My first ebike build: custom-built KMX-based electric velomobile

[The Toecutter]

Update 11/14/21

Have since changed the battery to twin Greenway 46.8V/15.6AH packs in parallel, running the Phaserinner controller and Leafbike 4T motor at 2,500W. Currently get 150-200 miles range at 30-35 mph. Total unladen vehicle weight is 90 lbs and it is still very pedalable with the battery disconnected and the hub motor turned off, able to cruise 20+ mph on flat ground and sprint to 35 mph with no electric assist in use and the motor's cogging losses making themselves known. Turn the motor on, and top speed on flat ground is 45 mph +/- a few depending on state of charge, with car-like acceleration, and the ability to climb moderate hills at top speed.

Original Post

Pics of the bike:

So far here's the realized specs:

-46 mph top speed on flat ground on a full charge with the motor enabled
-1,500W peak power, plus rider effort
-7 watt hours per mile @ 30 mph w/light pedaling
-9 watt hours per mile @ 35 mph w/light pedaling
-12 watt hours per mile @ 40 mph w/moderate pedaling
-12 watt hours per mile in the snow with enough effort to do 30 mph on flat ground, but riding at speeds from 20 to 40 mph with a 20 mph rolling average
-While powered yields a consistent 40-50 mile range with enough effort to cruise at 30-35 mph on flat ground in freezing weather. I get about 8.5AH from the 46.8V pack in the cold, although it is rated to 10.5AH. This should improve to perhaps 60 miles range at the same speed in warmer weather due to reduced air density and increased battery capacity.
-With the motor disabled, I can pedal it to 35 mph in a sprint and cruise on flat ground around 23 mph. I lost a few mph unpowered after installing the motor due to its additional weight and cogging losses.

At any operating speed with which it is capable, it feels stable and tracks straight. It has been up to 60 mph downhill before I motorized it and felt stable. It is very controllable at 30-40 mph in 30+ mph crosswinds with 50 mph gusts, although one will need a bit of effort to keep it straight in those extreme conditions.

Regarding what specs this will eventually have, I'm in the process of building another battery pack. It's going to be 72V 21AH, capable of delivering 5 kW peak, and is liable to give me a 150-200 mile range @ 30-35 mph, acceleration like a car(simulation shows 0-30 mph in about 5 seconds with a 96A phase current, but a better controller and a 150A phase current could allow 4 seconds), and a 70 mph top speed. I’m also going to finish a set of wheel fairings, a windshield, roof, and rear wheel cowling to cut drag further. I also plan to add brake lights, permanent running lights, headlight, and turn signals, as well as solar panels and a charge port for electronics. Later upgrades may also include hydraulic disc brakes, thicker brake rotors, and perhaps a rear suspension.

Here's some additional specs:

-Weight: 82 lbs
-CdA: Unknown
-Trike Frame: KMX framekit
-Body shell: Custom corrugated plastic bodywork with aluminum ribbing, mounts, and supports
-Battery: 46.8V 10.5AH pack of 13S3P Panasonic NCR18650GA cells, 490 Wh rated capacity, 1,500W rated peak power
-Motor: Leafbike 1500W 4T wind brushless PMDC motor, w/10k NTC thermistor and 10-12 awg bullet terminals, laced to 26" rear wheel w/13ga spokes and double-walled aluminum rim
-Controller: Phaserunner v3, currently set to 1,500W maximum power output, 750W maximum regenerative braking
-Computer: Cycle Analyst v3 DP, currently set to 1,500W maximum power output
-Charger: GrinTech Cycle Satiator HV model
-Torque sensor: Sempu 2-wire w/115mm spindle width, no throttle installed at the time
-Front wheels: Velocity 20" 32-spoke w/13ga spokes laced to double-walled aluminum rims
-Front tires: Schwalbe Marathon Greenguard 20x1.5" @ 95 psi
-Rear tire: Schwalbe Marathon Plus Tour 26x1.5" @ 60 psi
-Front crankset: Suntour XCT Jr. with 152mm crankarms, modified with 26/39/53T chainrings
-Rear gearing: SunRace 34-11T 7 speed cassette
-Front derailleur: Microshift Mezzo
-Rear derailleur: Shimano Tourney RD-TX35 7/8sp
-Chain: KMC Z7, lubricated with Boeshield
-Torquearms: custom KMX torquearm designed and made by ES member dhwahl(left-side), GRIN Torquearm Kit v4(right-side)
-Right shifter: Shimano Revoshift 7 speed index grip shift
-Left shifter: SRAM 3-speed index grip shift
-Front suspension: EXA gas shocks on upper/lower swingarms at 65 psi, suspension designed and made by ES member adam333
-Brakes: Avid BB7 calipers with 165mm disc rotors, cable-pull operated
-Right Brake lever: Terratrike dual-pull lever with parking brake
-Left Brake lever: GrinTech ebrake, operated for regenerative braking in the rear

