Metal 3D Printed RC Friction Drive - Development / Build Log

jd_3d

1 mW
Joined
Oct 18, 2017
Messages
17
Hi All,

I'm planning to build a friction drive system for my hardtail commuter bike and figured it might be good to document the process here. I am very new to the RC scene and have little experience with electronics, so it will be a learning process for me. However, I have a lot of experience with CAD, FEA, laser cutters, 3d printers, and 3d metal printing so I plan to utilize those in the development process. I plan on using an RC motor and my main goal is to keep the weight down as low as possible, and have around 200-300W of power. Range is not an issue for me, as I really only want 2-3 miles of assistance to get through some hills. I'd also like it to be as quite as possible (will probably go with VESC-X/FOCBOX eventuall for FOC motor control).

I'm in this for the journey as much as the final result, so I'm trying not to get too bogged down on the research and instead learn by doing and learning from my mistakes. That being said, I have read quite a bit here over the past few weeks such as Kepler's great friction drive setup, and others.
 
To gain some confidence on the electronics side of things, I ordered a cheap 2204 brushless motor, lipo battery, 20A esc, and a servo tester. I really suck at soldering but managed to get everything wired up and the servo tester worked great to control the ESC. I won't be using this motor/esc on the bike, but I was amazed at the power of these BLDC motors. Just holding it in my hand and pressing it against the rear tire (with bike upside-down) the rear wheel spun up to 25+mph very quickly at only about 20% throttle. Impressive for the size.

Based on my tests I've ordered some larger parts (63mm motor and 70A ESC, and another 3S battery) so I can get to prototyping. I was going to go for the FOCBOX, but after tax and shipping it was over $180. That was too big a pill to swallow, so for now I'll stick with conventional parts to get everything up and running in prototype form.

e-bike_tester.jpg
 
jd_3d said:
However, I have a lot of experience with CAD, FEA, laser cutters, 3d printers, and 3d metal printing so I plan to utilize those in the development process.

Do you have economic access to a metal 3D printer; or will you be contracting the work out?
 
Buk___ said:
Do you have economic access to a metal 3D printer; or will you be contracting the work out?

I don't have any direct access to metal 3d printers, but I could ask for favors to maybe get it printed for free. However, in the past I've just used print service companies to get parts for personal projects. I plan to do all prototyping on my FDM printers (Type A Machines, Series 1) and possibly some laser-cut parts on my laser cutter (Epilog). If I can keep the weight of the metal parts to under 100 grams, the price shouldn't be too bad relative to the rest of the components for the build. And I think 50 grams may be all that's needed.
 
jd_3d said:
I plan to do all prototyping on my FDM printers (Type A Machines, Series 1) and possibly some laser-cut parts on my laser cutter (Epilog).

Sounds ideal. Prototype cheaply in plastic and only commit to metal once you're reasonably sure of a good design and fit.

I'm envious :) I'm also contemplating an RC motor based friction drive; and having worked my way through the Kepler drive threads and adrian_sm's Commuter booster threads, amongst others, the biggest flaws seem to be down to the limitation of driving directly onto the tyre tread surface.

1) if you run the bare can; you get an almost complete loss of function when the tyre gets wet, even from just surface water.

2) running a tyre covered in road grit directly against the motor can wears the can which is only thin aluminium and also adds side loading onto bearings not designed for it.

3) You can glue sander belt material to the can to improve grip, but at the cost of increased tyre wear.

4) using the start-up torque to lift the motor into engagement is a cute process, but unless you have/find/engineer a 'soft-start' ESC, the engagement will be quite brutal. Even if you limit use to only engaging once the bike is moving, there will be inevitable slippage as the two equalise speed; wearing tyre, can or both.

My thoughts -- and they are nothing more than that at this point -- is to use two smaller motors and engage them onto the rim brake surface (via a rubber tyre added to the can). The idea such as it is, is to mount the motors onto the unused set of v-brake pivots on my front forks -- I have disc brakes -- and use a third brake lever to pull the motors into engagement. This would be done with the motors unpowered so slippage would be minimal and the power would only be applied once they have matched speed with the rim. The action of the v-brake would serve to clamp them onto the rims which should also help to overcome any rain induced slippage.

