PaulD's lightweight mid drive development

[PaulD]

Hey ES,
I've been working on this for a while, so I thought I should share a bit. I have been developing a mid-drive system with the intention of converting most road and mountain bikes with minimal weight increase. Specs as they are now: 250-350W, ASI FOC controller, torque sensing BB, 36V, 11ah, about 9lbs with battery. Motor is about 3.5 lbs. Planetary 1st stage, #25 chain secondary to crankset. Battery (should) fit where a standard size water bottle would normally go. On bikes with multiple bottle locations, multiple batteries could be installed.
Pictured is a beta prototype with all of the plastic bits just 3D printed, and no finish on the metal bits. I'll be working with my Industrial Design buddy to make it look less prototypie.
This is intended for riders that want a lightweight, street legal drive, and aren't interesting in motorcycle speeds. I'll be posting some more pics/info over the next few weeks, but I wanted to show a few of the kit installed before I head down to the Sea Otter Classic for the emtb race. Keep observers will note that I need to slide the battery up (the mount is slotted and adjustable) so that the swingarm doesn't hit the battery.
Cheers,
Paul

 

[liveforphysics]

Nice job buddy!! I pity anyone who rides against you, as I know even with an unpowered ebike you would beat most ebikes!

 

[PaulD]

Ha! Thanks Luke! There are some super fast guys that do the emtb race. Last year a former world champion won! .. funny you mention riding unpowered. Apparently last year, Adam Craig (super fast pro racer) just taped an extension cord to his bike and finished mid-pack. Lol.

Sent from my Pixel using Tapatalk

 

[PaulD]

Well, the race was a blast, and the bike performed great except for a jammed chain which cost me 4 minutes... That's an easy problem to fix. In my rush to complete the bike, I adjusted the chain line without realizing I made the perfect size slot between the chainring and powered secondary sprocket for a jam
With a short race with half of it on flat ground, and a 20mph electric cut-off, there was a lot of pedalling. Pro racers dominated. Only used 3.6ah out of a 10.5ah pack.
I figure that I had one of the lightest bikes, but it didn't make up for lack of fitness. The rules stated a power limit of 750W. My drive was programmed for 350W conservatively (hadn't had time to experiment with higher power and the associated stability tuning)... bumping this up would have helped make up for the lack of leg power. I figure there were a few bikes that were exceeding 350W, but its hard to tell when the riders are so strong.

Ended up charging in the registration building, as all of the outdoor plugs were off.


Having a lighter bike and a dropper post allowed for some fun times through the rock garden and railing corners:


Next year, I'll hire a pro racer and make sure the chainguide is set up right!

 

[liveforphysics]

You rock Paul! Congratulations! I would be happy to turn my Nyx bike down to 750w and race with you next year!

 

[spinningmagnets]

Congratulations! Glad to hear you had a good time...

 

[flathill]

torque spec?

helical or straight planetary?

 

[ScooterMan101]

+ 1

and

What make and model of Crank ? It does come a freewheel , or did you make a freewheel for the crank ?

What kind of Bottom Bracket , Square Taper or ?

Cost of the Crank , with BB and Freewheel, and chainrings ?

 

[dtx1]

Awesome build. Can you talk a little about the chainguards you used? They look very well done!

 

[PaulD]

Thanks Ron and Luke! I think its feasible to make a winning bike for next year.
Flathill - Theoretical torque at the crank is 60 Nm with the current settings. I haven't confirmed. The limiting factor will be the gear reduction. Gear reduction is a compound planetary 1st stage is helical, second is straight. I am using a motor from a vendor that builds geared hub motors, but with custom winding and lamination thickness. The gears will look familiar to anyone that's opened up a Q100.
ScooterMan - At the moment I am just using standard square taper freewheel cranks from Sick Bike Parts. The freewheel is from them as well. I am working on a lighter weight solution at the moment, but I haven't decided if it will be something I introduce later, after I make the first small batch production run. The NCTE torque sensing BB is by far the most expensive part. They retail for $240. The #25 sprockets are pretty cheap, about $12-$16 each.
Dtx1 - The front guard is 3D printed high impact PLA. Designed to prevent anything from getting sucked into the chain/sprocket. The rear is CNC'd Lexan for the moment. It's quite thin, so I didn't think a 3D printed part would be strong enough. It bolts directly to the #25 sprocket.

 

[ScooterMan101]

Paul,

The solution that I have been thinking about for a while now, to cut down a pound or more off the BB/Crankset , and , to have better quality BB Bearings and better quality Cranksets ,
Is
To Design a inner sprocket/feeewheel, that bolts right onto existing , Modern 2 piece Cranksets. ( BB30 / Shimano Hollowtec II / Sram Mega / )

They are all available in 3 chainring and 2 chainring configurations . now take off the inner stock chainring, and bolt to the BCD Bolt Pattern, a chainring/freewheel .

