Bob's very simple <1kW friction drive

[bobc]

I thought I should start a thread about this project.
The idea is to make the cheapest and simplest possible legal ebike conversion.
I want to cover the whole thing including electronics.
So here is a picture of the mechanical assembly: it is intended to fit onto the seat tube inside the rear triangle.
The separate drive roller is smaller diameter than the outrunner rotor in order to run efficiently at legal speeds.
The motor is a cheap outrunner from hobbyking or elsewhere. I've used a propdrive 50/60 270kV (from hobbyking $35)
The Controller is a 60A RC ESC with BEC 6s capable. I've bought a "mystery" one (£10.50 off ebay)
Initially I will mount a single 6s 5Ah lipoly battery - this should be good for 7 miles or so on the flat with no pedalling (15.5mph). Cost about £40.

Custom parts are
1) drive roller with taper lock fixing to motor shaft
2) swinging motor mount frame, laser cut stainless steel
3) seatpost clamp parts, fashioned from 10mm Ally sheet
I'll post drawings for the bits as I do 'em, so if it turns out OK anyone can copy it.

 

[bobc]

I made these today on our Zoltar CNC router. We used a 3.2mm single flute bit in the Kress & split the job into 3 parts;
1) "pocket" the rounded profile to clamp the seatpost (if you don't get this profile right you squash the tube when you tighten it and the saddle won't move any more)
2) mark the holes
3) profile the outline. I left 3 tags on the outline to stop the workpiece flapping about on the final cut.
Of course this part could be made many other ways - one could water jet the outline and holes then flycut the groove or simply go at it with hand tools.
I include 2 dxf files, one is a drawing of the part in dxf format, the outline one can be used by any automated profiling equipment e.g. water jet cutters or CNC routers

 

[EVTodd]

Looks good bob. One thing I would recommend is supporting that friction roller on both sides. The outrunner's bearings won't last long with it hanging out the side like that.

I built a similar (electronics wise) setup for my wife and we've had great luck with it. At kv270 on 6s with a 1.25 inch roller she gets a 19-20 mph top speed.

She has been using a mystery 60a esc, with one cap added to help with ripple, for a couple of years now (at least). She's doesn't weight much though. I would say for an adult male you might need a bit larger esc.

We started with 5s with this setup. It was actually fine for her, but again, she's light. I can't imagine 5s working well for anyone over 140-150 lbs.

Good to see someone not using the can of the motor for a change. Imo, the diameter of the can is simply too big for a useful assist at lower power levels.

 

[bobc]

Yeah 20mph is about as much as you'd want on a "hardtail" - particularly a roadie with narrowish tyres.
Others (SM) have also suggested more support for the roller, so I won't be surprised if there turns out to be an issue there. But the one sided system is so wonderfully simple & easy to make... I'll give it a go with a view to beefing up the bearings at the back of the motor if it all starts to fall apart. My hope is that I'll get away with it because I'll be running low power levels, low speeds & the motor is very lightweight - fingers crossed

 

[spinningmagnets]

For newer posters who aren't familiar with the 2010 "year of the friction drive", Todd was the first to make a friction drive using appropriate RC motors. Friction drives have been around for 100 years (using "chainsaw" motors and very large electric motors). The small RC motors are a perfect fit for this application.

In the middle of Todds original thread, Kepler posted he thought the motor-shell itself could be used as the roller, which...if possible, it could make the drive more compact. Kepler and Adrian_sm both developed tiny and light friction-drives that used the motor-shell as the roller.

No drive is possible that could be smaller or lighter. However, from the very beginning, road testing showed that motor-shell drives are most appropriate for mild-assist, which actually fits the demographic of a very light road-bike very well.

For a daily commuter, I agreed with Todd from the beginning that for the very small increase of size and weight, having the roller as a separate component was best. The roller diameter, motor diameter, and motor-kV could each then be chosen based on the best choice, rather than a series of compromises. However, friction-drives with separate roller have been tried that were only supported on one end. Its a cheap experiment, and for me...experiments are fun. But...every drive supported on only one side broke the shaft, even with light tire pressure.

When you apply power to the roller onto the tire tread, light pressure will lead to excessive slippage, and heavier pressure will lead to more rapid failure of the shaft. With the exception of the added outboard bearing, changing your design to add support to both ends would not significantly increase the cost, size, or weight of your drive.

I am willing to be proven wrong, and I wish the best for you...

 

[EVTodd]

For newer posters who aren't familiar with the 2010 "year of the friction drive", Todd was the first to make a friction drive using appropriate RC motors. Friction drives have been around for 100 years (using "chainsaw" motors and very large electric motors). The small RC motors are a perfect fit for this application.