 

[Dauntless]

Are you sure you want to go 46mph in that? NO protection there. I see people with tadpoles who have a flag above eye level for the cars.

How are the steering and brakes?

And the third pic. Is there another vehicle in the distance?

 

[The Toecutter]

Are you sure you want to go 46mph in that? NO protection there.

Of course I do!

What I don't want to do is WRECK at 46 mph, or really any speed for that matter. If one were to jump off of a 6 story building, they'd smack the pavement at close to 46 mph.

I see people with tadpoles who have a flag above eye level for the cars.

The wind points the flag straight back minimizing the visual cue exposed to any drivers from behind. My vehicle has about the same ride height as my electric Triumph GT6, and is about as readily seen. The flag in my case would be next to worthless and add drag. The real threat is boobus Americanus not looking at the road and texting on their phones.

How are the steering and brakes?

The steering is perfectly adequate, although an inexperienced rider might find it a bit "darty" since it is direct steering using the stock KMX steering bars. I keep my hands on both bars the vast majority of the time, and one must be mindful of possible loss of control with only one hand on the steering if road imperfections are encountered.

The brakes are solid and will panic stop reliably and straight up until about 35 mph with minimal worries, with my once a month inspection/adjustment. They get a bit questionable at anything faster than that, although I have done a panic stop at almost 50 mph before I converted it to electric. I was careening down a hill on State St into East St. Louis when the light at the intersection suddenly turned red with only a 1-2 second yellow. I had about 200 feet to stop, and got pulled towards a minivan in the next lane over at the first application of the brake lever, and had to carefully modulate application of the brake while adjusting my steering to stop quickly enough to avoid going through an intersection with 50 mph cross traffic, while at the same time avoiding a 3 foot concrete wall/divider/median on my left, and a minivan at 40+ mph on my right. The next chance I got, I pulled into a parking lot and re-adjusted the brakes, after letting the rotors cool down.

Installation of hydraulic brake calipers and dual-pull lever will correct this because of their self-balancing tendency. That is on the list of things to upgrade to when finances permit.

And the third pic. Is there another vehicle in the distance?

It's a plastic child's toy car with an open bottom, no pedals.

 

[goatman]

hitting bumps like train track crossings at 46mph, how does the front suspension handle it?

like your build :thumb:

 

[The Toecutter]

hitting bumps like train track crossings at 46mph, how does the front suspension handle it?

I haven't hit a railroad crossing at that speed. The rear wheel is unsuspended and there is a possibility that scenario could cause a loss of control. I plan to add a rear suspension when finances permit, which will entail.a redesign of the tail section to accommodate it..

For run of the mill potholes, speedbumps, and road debris, the front suspension exceeds expectations. As long as the rear wheel stays on smooth pavement, the front can hit anything and it feels like being in a cheap truck or a mid 20th century British sports car as far as ride quality.

 

[goatman]

weird how you lost the axle nut on the cassette side

i remember you saying you used a rusty bolt for the spacer inside the cassette, inside the drop out

i wonder if thats what made the axle nut loose

if you change the Regen torque ramp up rate in the phaserunner from 200ms to 400ms it wont hit on as hard, regen will roll on smooth

 

[The Toecutter]

I'm going to try that.

I plan to do long-distance touring with this thing, regardless of whether I soon end up homeless or not, so keeping this thing reliable is of utmost importance. It is neither easy to carry nor inexpensive to tow or ship.

 

[goatman]

have you tried downloading phaserunner suite to your cell phone so you can make changes on the road

if you want faster acceleration, change the first setting

Torque Up Ramp to 50ms from 200ms

 

[Chalo]

I'm going to try that.