Once engaged, it'd be nice if the mechanism was such that the motor torque kept them engaged until they slowed down, but I'm not sure if that is realistic. I've also got the notion in my head that the third brake lever would also act as the throttle, perhaps by using a strain gauge to measure the pressure applied. Of course, that might prove to be a pain in the fingers for extended usage; probably needs more thought.

Anyway, just a dump of my thoughts on the idea; I'll bug out of your thread now; but I'll be following along with interest.


Buk.
 
It's funny you should mention driving the rim directly, as I also thought of some alternatives to the standard friction drive. My first thought was to drive the chain/chainring directly but some quick math dictated a gearbox which would add a lot of weight. Then I thought of taking advantage of 3d printing and print out a huge flat ring with a GT2 tooth profile that I would bond to the rim brake surface (thus turning it into a huge gear), then I could use a small GT2 gear on a small motor (since the gear reduction would be huge) and drive the wheel that way. The main problem is how to engage and disengage which adds complexity. Maybe something like a dropper post lever (cable actuated) could be used to engage it physically.

I may still explore those ideas, but I thought it would be good to start simple and do a friction drive. I live in Southern California so rain isn't really an issue and I probably wouldn't ride if it was raining anyway. Since that is the main downside of a friction drive, it may be good for my purpose.

Oh, and my new parts should be arriving today!
 
New parts have arrived!

1) KEDA 56-63 195KV Brushless Outrunner ($40)
2) 70A ESC ($20)
3) Another 2.2Ah 3S Lipo along with a Y-cable to connect both in series ($11 per battery)

I put them on the scale to get an idea of the weight. For the main components I'm under 1KG which is great. Main components are under $100 too which is almost too low. Hopefully I get lucky with the quality of these parts.
 

Attachments

  • IMG_20171031_120500.jpg
    IMG_20171031_120500.jpg
    83.2 KB · Views: 3,671
  • IMG_20171031_120532.jpg
    IMG_20171031_120532.jpg
    60.3 KB · Views: 3,671
  • IMG_20171031_120601.jpg
    IMG_20171031_120601.jpg
    82.9 KB · Views: 3,671
  • IMG_20171031_120649.jpg
    IMG_20171031_120649.jpg
    122.2 KB · Views: 3,671
Some updates on my build:

I soldered some bullet connectors to the motor and ESC, as well as an XT60 connector to the ESC. Tested the motor out on both 3S and 6S, and the motor has some serious power on 6S. I laser-cut a simple bracket so that I could have a very early prototype (I call it Prototype 1). The main goal of this was to test the location and general feel/noise of a friction drive since this is my first time using one.

As you can see in the pics I picked the inside of the rear triangle as the mounting point as I think it has some advantages over the BB location:
1) Going up large curbs or over large objects could cause them to strike the motor in a below BB location. The rear triangle location keeps it safe.
2) I like the aesthetics of the location as it will be within the frame of the bike and hopefully look somewhat stealth in final form. Easier to work on and service too.
3) Because the location is much higher I may not need the complexity of a 'zero gravity spring' as its resting point may be close to vertical, meaning an easy jump onto the tire.

I took Prototype 1 out for a spin and it was sort of a success. The main problem I had is getting enough force on the motor to prevent slipping. I guess zip ties really don't have the stiffness to properly support the motor :D But if I got up to speed first the motor was able to pull me up some moderate hills (4-5% grade).

Can anyone who uses a friction drive setup comment on slippage? In the dry, does the motor slip up hills? How steep of a hill can you climb using motor only? In the dry does the motor slip first or simply stall-out if you start breaking to a stop but keep the power on?

I think at this point I'm ready to start designing the actual 3d brackets/pivots/end-stops. Still haven't decided the best way to do the end stops keeping in mind 3d printing and trying to use minimal hardware.
 

Attachments

  • IMG_20171102_152322.jpg
    IMG_20171102_152322.jpg
    108.2 KB · Views: 3,606
  • IMG_20171102_152348.jpg
    IMG_20171102_152348.jpg
    71.4 KB · Views: 3,606
  • IMG_20171102_154714.jpg
    IMG_20171102_154714.jpg
    74 KB · Views: 3,606
  • IMG_20171102_155038.jpg
    IMG_20171102_155038.jpg
    89.3 KB · Views: 3,606
Spent a few hours designing Prototype 2 which will be printed in plastic via FDM. Here's an image of the design so far, if anyone is interested I can write up some more details and attach some other pics. Any input is welcome.
 

