Powerful middrive in a carbon frame (abandoned)

areuz

1 mW
Joined
Jul 1, 2018
Messages
19
I've been researching middrive conversions for quite some time, but still haven't seen many people with middrives in carbon frames.
Is this generally a bad idea, even with a modern trail bike (stump fsr carbon)? Or is it just complicated to pull off and can work well with proper engineering? How would you mount it - 2 piece brackets that go around the frame and tighten or screw through the frame?

This build is giving me some confidence, that with sturdy mounting you can pull it off even jank/ziptie style: https://imgur.com/a/GGiIx#OGRNqfT

Any feedback is welcome, I'm currently deciding between an alloy and carbon stumpjumper, and 2-3kg weight saving on an ebike would be quite appreciated, I just don't want to snap it in half. Attached is my initial idea of how I can distribute enough of the forces and weight to not damage the frame, of course, once I decide on the bike, I'll make a proper 3D model as they have different geometries.

To get the sense of power I'm aiming for.
APS 8072S 50kv - 80mm motor, I will limit the current via the VESC to not overheat it, and implement active cooling if necessary.
ESCape by Stewii from esk8 builders.
12S 10Ah setup - 2x Multistar 10Ah 6S. I will try my best to fit this inside the frame, but I have have to, I'll sacrifice the looks and mount it on the outside.
Currently looking at planetary gearsets to bring the cadence to <80, though I might make a custom chain reduction.
 

Attachments

  • Quick mounting idea.png
    Quick mounting idea.png
    139.1 KB · Views: 3,380
Seems like no one has done something like this before here? That's kinda... scary.
Anyways, I bought the carbon bike, because of the insane (probably computer optimized) geometry of the BB, I'll get it 3D scanned and print a soft plastic\rubber mount with clamp anchor points - later maybe CNC'd with rubber coating. We'll see how it goes - will probably create a thread if all goes atleast kinda well.
 
I think you did already create a thread! Here it is...Your off to a good start. Good luck on your build but it should be fairly easy to buy a few pieces and slap them together if you have 3-d printing skills. I look forward to seeing the first and only carbon mid drive E build ever! Post some pics its what keeps this place alive.
 
areuz said:
How would you mount it - 2 piece brackets that go around the frame and tighten or screw through the frame?

Any feedback is welcome, I'm currently deciding between an alloy and carbon stumpjumper, and 2-3kg weight saving on an ebike would be quite appreciated, I just don't want to snap it in half.

Hi,
I highly doubt there is a 2-3kg difference between a SJ carbon frame and an aluminium frame. Maybe that is the difference between the top spec full carbon complete bike and the entry level model, but that means $5k-$6k differance in cost (new).

To prevent damaging a carbon frame I would avoid clamping anything to the frame.
Check out the Rune build in my signature for a ideas on how to attach to the suspension pivot and bottle cage holes.

Avner.
 
[strike]Nice build ferret, though isn't that motor kinda small for an offroad bike? You have to really sacrifice the maximum speed to get the torque required, or am I overestimating the forces needed to move the bike up a dirt hill?[/strike]
Nevermind on that motor comment, 3215 apparently means 3.2in and not 32mm diameter... ughhh

At first I was sort of confused by your mounting setup, but noticed that you are utilizing the pivot bearings as anchor points. Interesting idea, though I'm not sure the bearings are made for that kind of load. Although, trail bikes can generally take a lot of stress so I guess it should be fine.

Also, your build is quite different from mine when looking at the drivetrain. I want to drive the front spindle and use the bikes own chainrings, etc, to drive the back wheel. So a more classic middrive setup. The only problem is that my crankset is a 2-piece BB30, so I'll have to see how exactly I can mod it to add another freewheel chainring.
 