You are just taking a 3 speed or 2 speed crankset and making it into a 1x front to the rear, ( outer larger chainring 50-54 tooth to the motor ).

there are about 3 different BCD's , 130, 110 , and the other I forgot as it is used on Mountain Cranksets, so just 3 Different Spiders so as to fit up to make a freewheeling chainring/crankset .

>

I do not use torque sensing , I would guess that many here on E.S. do not, we use throttles. so we would not need that extra expense . just throttle hook up to the controller. and perhaps a PAS as a cheaper alternative for those who do not need torque sensing. So 3 different options for your customers . Throttle only, PAS and Throttle , and Torque Sensing. Three different price points .

ScooterMan - At the moment I am just using standard square taper freewheel cranks from Sick Bike Parts. The freewheel is from them as well. I am working on a lighter weight solution at the moment, but I haven't decided if it will be something I introduce later, after I make the first small batch production run. The NCTE torque sensing BB is by far the most expensive part. They retail for $240. The #25 sprockets are pretty cheap, about $12-$16 each.

 

[PaulD]

The solution that I have been thinking about for a while now, to cut down a pound or more off the BB/Crankset , and , to have better quality BB Bearings and better quality Cranksets ,
Is
To Design a inner sprocket/feeewheel, that bolts right onto existing , Modern 2 piece Cranksets. ( BB30 / Shimano Hollowtec II / Sram Mega / )

I do not use torque sensing , I would guess that many here on E.S. do not, we use throttles. so we would not need that extra expense . just throttle hook up to the controller. and perhaps a PAS as a cheaper alternative for those who do not need torque sensing. So 3 different options for your customers . Throttle only, PAS and Throttle , and Torque Sensing. Three different price points .

I am working on a solution to fit a SRAM direct mount crank. A freewheel is a significant engineering effort, but making the diameter larger should allow me to use hard anodized aluminum instead of hardened steel bits... keeping the torque sensing strategy to myself for now ... I think torque sensing is critical on a mountain bike when riding technical uphills. I want the complete bikes to be street legal, so that rules out throttles (mostly).

 

[Norton]

... The NCTE torque sensing BB is by far the most expensive part. They retail for $240. ...

Paul,
Your build here looks really nice and is now proven! Congrats!
I'll be following this thread!
I agree with you on Torque Sensing with a mtb. Just shifting in general sound goofy with a simple PAS system.

I looked at the NCTE BB. One model had "Direction of Rotation" as one of the sensors. I thought wth is that for?
Then it struck me ! Regen braking on a Hub Motor ebike !! Slow back-pedal = small regen, faster back-pedal = max regen! Ha, ha!!

 

[PaulD]

Your build here looks really nice and is now proven! Congrats!
I'll be following this thread!
I agree with you on Torque Sensing with a mtb. Just shifting in general sound goofy with a simple PAS system.
I looked at the NCTE BB. One model had "Direction of Rotation" as one of the sensors. I thought wth is that for?

Thanks Norton! Direction of rotation is needed for torque sensing middrives too, since this type of sensor only measures leftcrank torque. Otherwise you would get a throttle response when standing on the pedals (with your left foot forward) and rotating the cranks backwards a bit... Using it for Regen on hub motor would be hilarious though!

Sent from my Pixel using Tapatalk

 

[ScooterMan101]

PaulD,

What is the Progress level ?

What all will you be selling with the kit ? ( Sram Direct Mount Crankset as well ? )
( Any other sources for a Direct Mount Crankset besides Sram at this time ? )

Many of use 12 s packs, or even 14 s packs, so the controller will be adjusted for Voltage, and Amp draw
by which Means ... Bluetooth and a Smart Phone ?

The controller will be Sinewave ?

Smart Phone will be the Display for Speed/Watt Draw/Amp draw ?

 

[PaulD]

PaulD,
What is the Progress level ?
What all will you be selling with the kit ? ( Sram Direct Mount Crankset as well ? )
( Any other sources for a Direct Mount Crankset besides Sram at this time ? )
Many of use 12 s packs, or even 14 s packs, so the controller will be adjusted for Voltage, and Amp draw
by which Means ... Bluetooth and a Smart Phone ?
The controller will be Sinewave ?
Smart Phone will be the Display for Speed/Watt Draw/Amp draw ?