In the middle of Todds original thread, Kepler posted he thought the motor-shell itself could be used as the roller, which...if possible, it could make the drive more compact. Kepler and Adrian_sm both developed tiny and light friction-drives that used the motor-shell as the roller.

No drive is possible that could be smaller or lighter. However, from the very beginning, road testing showed that motor-shell drives are most appropriate for mild-assist, which actually fits the demographic of a very light road-bike very well.

For a daily commuter, I agreed with Todd from the beginning that for the very small increase of size and weight, having the roller as a separate component was best. The roller diameter, motor diameter, and motor-kV could each then be chosen based on the best choice, rather than a series of compromises. However, friction-drives with separate roller have been tried that were only supported on one end. Its a cheap experiment, and for me...experiments are fun. But...every drive supported on only one side broke the shaft, even with light tire pressure.

When you apply power to the roller onto the tire tread, light pressure will lead to excessive slippage, and heavier pressure will lead to more rapid failure of the shaft. With the exception of the added outboard bearing, changing your design to add support to both ends would not significantly increase the cost, size, or weight of your drive.

I am willing to be proven wrong, and I wish the best for you...

Year of the friction drive. lol. I almost spit my coffee out when I read that. Kinda true though, I remember seeing the top 5 or so posts on here being friction drive for a while.

But yeah, there's quite a bit of force on the roller, probably more than you think.

Looking at your design I don't think it would be too hard to bring the other arm of the pivot straight down and add a bearing to it. Might actually be easier than doing perfect bends to get that shape on the arm.

 

[bobc]

Interesting "every one sided one broke the shaft".
That's not the failure mode I was considering, (I had been worried about bearing failure) and looking at the motor there is of course a circlip groove (stress raiser) in exactly the wrong place.
I was hoping to do NO work on the motor at all. Replacing the shaft to remove the stress raiser is the same amount of work as putting a longer shaft in (what I'd have to do to install an outer bearing) although in the former case I could simply swap the shaft round & would not need to furnish myself with extra 8mm shafting.
hmmm what to do.
I'll swap the shaft round on the prototype (all the bits are in production already) and see how it goes. If it breaks or bends I will re-engineer with an extra outboard bearing as Todd suggested. The length of the lever arm to the supporting bearing from the tyre contact patch is about an inch, it would take a fair amount of pressure to make an 8mm hardened shaft an inch long yeild. I'm expecting those forces to be less than 300N in reality.
BTW the laser cutters I use have a nifty machine for putting those "perfect bends" in ;^) And it would only cost me about £5 to get an alternative "side plate" made along the lines you suggest - I even have the bearings here already (the motor swings on 'em).
It's just a shame I can't use my lathe for a few weeks (due to building work)!
Thanks guys!

 

[scriberman]

How about fitting a bearing to the inside of the roller, run a copy of the left hand side plate on the right, Use a cap head bolt as a spigot to support the inner race of the fitted bearing?

 

[bobc]

Like it scriberman, it's not as good as a shaft right through, but maybe it doesn't need to be...
Certainly might save messing with the motor shaft - food for thought there

 

[Hillhater]

Interesting "every one sided one broke the shaft".
... it would take a fair amount of pressure to make an 8mm hardened shaft an inch long yeild..

Depending on which motor you use,.. i would not bank on that shaft being very hard ! :wink:

 

[Kepler]

Good to see someone not using the can of the motor for a change. Imo, the diameter of the can is simply too big for a useful assist at lower power levels.

LOL, you keep on telling us that :?. Reality is that 2 years and 1,000's of kms of riding with a 63mm motor driving straight off the can has well and truly proven that useful assist even at low power levels is well and truly achievable.

 

[Kepler]

I thought I should start a thread about this project.
The idea is to make the cheapest and simplest possible legal ebike conversion.
I want to cover the whole thing including electronics.
So here is a picture of the mechanical assembly: it is intended to fit onto the seat tube inside the rear triangle.
The separate drive roller is smaller diameter than the outrunner rotor in order to run efficiently at legal speeds.
The motor is a cheap outrunner from hobbyking or elsewhere. I've used a propdrive 50/60 270kV (from hobbyking $35)
The Controller is a 60A RC ESC with BEC 6s capable. I've bought a "mystery" one (£10.50 off ebay)
Initially I will mount a single 6s 5Ah lipoly battery - this should be good for 7 miles or so on the flat with no pedalling (15.5mph). Cost about £40.

Custom parts are
1) drive roller with taper lock fixing to motor shaft
2) swinging motor mount frame, laser cut stainless steel
3) seatpost clamp parts, fashioned from 10mm Ally sheet
I'll post drawings for the bits as I do 'em, so if it turns out OK anyone can copy it.