UPDATEEEE

First of all, I made a giant spreadsheet. Parts, prices, drivetrain calculations, charts to choose the battery, motor stats.
https://docs.google.com/spreadsheets/d/1kn0fpVh8Y_zdPq9DOBgOfY6VqUiCEM-_8SLz82f15qg/edit?usp=sharing

And here we go, progress after 2 weeks:
Got most of the parts delivered, the main parts still left is the ESCape (damnit stewii) and BMS, although the BMS seems like it will arrive in 2-3 days. Also the pulleys... but that's not that important atm.
xh26QMo.png


You can also see the motor and planetary gearbox that I ordered already mashed together with a 3D printed part. The total weight of that unit is 2.8kg and is sturdy as hell - So I'm gonna remove some of the material to reduce weight and bulk, as seen in the 3D model.
sTl1OI5.png

TKhBY16.png

eHWV3Jk.png

i72Erd5.png

NaOlOci.png


General idea of putting the motor and gearbox in parallel - with a tensioning pulley that I made a cutout for in the CAD:
4fDB91H.png


Mounting it on the bike and rebuilding the crankset to accommodate a freewheel is proving to be the most challenging, though I have a guy that could 3D scan the lower bracket area for 75-100€. But I'm also exploring the idea of molding it and then adapting the mold to fit the drivetrain unit.
WuHIxab.png


Oh yeah, did I mention that the 3D printed drivetrain part will be "quickly" detachable with a few screws and a press fit? Yeah... I'm going there.

So far, pretty okay progress but I'll be really happy when I finally get the tools to disassemble the cranks and see If I'll have to completely replace them to fit the freewheel and 3rd chainring.
 
awesome progress! Tiny drive curious how much it will output when geared properly

Nicely done
 
Thanks skeetab! All of my planned gearings and final outputs are in the spreadsheet (in the "drivetrain" tab) linked above.
Of course those are just theoretical, but should give an idea what I'm going for :)
 
More photos will be coming tomorrow, too dark for any nice ones last few days.

So, I worked on the idea of modding the crankset to accept a freewheel with another chainring and I think I got a solution. And before you say anything, I know that I can just buy cranksets made for this purpose, but this is so much more fun, and maybe even cheaper :D

So, the first task was to find a freewheel. There are all the standard ones that SickBikeParts sell, but those require the special cranks that I talked about. I then found out about clutch-bearings. Holy crap how did I not know, they're awesome.
So here we go, I will remove the original 22T mounts that are protruding from the crankarm and remove all of the material that is higher than the spider as seen on the left prototype:
3B5NDri.png

Knm4B6d.png


[strike]This is to make enough room for my own mount. This mount will be a CSK6006 clutch bearing, inner side directly mounted to the spindle (see photo). Outer part of the bearing will be grinded by 1mm on 4 sides to create a shape that can lock in place really well, and inserted into my custom chainring holder.[/strike]

Okay, literally during the writing of this I've realized that I'll try something else first. I realized that even though the screws for the 22T chainring are 11mm deep, the screws actually go in only 5-6mm. So I might be able to use a 62mm wide clutch bearing. I also just wrote the manufacturer if they can't make a version with 2 or 4 keyways which would be much better for the forces on a bike, and make it from the 6006 bearing, which is 55 instead of 62mm wide and 14 instead of 16mm thick. We'll see, this would be absolutely ideal.
BjqSmxv.png


So yeah, the design is not going to change much though either way. But getting rid of the "CNC'ing and grinding a bearing" task would be quite helpful :D

For this bearing, I have made a chainring holder that I will get CNC'd (photos are a 3D printed part). It allows me to mount 3 chainrings onto the outer part of the bearing using a few keyways and (probably) steel keys. The bearing will be mounted onto the spindle (discussed later) using keyways and keys aswell.
wdhQsMi.png

sRtpTEa.png


Here is the fully assembly, going from bottom to top (kinda) is the spindle, custom chainring mount, clutch-bearing and finally on the splines the original crankset spider, modified and not fully modeled cuz I don't need to.
Im8X7c3.png


The mount that I designed has a two-sided spider, the original chainrings mounted on the inner side of the cranks, and 1 outer, that will be for the motor drivetrain. (This is a old model without inner keyways)
1nS2Y4F.png

2YbdVlC.png


So. All of this is relatively simple, all the tolerances are great, and the increased spider arm thickness (by ~3mm) should account for the increased stress from 2 chainrings mounted on it. It also (probably) has a advantage, when using the bigger chainring with the motor, the torque on all the other components will be lower as the chainrings are pretty much mounted rigidly together - at least that's a guy that has never done something like this nor has any skills in proper engineering thinks :).