Attachments

  • E-bike_Prototype_1.png
    E-bike_Prototype_1.png
    59.7 KB · Views: 3,557
I spent the weekend updating and refining the 3d printed design. The main issue is packaging as there is so little space where I'm mounting it, yet I need the parts very thick to make them stiff enough in plastic. After printing out 4 different designs I finally have something compact enough to fit and still plenty stiff. I'm trying to go for a minimal look, and also to minimize the hardware/parts. Took it out for a test ride and it performed great. I was hitting over 20mph on flat ground at about 75% throttle. My main complaint is the noise as it is very loud when climbing hills at slower speeds. Not sure how much of a difference a VESC would make, but I'm hoping a lot.

All the 3d printed parts weigh 75 grams, so the total weight with batteries is 1,025 grams. The bike weighs 8kg (17.6lbs), so with everything added it's about 9kg (under 20lbs) total.

I think my next step is to get a throttle/control mounted on the handlebars as its pretty sketchy right now adjusting the power while riding one-handed. Then I need to work on a housing for the ESC to clean up the look at bit.
 

Attachments

  • IMG_20171106_162205.jpg
    IMG_20171106_162205.jpg
    113.4 KB · Views: 3,495
  • IMG_20171106_164924.jpg
    IMG_20171106_164924.jpg
    103.5 KB · Views: 3,495
jd_3d said:
Then I thought of taking advantage of 3d printing and print out a huge flat ring with a GT2 tooth profile that I would bond to the rim brake surface (thus turning it into a huge gear), then I could use a small GT2 gear on a small motor (since the gear reduction would be huge) and drive the wheel that way.

The worst problem with this is the debris/dirt/etc that can get in the teeth, potentially breaking things. If the motor is on a springloaded mount then it wont' break, but it would still "jump" teeth whenever debris gets between the gears.

You can make a housing that encloses the whole thing, with thin sheet metal or plastic, but that adds weight and complexity.

Even riding on the street debris gets thrown up that could end up in the gears; could be worse on unpaved areas.

For engagement I'd use a set of cams on levers, actuated by a cable on an old friction shifter. The whole thing would be mounted on a pair or quad of posts with springs, and the cams would push against the springs to cause engagement. The springs would still allow disengagement even then, so debris couldn't break gears (assuming the debris is smaller than the disengagement space).
 
jd_3d said:
Can anyone who uses a friction drive setup comment on slippage? In the dry, does the motor slip up hills? How steep of a hill can you climb using motor only? In the dry does the motor slip first or simply stall-out if you start breaking to a stop but keep the power on?
My only successful friction drive used automotive radiator fan pancake axial-flux brushed motors, mounted on a big plate that was attached to the frame (with hose clamps IIRC), so motors were always engaged.

I used rollerskate wheels, which were ok but pretty terrible at slipping under non-ideal conditions.

I got some skateboard wheels (stickier) to use but IIRC I broke a motor shaft before I got to try them, and without both motors it wasn't powerful enough to do what I wanted (because I hadn't made any considerations about how fast the motors needed to spin for their best power, etc.; I just used what I could find at the junkyard).

If you're interested in it, you'd have to go to the old http://electricle.blogspot.com site and look at the DayGlo Avenger stuff, because by the time I found ES and started posting here, I'd moved on from the friction drive already to a middrive on CrazyBike2. :)


For the roller cans to not slip, there's a bunch of things people hae used, whihc you've probably read of in the Kepler thread, or in the other friction drive RC motor threads.
 
amberwolf said:
For the roller cans to not slip, there's a bunch of things people hae used, whihc you've probably read of in the Kepler thread, or in the other friction drive RC motor threads.