Looks like a nice design. I would be a little concerned that you would knock the back of your leg on the motor as it is going to stick out a bit.

Here is an idea, why not just turn the motor around to get it away from your leg and drive it off the can. :lol: Worth a test. If you dont like it, spin it back around and install a roller.

 

[EVTodd]

Good to see someone not using the can of the motor for a change. Imo, the diameter of the can is simply too big for a useful assist at lower power levels.

LOL, you keep on telling us that :? yet 2 year down the track and 1,000's of kms of riding with a 63mm motor driving straight off the can, has well and truly proven that useful assist even at low power levels is well and truly achievable.

Sorry, you're just not going to convince me that having that large of a roller is the best way to go. I suppose saying it's not useful may be a bit of an overstatement but a smaller roller is easier on the esc and motor, especially at low power levels. Running at 6s with a 60 amp esc and that big of roller wouldn't last for long unless you're being very careful with it.

Like I said though. That's my opinion, based on thousands of miles of using my drive just like you. I'm also not trying to sell drives, I'm just throwing ideas out there based on what works for me.

Just out of curiosity Kepler, have you ever tried a smaller roller? When I first started messing around with friction drive I tried larger rollers. It became very apparent, particularly with outrunners, that a smaller diameter helped to keep the motor in it's sweet spot, especially climbing hills and from dead stops. Try it, you'll like it!

Marketing and packaging wise, using the can of the motor is cool as heck, I think it's a great idea myself. It's just too bad there arn't any outrunners with that kind of power in a smaller diameter.

But enough of that, I don't want to keep cluttering up this build thread with an debate that's been going on for 3 or so years.

 

[bobc]

It's interesting & informative clutter to me guys
The can is well out of the way on my road bike but I'd better double check on the folder.
I went the separate roller route because it gave me the freedom to get useable power levels out at reasonable efficiency from the cheapest practical motor, without (I thought - starting to wonder....) having to try too hard.... My design speed of 15.6mph doesn't hit the most efficient operating point of an small outrunner can.
I started looking at a "roller held on both sides" design, based on a bent 3mm stainless laser cut profile (bent into U section). Looks doable and almost as cheap but would mean sourcing an extra long shaft.
mk1 will be single sided but with the circlip groove moved somewhere safer. mk2 will happen if necessary.

 

[Kepler]

Good to see someone not using the can of the motor for a change. Imo, the diameter of the can is simply too big for a useful assist at lower power levels.

LOL, you keep on telling us that :? yet 2 year down the track and 1,000's of kms of riding with a 63mm motor driving straight off the can, has well and truly proven that useful assist even at low power levels is well and truly achievable.

Sorry, you're just not going to convince me that having that large of a roller is the best way to go. I suppose saying it's not useful may be a bit of an overstatement but a smaller roller is easier on the esc and motor, especially at low power levels. Running at 6s with a 60 amp esc and that big of roller wouldn't last for long unless you're being very careful with it.

Like I said though. That's my opinion, based on thousands of miles of using my drive just like you. I'm also not trying to sell drives, I'm just throwing ideas out there based on what works for me.

Just out of curiosity Kepler, have you ever tried a smaller roller? When I first started messing around with friction drive I tried larger rollers. It became very apparent, particularly with outrunners, that a smaller diameter helped to keep the motor in it's sweet spot, especially climbing hills and from dead stops. Try it, you'll like it!

Marketing and packaging wise, using the can of the motor is cool as heck, I think it's a great idea myself. It's just too bad there arn't any outrunners with that kind of power in a smaller diameter.

But enough of that, I don't want to keep cluttering up this build thread with an debate that's been going on for 3 or so years.

I certainly know I am not going to convince you nor am I trying to say a can drive is better than a roller drive. Totally agree that a smaller roller is better off the mark and allows you to spin the motor harder. Both systems work and have their merits and drawbacks. For me the balance between merit and drawback goes in favor of the can drive and for you the opposite. All good.

Yes I have tried a small diameter roller drive. Its performance was as predicted with good low down performance but above 10kph, I could see little advantage. For me below 10kph rarely needs assist anyway. Definite advantages for a small roller with regard to hill climbing but there really are not too many hills I can't tackle driving off the can. 12% is about the max for a can drive which still isn't too bad. The small diameter roller would know doubt do better but low speed climbing up a steep hill is really hard on tires.

 

[EVTodd]

I certainly know I am not going to convince you nor am I trying to say a can drive is better than a roller drive. Totally agree that a smaller roller is better off the mark and allows you to spin the motor harder. Both systems work and have their merits and drawbacks. For me the balance between merit and drawback goes in favor of the can drive and for you the opposite. All good.