The final problem now is, that the thickness of the spider just increased by 15mm (13mm). 5.5mm for the new chainring, 7mm for the spider thickness, 2mm for the bearing (maybe not, we'll see), and 0.5 rest for clearance with the original spider. So I'll have to figure out what to do about the spindle. There are a few ideas in the works.

1. Extend the original one. [strike]The basic principle is simple, I even made a model. But the issue is where the extender goes over the spindle splines, the material thickness is as low as .5mm at the edge.[/strike] Yeah.. no, I can't really do that after I changed a few things about the model. And It wasn't a good idea anyway.
2. Remove the spindle and get a new one made. No idea how it is attached nor If I can get someone to remove it, then assemble the new one in place. Also getting it CNC'd will probably be the only way. Maybe in combination with a lathe, but the attachment interfaces on both sides seem very complicated to machine.
3. This is the easiest from the bunch, but probably the most expensive? Getting a 3 piece crankset with a longer spindle. Then getting the spindle modified with keyways (relatively easy task compared to the other ones) and just mount it normally. Although, I will have to see if I can modify the new spider the same way as the one I currently have or use the old one with the new spindle and crank (which might look ugly), it's kind of a gamble really.

The modified spindle should look like this, either way I go.
g6jEnNo.png


Now to more good news, I got the alu pulleys! I'll have to make a 14mm shaft for one of them, so I can't mount it, and made a temporary 3D printed prototype with 8mm bore. Also, I ordered shorter belts as this one is way too long, even with a tensioner it's a little too much, and I will move the gearbox and motor 2-3mm closer together which will just extend the belt.
HB5sBT8.png

JCbxctg.png


So there, this is the progress of 2-3 sleepless nights and full 10 hour days of modeling, researching of prototyping. Why do I enjoy this so much is a mystery I will never solve.
 
After a good nights sleep and some life contemplation I decided to find a 3-piece crankset without a spider so I don't have to modify it nearly at all. Also, after some research about shaft keys.. yeah 1 keyway is enough, so I'll be getting a conventional CSK30PP clutch bearing with 2mm keyways. Though that increased the chainring mount size to 17mm and its (predicted) weight from 130g to 155g before adding the bearing, screws and chainrings.

Oh yeah, here is (hopefully) the last model of the freewheel chainring mount :)
vcYml1j.png

7ast6Qo.png

UHJDlQQ.png


The 22T chainring is not there because I can't find a nice model of it anywhere, so use your imagination and the previous post :)
g0IwTWD.png


So yeah, I just saved myself a headache trying to manufacture a bunch of custom parts. I now just need to CNC the chainring mount and machine a single keyway in the spindle.
 
Alright, seems like everything is turning from "ughhh, what have I got myself into" to "yay, this might actually work better than I thought".

So, I contacted a bunch of aliexpress sellers about the custom bearing, only one replied that they (can) make such a bearing.
e2msrjN.png

https://www.aliexpress.com/item/One-way-bearing-CSK6006-PP-The-overrunning-clutch-two-keyway/32902919828.html

With this bearing option, the OD is smaller than the 22T chainring bolt threads, so I can shorten the whole thing by ~7mm. This gets me back in range of normal spindle lengths - 73mm (BB) + 21mm freewheel chainring = 94mm which is much better than the original of 100mm+. I still have to find a crankset with a ~122mm spindle. White Industries does seem to have something like that (their M30), although local sourcing would be much more comfortable.