Progress is slow right now, working out issues with BMS, wiring, motor assembly, etc. The display is TBD. I initially decided I didn't need a display, but I think the ability to change modes is important enough to warrant a display/switch.
Battery will be 10S. I don't see the need to go higher. Controller is from ASI, so it is FOC sinewave.
As for other direct mount cranks, I think Praxis makes them. Unfortunately most brands have their own "standards" like RaceFace, FSA and Cannondale. SRAM seems the most common though. Making a freewheel thin enough to fit in the space of direct mount crank spider is challenging at best.

 

[ScooterMan101]

Now that I am using Lipo's I have no problems with a BMS, they have none to mess things up .
( There have been a few fires from malfunctioning BMS's on E-Bikes with 18650 cell packs , look at the battery section of E.S. )

I use a cell checker/alarm for LVC with my Lipo's so no problem with having the batteries going to low, besides that is another reason to have a display like the C.A. or smart phone on the bike , with it's Voltage Ranges and LVC , etc. , functions .

Since you are designing a Motor Kit , Just Design around the newer make it your self battery pack , ( VRUZEND , Mlt34 ) so that you can have a quick disconnect on the BMS , so that the user can swap out a BMS when it does not work properly, and while you are at it , solder up some Balance Wires to each cell so that if a BMS malfunctions during charging , then using a Charger with Balance Function / JST-SH Connectors, will not allow it to over charge in the first place .

Does the ASI controller have a Bluetooth Dongle or can communicate with a Smart Phone/Small Tablet via Wi-Fi ?

You are using only 10s pack ? , if so I hope the controller and motor does at least 25-30 amps continuous, at the least. There are some steep hills I need/want to go up. I already have experience with trying to go them at 25 amps and that is just not good enough 30 amps would be more like it, if going up paved roads then up to 40 amps.

Yea I know that we would need a few less amps on a Mid-Drive, over a Hub, But still 25-30 amps at least .

Since Bicycle Cranksets are not made for having a freewheel, although if you use an older 3 chain-ring Crankset , that is not direct mount, isn't there enough space if you get rid of 2 of the chain-rings , put the freewheel on the outside near the crank arm and chainring on the inside to have a better chain line ?
If not then it would probably be best to contact some bike component Mfg in Taiwan or China to make a ( modern Lite Weight ) crank set that is designed around a freewheel. Hint: have the drive side crank arm with a arch or notch cut so as to allow room for the freewheel .
For BCD's a 110 BCD will allow us to buy off the shelf Cyclocross Chainrings up to 48 tooth and up to 52 tooth Road Charnrings, mostly for FSA / Sram Cranksets/Chrinrings , and
130 BCD for Shimano and Aftermarket Road Size Chainrings.

Progress is slow right now, working out issues with BMS, wiring, motor assembly, etc. The display is TBD. I initially decided I didn't need a display, but I think the ability to change modes is important enough to warrant a display/switch.
Battery will be 10S. I don't see the need to go higher. Controller is from ASI, so it is FOC sinewave.
As for other direct mount cranks, I think Praxis makes them. Unfortunately most brands have their own "standards" like RaceFace, FSA and Cannondale. SRAM seems the most common though. Making a freewheel thin enough to fit in the space of direct mount crank spider is challenging at best.

 

[PaulD]

Since you are designing a Motor Kit , Just Design around the newer make it your self battery pack , ( VRUZEND , Mlt34 )

I want a pack that will fit where a standard size water bottle fits... and I want it lightweight. Since this is a low power system, I can get away with smaller 30 cell battery. I can't find anyone making a pack like this.

Does the ASI controller have a Bluetooth Dongle or can communicate with a Smart Phone/Small Tablet via Wi-Fi ?

Yes, the ASI has bluetooth, but it doesn't have an end-user interface.

You are using only 10s pack ? , if so I hope the controller and motor does at least 25-30 amps continuous, at the least. There are some steep hills I need/want to go up. I already have experience with trying to go them at 25 amps and that is just not good enough 30 amps would be more like it, if going up paved roads then up to 40 amps.

The intention is for this to be a lightweight 350W max assist. This more than doubles most peoples power output. There are plenty of high power kits out there for those that don't want to pedal hard. My kit is for people that just want electric to augment their efforts, not replace them.

Since Bicycle Cranksets are not made for having a freewheel, although if you use an older 3 chain-ring Crankset , that is not direct mount, isn't there enough space if you get rid of 2 of the chain-rings , put the freewheel on the outside near the crank arm and chainring on the inside to have a better chain line ?

The spider of a non-direct mount crankset gets in the way. There generally is very little room between the chainrings and the crank arm to fit anything in there. If someone wants to use a big chainring, it usually has to go on the outside (big ring) position for chainstay clearance.