Yeah, with my latest design I'm getting zero slippage which is great, even up steep hills. I think this rotational design is good because the more resistance you put into the tire the more the motor rides up it and really digs into the tire. Previously I was thinking of gluing a tube around the motor housing, but now I realize its not necessary. The main trick I found is the motor mount/bracket has to be extremely stiff. My latest design bends less than 1mm at approx 100lbs of force which really helps to transfer the load into the tire instead of flexing the brackets. I'm sure in the rain it wouldn't work, but I don't really ride in the rain anyway :)
 
How is that hobbyking esc holding up? I'm planning on building an ebike. At first I thought of a friction drive, but now I'm thinking of a Tom Stanton design. I have a keda 63-64 190kv motor, but i'm looking for an esc ( low cost).
 
vagosofron said:
How is that hobbyking esc holding up?

It's actually holding up great and has made me put off buying a VESC, since the VESC would cost more than my entire build. Someone who weighed about 200 lbs was doing some repeated runs up a steep road on my bike, and afterwards I checked the temp of the motor and it was extremely hot (close to demagnetization temp) but the ESC was fine, so I'm pretty sure my motor would fail first.

I haven't posted an update in a while but since my last post I got a larger battery from Hobbyking (6S 6.6 Ah) which was about triple the energy but only double the weight. I can now do 10 miles easy and haven't yet done a ride where I've used up the battery. I've also done a top speed run and I hit 24 mph on flat ground, or 26 mph if I get into a tuck position.
 
jd_3d said:
vagosofron said:
How is that hobbyking esc holding up?

It's actually holding up great and has made me put off buying a VESC, since the VESC would cost more than my entire build. Someone who weighed about 200 lbs was doing some repeated runs up a steep road on my bike, and afterwards I checked the temp of the motor and it was extremely hot (close to demagnetization temp) but the ESC was fine, so I'm pretty sure my motor would fail first.

I haven't posted an update in a while but since my last post I got a larger battery from Hobbyking (6S 6.6 Ah) which was about triple the energy but only double the weight. I can now do 10 miles easy and haven't yet done a ride where I've used up the battery. I've also done a top speed run and I hit 24 mph on flat ground, or 26 mph if I get into a tuck position.

Did you ever get your design printed in metal or are you still using the plastic setup?
 
jd_3d said:
vagosofron said:
How is that hobbyking esc holding up?

It's actually holding up great and has made me put off buying a VESC, since the VESC would cost more than my entire build. Someone who weighed about 200 lbs was doing some repeated runs up a steep road on my bike, and afterwards I checked the temp of the motor and it was extremely hot (close to demagnetization temp) but the ESC was fine, so I'm pretty sure my motor would fail first.

I haven't posted an update in a while but since my last post I got a larger battery from Hobbyking (6S 6.6 Ah) which was about triple the energy but only double the weight. I can now do 10 miles easy and haven't yet done a ride where I've used up the battery. I've also done a top speed run and I hit 24 mph on flat ground, or 26 mph if I get into a tuck position.

That's great news! I want to use a 10s battery though and finding a cheap esc is hard. I might settle for 8s probably and get a cheaper one from hobbyking, or a used focbox.
 
Buk___ said:
Did you ever get your design printed in metal or are you still using the plastic setup?

I had assumed my plastic design wouldn't hold up to the abuses of daily use, so I had planned on using my prototype until it failed and then doing any re-design needed and then getting it printed in metal. However, the plastic version is holding up great and the weight is very minimal (not sure if going to metal would save any weight really). So for now I'm sticking with things how they are. I think I'd rather spend any additional funds on a VESC first to see if I can get the noise levels down and be able to current limit so I don't have to worry about the motor overheating on steep climbs.
 
jd_3d said:
I had assumed my plastic design wouldn't hold up to the abuses of daily use, so I had planned on using my prototype until it failed and then doing any re-design needed and then getting it printed in metal. However, the plastic version is holding up great and the weight is very minimal (not sure if going to metal would save any weight really). So for now I'm sticking with things how they are. I think I'd rather spend any additional funds on a VESC first to see if I can get the noise levels down and be able to current limit so I don't have to worry about the motor overheating on steep climbs.

Impressive. And I don't blame you. And if once you've sorted the noise level it still holding up, you could just re-print it in black (and spray your can) to make things a little more stealth. Add a small rack and pair of tiny panniers and it could disappear completely.