I realized that I never actually properly explained why I'm going this route.
There are a few main advantages:
- The freewheel (bearing) is relatively cheap and sturdy compared to other ready made freewheels.
- The custom spider is designed to take regular old chainrings instead of the expensive proprietary ones, that I'd have to source.
- I get to keep the 2x10 setup the bike had, and don't have to deal with alignment because I'm quite literally copying the original spider design and it's position.
- Original bike parts, if, are only slightly modified, so repairing or replacing parts will be relatively easy and quick.
- I get to brag about designing it myself :p

So yeah, that's that, a lot of theoreticals. Currently printing prototypes for a bunch of things, so photos maybe later today or tommorow. I'll still have to wait for the bearing to arrive from godknowswhere, until then I can't really get anything machined, but I'll look for the right spindle.
 
Hi Areus, watching this thread with much interest.
I hope you can sort out the problem of freewheeling chainwheels on a strong crankset, because many systems are still based on square-taper cranksets which I believe are too weak for serious offroad riding.

I like where you are going, but have you checked the torque capability of that sprag clutch? I only looked at one, but it had a torque rating of 110Nm.
If you relate that back to your situation: I don't know how heavy you are, but with full riding pack I weigh around 100kg, which is a "weight" of approx 1,000N. Multiplied by a lever arm (pedal crank) of 175mm (0.175m) this gives a static torque into your sprag clutch of 175Nm. This doesn't allow for dynamic loads caused by bumps or pulling up on the handlebars so you can push down harder on the pedal.
Hopefully the manufacturer has allowed a safety factor, and it is possible that a sprag clutch from one of the world class manufacturers might have a higher torque rating.
If you only weigh 60kg you might be ok, but I think I would overload the clutch and it might slip or be damaged.

The Tongsheng TSDZ2 mid-motor uses a CSK30p sprag clutch as a freewheel driving the chainrings, and several people have reported failure of that clutch.
Dave
 
Hi Drum, glad you are enjoying my build progress :)

And yes, I also share that view, the clutch bearing is the weakest point of my design so far, but I'm hoping, that as the cranks are rigidly mounted to the spindle, it will take most of these downwards forces, and the bearing will only experience the rotational torque. I did order 2 of these bearings, so I will have a spare, and the replacement will be quite easy requiring only a vice.

Also, I don't know about the bearings I got made, but the CSK30PP (6206 bearing) have a torque capacity of 180Nm and a dynamic load rating of ~12kN. Of course, these are numbers from a Chinese manufacturer, and I will be using a slightly smaller bearing so I take everything with reserve.

Source: https://www.aliexpress.com/store/product/CSK30PP-One-Way-Bearing-Clutches-30-62-16mm-1-PC-With-Keyway-CSK6206PP-FreeWheel-Clutch-Bearings/511615_32828413883.html
PrHw9Ww.png


I guess we'll see what breaks first :D There is always something that breaks
 
I'm kinda spamming updates at this point, but for anyone who likes renders, here is the latest motor\gearbox mount, with a belt tensioner made from two 6000-2RS bearings. Will post photos of how these work tomorrow when I print and assemble them :) So far, the tensioner will be just screwed in place, but I might make it spring loaded if it's worth it.

cYeLdG3.png

of3Hqph.png

wcvl4Li.png


Sneak peek of a "few" parts I'm printing. So far a 18 hour print, expect 26+.
jY8AMGR.png
 
I'll do a proper update soon, but here is a preview of what I've been working on :)
2d6UuH1.png
 
querlenker said:
Which gearbox are you using? Thanks

It's a 16:1 planetary gearbox I got from ebay. GAM FP-F-070-016H. It's probably a robotics part, as it's designed with minimum backlash, but datasheet says it's good up to 6000RPM, so my measly 3000 won't even push it.

EDIT: Yeah, also, you can get it from this guy pretty cheap. He said his minimum is 65$ but got him to 60$.
 
massiveb said:
watching this thread with much interest!

