fail. After fixing that, I put it all back together and  tested it and got another few seconds of power, then nothing. This time it was the other side of the fuse holder, the wire had corroded all the way back through several inches of wire. Cheap junk from China. Made my own fuse holder using spade lugs and put it back together.

Today I had a nice ride down to the parade and the e-system seemed very peppy. I am a happy e-biker.

Continued on the 18650 page...


Back to the e-recumbent page