Yes I have tried a small diameter roller drive. Its performance was as predicted with good low down performance but above 10kph, I could see little advantage. For me below 10kph rarely needs assist anyway. Definite advantages for a small roller with regard to hill climbing but there really are not too many hills I can't tackle driving off the can. 12% is about the max for a can drive which still isn't too bad. The small diameter roller would know doubt do better but low speed climbing up a steep hill is really hard on tires.

Fair enough. I suppose for MY use I need a smaller roller. I'm not a small guy and I'm not using a super lightweight road or hybrid bike. I don't want to have to think about babying the drive from a dead stop either. I'm not easy on my bike at all.

I'm not 100% sure but I think I'm running more volts on my drive than most others on here too so I think in my case a big roller would pop escs all the time if I went that route.

Actually, that's one of the great things about friction drive. It's really easy to match it to your personal needs.

My drive, however, hasn't been hard on tires at all. I live on top of a river valley. To get to work I have to ride down and along the river. On the way back I have to pick one of the many long, steep hills to get back up and there's no problem. Since 2009 I've gone through one set of tires (front and back). With the amount of miles I put on my bike I don't think that's bad.

 

[bobc]

I'm a bit held up until I get my workshop back together; meanwhile here are some of the laser cut parts

New workshop walls are half built - windows & roof ready & waiting - just need another dry day or two.....

 

[spinningmagnets]

I have recently become fascinated by water-jetting, laser-cutting, CNC milling, and 3D printing. Even though the 3D printing is plastic, you can use the part as a core for a silicone mold, and cast aluminum. Draw a part on a computer screen, paypal someone, and a complex part that started out as just an idea...is sent to your house FedEx/UPS, and very professional-looking parts too.

Looks great! Best of luck.

 

[EVTodd]

Looking good! I can't wait to see it done.

I wish my laser engraver could cut metal but unfortunately it's not quite that powerful :lol: .

 

[whereswally606]

Hey bob, that looks tidy, did you do an extra motor mount for my cadenza? i'm away this weekend but might be available next to drop by.

 

[bobc]

Hi Andy,
I thought it might be a good idea to put a few tens of miles on it on my folder before making a second or third one. Actually another approach would be to put one together with 2 sided roller support & you could be guinea pig for that if you want.... You'll have to let me know what the outside diameter of your seatpost is ;^)
It generally is possible to touch the motor can with your heel, a thin stainless guard of some description might not be a bad idea.
EVtodd - seriously, get in with your nearest laser cutting (small) enterprise. You get proper accurate finished parts for what the material would cost you to DIY. I never ever make anything out of sheet these days. All the stainless bits in that picture cost me £7

 

[spinningmagnets]

For the seat-tube clamp, I used an inside-diameter (ID) of 1-1/8-inch, because that allowed me to use pre-existing handle-bar stems (with adjustable angle), and the difference between the 1-1/8-inch clamp and the common sizes of seat-posts are easily and cheaply fitted by adding a $7 "seat-post shim" (sleeve/tube adapter?)

Even if a friend knew ahead of time which bike he would mount the drive to, this would allow him to swap the drive between several bikes with differing seat-tube diameters (or for future sale?). Just a thought...

Common seat-post diameters http://problemsolversbike.com/products/seatpost_shims/. Since you are making the clamps from scratch, I'd recommend 30.9mm as the ID for the clamp on your drive (I-1/8-inch is 28.6mm)
25.4mm (one inch),
27.2mm,
30.9mm

http://endless-sphere.com/forums/viewtopic.php?f=28&t=21365&start=60#p324661

 

[EVTodd]

EVtodd - seriously, get in with your nearest laser cutting (small) enterprise. You get proper accurate finished parts for what the material would cost you to DIY. I never ever make anything out of sheet these days. All the stainless bits in that picture cost me £7

I'll have to look into that. I have a couple of small cnc machines but they're not powerful enough to cut stainless and it sure would be nice to just send a file over to someone and pick the part up a couple of days later.

How does the laser help with bending the material?

 

[bobc]

The laser place on the industrial estate where I work has a pretty racy computer controlled bending machine, If I can draw it they can usually put bends in for not much money.
SM I was aiming for the seat tube of the frame rather than the seat post itself, & that's 33mm diameter on my folder. I don't know if there's any sort of standard. Yeah, shims will make up for a multitude of sins ;^) Great idea with the handlebar stem clamp! I guess you're using a much bigger motor in that picture which would always catch the heel if it were put in the triangle, so you're mounting above the back wheel.

 

[krik]

Hello folks,
there is mention of speeds of 15 & 20mph for this kind of drive - what happens when you are going down hill 30+mph and the drive is still engaged?
Does something get fried or it just not pull any amps?!