I plan to do long-distance touring with this thing, regardless of whether I soon end up homeless or not, so keeping this thing reliable is of utmost importance. It is neither easy to carry nor inexpensive to tow or ship.

Recumbents generally are a bear to push for any distance, too. At least your pedal drive and motor drive are independent of each other.

 

[The Toecutter]

have you tried downloading phaserunner suite to your cell phone so you can make changes on the road

I do not own a smartphone.

Recumbents generally are a bear to push for any distance, too. At least your pedal drive and motor drive are independent of each other.

This is the most significant motivating factor for adding a throttle. This way, if the bicycle drivetrain fails, I'm not stranded.

 

[goatman]

have you tried downloading phaserunner suite to your cell phone so you can make changes on the road

I do not own a smartphone.

Recumbents generally are a bear to push for any distance, too. At least your pedal drive and motor drive are independent of each other.

This is the most significant motivating factor for adding a throttle. This way, if the bicycle drivetrain fails, I'm not stranded.

smartphone doesnt have to be connected to a network, if you get one of your friends "old" smart phones you can hop onto free wifi somewhere to download it, then its just a matter of plugging the cell phone into the phaserunner

if you got a PAS potentiometer with your kit, it can be plugged into the throttle connector of the ca3 if your stuck in a pinch, it doesnt spring back like a throttle but it will get you home

https://ebikes.ca/ca3-aux-al.html

 

[The Toecutter]

smartphone doesnt have to be connected to a network, if you get one of your friends "old" smart phones you can hop onto free wifi somewhere to download it, then its just a matter of plugging the cell phone into the phaserunner

if you got a PAS potentiometer with your kit, it can be plugged into the throttle connector of the ca3 if your stuck in a pinch, it doesnt spring back like a throttle but it will get you home

https://ebikes.ca/ca3-aux-al.html

I have a thumb throttle and briefly tested it on the trike in order to determine energy consumption figures at 30 mph without pedal input. It uses about 400W from the battery to do 30 mph on flat ground. So my pedaling is accounting for almost half of the energy requirement at that speed. It has since been removed. I'm trying to clean all the wiring up and am going to re-route everything so that none of the EV components are readily visible, with the exception of the CA3. I have a keyed on/off switch now(viewable in the 5th photo with the hood piece lifted; it's red with a black mount) so that the CA3 shuts off as well.

I'll see if I can find an old smart phone. A year and a half ago I bought one in anticipation of doing Doordash deliveries to make money, but the company said I was not needed as my zip code was saturated with employees. I've since given it away to a friend who needed one, but that person might hand it back. I'm not a fan of them, but keeping it in a faraday cage will nullify its objectionable traits, and it will be available when I need it.

 

[The Toecutter]

I took a 30.4 mile ride today. Total energy consumption was 7.8 wh/mi from the battery according to the CA3. I was cruising at 30-35 mph as normal, and don't recall catching any red lights during the ride and had very few stops. This thing is very efficient.

I'm about to make a pack of two used Greenway 46.8V 15.6AH packs with BMS wired in parallel with each other and increase my range to 150 miles while only adding 9 lbs to the bike. I'll also be able to increase the power to 3 kW.

 

[The Toecutter]

The batteries arrived a week ago. A printout of their delivered capacity during testing from the seller was provided. They were 15.51 AH and 15.35 AH respectively. This means 30AH when I parallel the two together is a given.

I hadn't had much in the way of free time so I didn't get to fully install them. I need to make a wiring harness to parallel them together AND I need to make a custom enclosure to fit both of them lengthwise underneath the boom with their narrowmost dimension placed between the rotating pedals. Being a holiday weekend, all the places I could get the needed parts were closed, so I decided to devote the free time I had after 3pm today to testing the vehicle.

I was able to fit ONE of the battery packs underneath the boom. I used the higher capacity pack of the two, held to the boom with zip ties. To say I was pleased with the results would be to understate things. I rode it 83.6 miles, with cruising speeds of 30-35 mph plus pedal input. At the end of the ride, I had used 15.5 AH, and had consumed 8.7 wh/mi. The battery wasn't dead, but it was on the verge of being empty. I plugged in my charger for about 20-30 minutes at an outlet at a pavilion in a park, got a few AH back, and then rode it another 20 miles.

I now have a decent range EV!