Thanks! Been a while but I'm still working on this, just haven't had that much progress since starting college. I'm really close to first tests because I only need to redesign a few parts, print them and finally mount on the bike for the first time! Probably will get to it in the next month or so. But I hope to post progress before that aswell, mainly because I finally received the VESC 6 (ESCape) over from sk8 and can begin testing my drivetrain :).
 
spinningmagnets said:
Well done, very nice

Thanks you :) I'm glad people like my progress even though I don't quite know what I'm doing :DD First time I've done something like this. I'm more of a IT guy.
 
I just today found this thread , because it is not in the Mid-Drive section.

I love what you are doing, I have been waiting for someone to do this for years.
( See the Article from over 10 years ago about the Retired Electrical Engineer in Switzerland, that put a small motor on a carbon hard tail Mountain Bike )

https://www.electricbike.com/lightest-bike/
This article was from 2012, however there was an earlier article on his other bike before the 2012 one.
he only added a couple of pounds, it was very simple, and it has been the the type of DIY design I have been wanting for since before 2012. With a freewheel/clutch though , like what you are doing.

I am very interested in you making a freewheel/clutch to fit onto a off the shelf , newer , lighter weight crankset like BB30/Shimano Hollowtec II
So that we can use off the shelf chainrings.
You probable do not know this yet, so here you are ... when having electric assist/throttle and higher speeds, it is good to use Road Chainrings so that you have better reduction ratio's .
So
Make your spiders with the BCD patterns that Road Chainrings use .
a spyder that accepts a large 53 tooth road chainring to the motor,
and a BCD to the Mountain chain that goes back to the rear cassette.
So what I am saying here is a spider that has both BCD bolt holes on it.
In a couple of versions,
130 Road BCD and which ever mountain BCD you want
110 Road BCD and which ever mountain BCD you want.
BTW
There is now a new 12 speed Eagle NX Drive Train that has a 50 tooth large cog on the cassette, so even less reduction from the motor with that drivetrain. Sram says that the new 2019 , sold now, Eagle NX drivetrain is E-Bike ready , made for the higher stresses of E-Bikes.
And
There are more and more motors that have a KV of less than 100 .


When bolting anything to a carbon frame, have some material between the clamp and frame to absorb some of the compression and vibration.
I think the Carbon frame can handle the vibration better, however with a carbon frame I think that a clamp should be 3 to 5 times or more wide , in order to spread out the clamping force over a larger area.

Where in the World do you live ? just the general area.
Thanks
 
ScooterMan101 said:
I just today found this thread , because it is not in the Mid-Drive section.
Yeah, at first this wasn't a build thread, but just a research question about a few things and I went with it. I might create a new thread when I finish so I can explain everything in one post and people can ask questions and whatever.

ScooterMan101 said:
I love what you are doing, I have been waiting for someone to do this for years.
Thanks! I didn't actually think this was uncommon, but I now see why it is - it's quite a lot of R&D to plan it well enough to get it to work properly. I actually tried doing everything in one summer, but didn't quite have enough time, and a bunch of parts didn't show up at time.

ScooterMan101 said:
You probable do not know this yet, so here you are ... when having electric assist/throttle and higher speeds, it is good to use Road Chainrings so that you have better reduction ratio's .
I actually do, but I planned out everything so that I will have a minimum of ~150/200Nm at all times and a max speed of 60km\h, so a nice balance. And using the original chainrings brings down the cost, and the cost of replacement.

ScooterMan101 said:
When bolting anything to a carbon frame, have some material between the clamp and frame to absorb some of the compression and vibration.
I think the Carbon frame can handle the vibration better, however with a carbon frame I think that a clamp should be 3 to 5 times or more wide , in order to spread out the clamping force over a larger area.
This is something I thought about the most. I found a spot, where I can use an existing bearing mount, and bolt the whole drivetrain to it using a 90mm M5 screw (I've already tested this and it works quite well). To bolt to the frame I'll use flexible 3D printed clamp, that will have a bunch of surface area, and being slightly flexible, absorb vibrations and some stress.

ScooterMan101 said:
Where in the World do you live ? just the general area.
Thanks
Bratislava, capital of Slovakia.
 
Back
Top