 

[briangv99]

A freewheel is a significant engineering effort

Hi Paul, very elegant design, I think it'd go great on my Yeti SB5.5. Regarding a freewheel, since you're going PAS only, could you not bypass having a freewheel altogether. A torque sensor could be added to the lower/return side of the #25 chain where it would only be feel the torque from pedalling, not the motor. You'd only need to adapt a #25 sprocket to the crank spider. At 350 watts a motor runaway should be less of a concern, and other safety measures/motor cut off could be built in. Just a thought.

 

[PaulD]

Hi Paul, very elegant design, I think it'd go great on my Yeti SB5.5. Regarding a freewheel, since you're going PAS only, could you not bypass having a freewheel altogether. A torque sensor could be added to the lower/return side of the #25 chain where it would only be feel the torque from pedalling, not the motor. You'd only need to adapt a #25 sprocket to the crank spider. At 350 watts a motor runaway should be less of a concern, and other safety measures/motor cut off could be built in. Just a thought.

Thanks Brianv99. I think think riders with nice lightweight mtbs and road bikes will like a kit like this.
Since there isn't any power (torque) transfer from the rider through the #25 chain, you can't use it to measure torque. All of the riders power goes to the bike drivetrain. When starting up, the rider's power spins the #25 chain, but there is only the resistance of the motor freewheel to overcome. The new Rocky Mountain PowerPlay senses in a similar way you describe, but the bike chain goes around the crank and motor sprockets rather than having a separate chain for the motor.. this allows torque sensing on the section of chain between the crank and motor.

Regarding whether a freewheel is even needed.. well I had this debate with Justin, and it's worth a shot. It will take some clever controller tuning to make it seamless. I need to spend some time experimenting to find out if it is possible.

 

[briangv99]

thanks Paul, of course, I didn't think that through :?

Rocky Mountains drive is interested, could lend itself to a kit option, but would probably need a lot of reduction at the planetary to bring the rpm down

 

[PaulD]

It's been a while since my last post on here. Not for lack of effort on the development side, just multiple changes in direction to get to a better solution. With some Sempu T4 bb sensors on the way, I am working on freewheel solution. High pole count cadence sensors and torque sensors that sense both sides lead to a very intuitive ride, very close to the best commercial systems out there. Here is a bit of progress on a 3D printed freewheel proto. It has to be very large diameter to clear the crank and BB cup, so no off-the-shelf solutions are available. Hoping to have CNC'd parts in a few weeks. Any machinists out there interested in making these parts?

 

[ScooterMan101]

Paul,

What Bottom Bracket / Crankset ( Make and model )
is that BB Cadence Sensor made for ?

B.T.W. I have for quite a while now been wanting a Freewheel system for the very popular Shimano Hollowtec II Cranksets, that are used on both Mtb's and Road bikes, ( 24 mm spindle )
and
perhaps for the somewhat new Race Face Cinch cranksets .

However as if that is not enough design work to do, now comes ...

Sram Eagle NX , the first , somewhat affordable , Eagle 1x 12 drivetrain.

Sram makes it to be e-bike capable ( heavier than the XX1/Xo1/GX Kits )
and very important somewhat affordable price of $ 380 for the entire drive train ( Shifter/Derailleur/Crankset/Cassette 11-50t /Chain ) except for the BB since there are many different BB shells on bikes these days.

Looks like Design work for a Freewheel crankset needs to shift over to Shimano and Sram Eagle .

 

[ScooterMan101]

Paul,

Design a freewheel for the Shimano Hollowtec II 2 and 3 front chainring cranksets
and
for the Sram Eagle Crankset .

for those two freewheel cranksets, I will talk to a Machinist I know to make the BB Cadence Sensor.
He is busy, as most all Machinist's , however he is starting to get interested in e-bikes . He needs to make at least 4+ at a time , count me in for 1 as long as it is for a Shimano or the Sram Eagle NX .

 

[Drum]

Hi Paul, I am interested in your design as I also really like the torque sensing control for e-mountainbikes.
Talking about a system with no crank freewheel, in many ways I find the thought a bit scary, but on a bike without a throttle and with a relatively low powered motor it should be possible.
I recently rode a bike with a "backpedal brake" (coaster brake) version of the Tongsheng TSDZ2 motor which has this setup: the chainring is bolted to a large output gear that is directly driven by the pedals (through the torque sensor) AND by the motor (through a freewheel). The only thing causing the motor to run is output from the torque sensor.
When I pedalled hard, then stopped suddenly, I could feel the pedals kick forward briefly as there was a minor delay in the sensor / controller dropping output to zero, but though it was noticeable it was not dangerous and I quickly got used to it.
I would prefer a crank freewheel, but as long as the sensor and controller combination could be set up to drop output to zero very quickly, it could be worth a try.
It looks as though you are well on the way to developing a crank freewheel, which is great, but I thought I would add my experience to the mix in case it is useful.
Looking forward to developments,
Dave