Once I put both packs in, a 150+ mile range is assured. Plus it's very pedalable with a dead battery and I can leave spandex fetishists on roadbikes behind in most settings. I won't have any range anxiety using it to visit job sites anymore and won't have to charge it every day, and can opt to just change when it is convenient instead.

 

[thundercamel]

Very nice efficiency and range!

 

[The Toecutter]

I've been working so damned much with this new job I got 2 months ago that I've had no time to work on this between my job and caring for my mother who can't walk. I use it as my daily transportation and cannot currently take it out of commission for the multiple days I will need to make it to my friend's shop, hack the bike apart, and install a rear suspension. The amount of miles that have been placed on it since converting it into an EV early this year is approaching 6,000, and the KMX chassis is getting pretty damned tired as it will be hitting 60,000 miles sometime this year at the rate I'm racking up miles. Something's going to fail sooner rather than later, and all the creaks and groans the frame and seat rails and boom make when I'm riding it at speed are making themselves heard. adam333's front suspension is still holding up rock solid at least.

I am considering swapping in a 3T wind 1000W Leafbike motor to cut down on cogging losses when operated with the motor shut off, which will free my current 4T wind motor to be put into another trike. I do not have another trike yet, but am seriously considering purchasing a Milan SL Mk II or Mk III velomobile if I can find one(not many of them in the U.S. or Canada), mostly to use as a backup trike to get me to work so I can actually take the KMX out of commission to work on it, and also to study for design cues on the next iteration of my body shell being that the Milan is among the most aerodynamically efficient human powered vehicles commercially available. My current machine uses 400W to do 30 mph, whereas a Milan does it on 150W. However, I can't fit 2.5" wide solar race car tires on a Milan nor is the front track wide enough for it to be able to corner as well as a car without tipping over, nor is the suspension as robust as I would like, so it's not exactly the "ultimate" velomobile for my application. That one, I still have to build, but I think getting my wattage requirements down to 200-250W @ 30 mph is doable with improved aero, even with a wide track and lossier tires. The outboard wheels and panel gaps are really dirtying up the airflow and mine probably has 5-6x the aero drag of a Milan. That said, both the Milan and the WAW are the two velomobiles that most closely fit what I want to do, but the WAW's aero is nowhere near as good as the Milan and I'd probably not be satisfied with it.

I believe in going full retard on efficiency. The math I've done says a Milan could be set up with a street legal 750W 28 mph limited system and STILL be faster on an overall basis than my corovelo as it currently is when operated as a PAS. My CA3 suggests I'm putting out 200-250W with my legs when cruising at 35 mph with the motor doing the rest, whereas with a Milan, in a 28 mph limited system, I'd do like 35-40 mph on flat ground with 200-250W of pedaling and the motor adding nothing with a 28 mph cutoff. In a Milan, the motor would mainly be for going up hills and accelerating, but on flat ground, once I'm at speed, my legs will do all the work. I'd probably only need about 3-5Wh/mile in a milan, which would allow me a 150-200 mile range with just ONE of the Greenway 15.6AH packs in my possession. But the WAW would not be nearly as fast set up like that, although if I set up a WAW as I have my current corovelo with no speed limiter and a 1500W limit, it would then be faster than the corovelo. A WAW is obviously easier to work on than a Milan, so it has that going in its favor. My KMX velomobile has removable tail and nose, inspired by the WAW, which indeed makes it easy to work on, although unlike what I've read about the ease of removing body pieces on the WAW, it still takes me a few hours to remove the nose and tail pieces on my trike.

But now that I have the money to finish the KMX, I have no time to actually work on it. If I hack it apart to install a rear suspension and remove all the body pieces, it's going to be out of commission for at least 2 days if I actually had two contiguous off days to work on it uninterrupted, which would not be a good decision when I use it to get me to work and everywhere else. Nevermind the fact that I've never had 2 days of leisure time in the last two months anyhow and would probably have to split the work up over the course of weeks or a month. So an expensive velomobile purchase, just to have a backup trike, is very tempting right now...