If I didn't need fenders year around and a rain-proof solution I be tempted to ask you to sell me a set :)
 
I have a road bike with a BB30 Bottom Bracket so Kepler's system would not work for my bike. As well as I can hit a top speed of 21-22 mph with my little Q100c CST rear hub motor,
which is just what his latest friction drive can get up to, so as much as I like Kepler's excellent work, the BB mounting that his design uses just keeps me from getting one.

I would like to lower the weight of my bike , as the hub motor itself weighs about 5 pounds.
And your system Might work if the mount could be made to sit above the seat stays, connected to the seat post. ( Most Road Bikes do not have the space your bike does where you put your motor . )

If you could use 7s or 8s packs and get that speed up at least 25-30 % faster than what you have now ,
and
have a mount that connects to the seat tube/seat post and seat stays , or under the chain stays.
On a Road Bike we can use the space down there, especially when you factor in the large 50-53 tooth outer chainring , that would hit anything way before a little RC motor down there.
If you would design and make a mount for those places, there will be allot of people with performance road bikes that would love to have such a light weight system.


jd_3d said:
I haven't posted an update in a while but since my last post I got a larger battery from Hobbyking (6S 6.6 Ah) which was about triple the energy but only double the weight. I can now do 10 miles easy and haven't yet done a ride where I've used up the battery. I've also done a top speed run and I hit 24 mph on flat ground, or 26 mph if I get into a tuck position.
 
ScooterMan101 said:
And your system Might work if the mount could be made to sit above the seat stays, connected to the seat post. ( Most Road Bikes do not have the space your bike does where you put your motor . )

If you could use 7s or 8s packs and get that speed up at least 25-30 % faster than what you have now ,
and
have a mount that connects to the seat tube/seat post and seat stays , or under the chain stays.
On a Road Bike we can use the space down there, especially when you factor in the large 50-53 tooth outer chainring , that would hit anything way before a little RC motor down there.
If you would design and make a mount for those places, there will be allot of people with performance road bikes that would love to have such a light weight system.

Have you seen this system:

http://www.instructables.com/id/Bicycle-BoosterPack-a-3D-Printed-Portable-Electric/

Sounds like what you are describing. I decided to do my own design as the parts count for that setup seemed overly complicated (over 20 3d printed parts, vs my 3 parts). Also, it requires a large aluminum frame to carry the load, which looks non-standard and not easy to source on your own. But I guess it has more versatility in mounting options. I could probably make a version of mine that mounts to the seat tube, but it would probably be heavier and I'm not sure if I could get it to dig into the tire as well (angles aren't as good).

Mounting it to the chain stays would require a spring to counteract gravity and I think it would 'bounce' a lot as you go over bumps. I really liked the angles on the mounting location I chose as it sits in a neutral hanging position without stressing the brackets, and as you go over bumps/curbs it does not upset the motor. But yeah, it probably won't fit most modern road bikes.

Just out of curiosity, how much would you be willing to pay for say *just* the 3d printed brackets of a seat stay / seatpost mounted setup?
 
Thinking about it more , For a Road Bike connecting to a seat post would be a bad idea, way to much aerodynamic drag.
for a slower and not aerodynamic bike like a city bike or cruiser it might be better, but then the space for a rear rack and/or seat bag would not be possible.

I think the two best options is like what you have for hardtail mountain bikes that have enough space in that area, and under the chain stays , right next to the bottom bracket for Road Bikes .

Under the seat stays and yes you would have to have a spring just like what Kepler's system has, as far as bouncing when operating I have not read anywhere where the motor bounces off the wheel, once engaged and running the motor is forced onto the tire.
The only difference would be the 3-D printed part would have to be angled back towards the BB area , and when the spring engaged to put the motor onto the tire .

For such a mount the customer would have to send you some kind of pattern to make the 3-D plastic mount like with clay or some other lighter weight mold that would be sent to you.
However once you have done one that fits right then others with the same exact bike do not have to send you a mold.

Casual riders could use the same set up you have.

Road Riders would want a faster / stronger motor that can get the bike to speeds that the fast riders go whole using 70% or so of the available power / speed from the motor.

I am only guessing here but I think people would pay around $ 60- $ 80 for a mount .

I am becoming more and more interested in a mid-drive for myself, however if a low cost , very aerodynamic and light weight road version could be made that fits under the BB , and gets up to Pro Peloton Speeds then you might have something really good there.
 
Back
Top