 

[goatman]

did you get a pattern online to cut the coroplast body

 

[The Toecutter]

did you get a pattern online to cut the coroplast body

No, but I do have a notebook full of pages with the dimensions of all of the cut pieces. That design is far from optimized. I was drawing up a 3rd design iteration that does away with the outboard wheels and will allow the airflow to be flush with the wheels, albeit the wheels won't be covered with the body. The sketch sort of looks like a mix between a Milan, a DF, and a WAW. It's designed with a removable nose and tail as well as easy serviceability in mind. AND because I want to add a gas shock in the rear and recline the seat back further to reduce frontal area and lower the center of gravity while extending the wheelbase, I will need to redesign the whole thing, but I can't do that until I have the trike frame physically in front of me assembled with the rear suspension and new layout to take measurements with which to make my next design.

My next velomobile may be a Milan SL. I'll be able to use that while I completely overhual the KMX.

I don't expect my next build will be as slippery as the Milan, BUT it can come close and I'll be sure to make it ready to handle highway speeds and 10 kW of power and it's going to have the storage space of a compact car. My next design will be more practical as well as faster than my current one. The Milan will act as a design guide should I obtain one. A friend of mine is currently setting a 3D printer up large enough to print a body shell in thirds, so I'll be able to make a nacelle with a roll cage that bolts around the KMX frame, and then bolt my nose and tail pieces to it.

As for the Milan, if I make the purchase, since I don't want to destroy it, I'm going to add a MAC geared hubmotor and restrict it to 750W, 28 mph, and 40A phase current. It will keep it street legal, and most of the time, I'll be on 100% pedal power somewhere over 30 mph, with the assist mostly used to get up hills and aid in accelerating up to speed. A 700Wh pack would probably give it a 150+ mile range in that usage case and it would still be very pedalable with the motor disabled without the cogging losses of my Leafbike motor. It would be every bit as fast as my current 1500W corovelo in that configuration, maybe faster.

 

[The Toecutter]

I bought a Milan SL:

I just need to find out where I can get a MAC 6T, because they seem to be discontinued. Any suggestions on a geared hub motor with a clutch that can decouple the motor from the pedal drivetrain? A GMAC won't work. If I'm going to have to deal with cogging losses, I may as well order a Leafbike 1000W with a 3T wind.

Here's some more pics of the custom build, which has almost 9,000 miles added since conversion to electric in January:

 

[HK12K]

Niiice. Lucky duck.

Fwiw I have a standard 4t 1500w leafmotor in a 20" wheel, 18s lipo. Hot off the charger it does 47mph without field weakening and well over 50 without. On a Giant Revive. I haven't really explored the upper limits as of yet. Perhaps once I get a motorcycle helmet and armored jacket.

Leaf actively discouraged me from ordering the 3t, and I can't say that I'm disappointed in the 4, but ymmv.

 

[goatman]

The Green Machine

 

[ZeroEm]

by The Toecutter » Aug 01 2021 11:39am

I bought a Milan SL

That is nice, well done.

Have you thought about going up in Voltage with a mac or eZee motor. You would not need all them amps that a low T motor needs and still get the speed.

 

[The Toecutter]

That is nice, well done.

Have you thought about going up in Voltage with a mac or eZee motor. You would not need all them amps that a low T motor needs and still get the speed.

I was going to run a MAC 6T at 72V. I'd have an "off road" mode for locations where I can get away with it and fly around at 65-70 mph. But where I need it, I could have it remain "street legal" with a 750W/28 mph limit, where above 28 mph, it is 100% human powered, wherein with the cogging torque eliminated by a clutch mechanism, I'd be able to maintain 35+ mph on flat ground or downhill, while the motor would be there for climbing up hills or accelerating from stops. With an FOC controller like the Phaserunner, the voltage is varied to match the operating point, so as to prevent the motor from running inefficiently. I don't need stump-pulling acceleration in the Milan, and don't want it because I want to avoid destroying the rear swingarm. In my custom build, 350W to the Leafbike 1500W 4T wind plus hard pedaling is enough to keep up with traffic! So 750W in a geared hub in a Milan that is much more efficient at transfering pedal power to the drive wheel would be more than enough.

As it is without a motor in it, I have a higher top speed in the Milan on flat ground than I do in the KMX at a full 1500W of assist. 350W of assist in the KMX makes it comparable in performance to 30 mph with the Milan as long as the gradient is < 5%, but above 30 mph when terrain permits, the Milan is still faster. I can reach almost 48 mph on the flat in the Milan. I'll get stronger. Some people can do 60+ mph in one on flat ground with a long enough stretch of straight road, unmotorized.

Once the Milan is usable as a daily driver, I can then proceed taking the KMX apart.

I'm eventually going to build a new shell for the KMX based on the Milan, add a roll cage, make some custom spindles that can accommodate cotter pin axles, add a rear shock, have Mitas MC2 moped tires all around, and turn it into a pedal "car" of sorts that can do 100+ mph, but still be pedaled to 40+ mph in a sprint with the motor disabled. This one won't quite be legal, but I'm going to have everything hidden so well no one will be able to tell it has a motor without taking it apart. I want to make a transportation appliance that gets around 20 Wh/mi at 70 mph, and can do 0-60 mph in 5 seconds or less, where human power can add to its thrust at all operating points.

Eventually, my plan is to build a completely custom platform dependent upon the success of this project, one that can accommodate a pure unmotorized velomobile of about 70 lbs, a "street legal" e-assist velomobile of about 85 lbs, an illegal high performance e-assist velomobile of about 100 lbs, and an AWD sports car with more than one horsepower per pound of vehicle with no bicycle drivetrain of about 120 lbs. The idea being to broaden the potential market for the same platform to help get the per unit cost down, in the interest of making an affordable velomobile.

 

[ZeroEm]

I'm here in Texas, have my speed limited to 46km/28mph with a three way power switch. Keep it on low for hills and power savings.

I run a 7T leaf motor at 72V in a 26" wheel, it will cruse at 33mph and it's most efficient at 15mph. look where your torque curve crosses your power curve, this is were you want to ride. You can set it up working with different windings, Voltage and tire size for any setup you want. I'm heavy and to tall for the Milan, set my trike up to pull hills not speed. Thinking about going to a 6T leaf. I don't feel cogging with my motor but I don't try to ride at 20mph with it off.

 

[The Toecutter]

look where your torque curve crosses your power curve, this is were you want to ride.

With modern FOC controllers I don't have to worry about this. It will adjust the voltage to match whatever speed I'm riding at. This way, I can run a 72V pack without being penalized when I'm doing 15-20 mph, but have all of the performance it is capable of whenever I demand it.

About to order a rear suspension and take this KMX out of commission for a while. It will be back in commission minus body shell once everything is installed, and I'll need to go through the laborious process of re-designing everything to accommodate the new specs imposed by the upgrade. It will be more reclined with a slightly lower center of gravity, slightly smaller frontal area, and a 2" longer wheelbase or more. Looking to install Mitas MC2 16"x2.5" tires on all three wheels, will be swapping out the 4T 1500W Leafbike motor built into a 26" rear wheel for a 3T wind 1000W version built into a 16" cast scooter wheel, adding hydraulic disc brakes for the front wheels with a cable pull emergency brake on the rear, and will be running a 20S7P pack of Panasonic 2170s. I'm getting the freewheel version of the motor to keep the axle length the same on each side, and will be running a DNP Epoch 34-11T 7sp freewheel, and adding a Schlumpf HS drive with a triple crank up front, and replacing the torque sensor with a model that accurately reads the torque at the crank, and not merely estimates it based on tension.

The new body shell is going to resemble a sort of cross between the DF velomobile, the WAW, and the Milan, and I have a Milan right in front of me to reverse engineer. Hopefully I get it efficient enough to cruise at 30 mph on 250W, 60 mph on 1.5 kW, which would be about 2x the drag of the Milan. There will be at least 150W of flexible solar panels on it. Aiming for a slightly higher 5-6" ground clearance with the exception of the chain tunnel and rear derailleur cover.

IF it proves to be too cumbersome to pedal with the motor turned off, I may remove the bicycle drivetrain altogether, install some Grin all-Axle motors up front, strip a Hubmonster out of a scooter for the rear wheel, and turn it into an AWD hoonabout with ridiculous acceleration. We'll see. I'd prefer to keep it a pseudo-bicycle and retain the bicycle functionality so that there's never any range anxiety and so that it is versatile(after all, there's a lot to be said for not being reliant upon the EV drive to use the vehicle), but if the bicycle bits prove to be decorative instead of functional, then it's of no use to keep them.

In any event, I'm still planning to build an AWD sports car version of this idea at some point, for which the bicycle drivetrain will be useless and therefore absent. I'd just prefer it to be a separate vehicle, as the human-electric hybrid concept has an appeal all its own.