An Advanced Friction Drive System
- Thread starter Kepler
- Start date
Kepler
10 MW
Ok, all measurements and geometry have been sorted and I am now busy fabricating parts. I have made every effort to make this drive as small as possible but also very ridged. The mock-up of the new drive on the bike looks ridiculously small and looks about the size of tube replacement kit. Not including the actual motor and pivot, the drive is only 38mm wide will all electronics completely hidden.
mwkeefer
1 MW
Kepler said:
Kepler,
You do very nice work, better than I could with my limited access to metal shop tools of even the most basic nature... Now working with wood, I can do (and I do)... I am now beginning working with carbon fiber at home too.
I do know basic steel, aluminum work, mig, tig, OA, stick and I can work a drill press, recip saw (bench type) and to some extent a lathe (not very good at it) but the only stuff I can do at home is a cruddy hack of a benchtop recip saw made from a portable b&d unit but powered with a supply line... a dremel drillpress and a circular saw which is strong enough to cut 4-10mm aluminum.... that said, it takes forever to make this crap by hand without the right tools so I have the highest admiration for people like you, Timma (what he can do with a drill press), etc...
Now about your newest design, I think it is great and will not only succeed in you goal of reducing the footprint but also I believe it may endup a bit lighter and surely more sturdy... That said, if your open to suggestions and slightly twists on the design - I have one totally out of place suggestion (which you may have posted in another thread) which would require a full redesign (mostly) and then based on your drawing I have some ideas to make it smaller, lighter and stronger (but not neccessarially more rigid in all places).
First the off the cuff bit:
Given the one real drawback of friction drive is the poor performance or just inoperability in rain and other wet weather, wouldn't it be ideal to drive the rim (it would have to be a V machined rim, but they are commonly available as they make a better surface for vbraking) using a pair of spindles which drop down to align with the rim properly when actuated (closed down on the rim)?
I would think a softer polyurathane spindle (or capstan if you prefer) would engage the rim wonderfully.... if you mount the motor transaxialy then you could run small belts or chain and a simple jackshaft to get both rollers engaged but running off one motor and both spinning in the proper way at a syncronized speed? If you use a radius slot type slide / pivot them the belt tension and or chain lengths would remain constant as the rollers actuated and locked into position.
This could also make the unit much more ergonomic...
Don't worry, if you don't give this a shot I will from wood if I must!
Now with the current design, if I am reading the tech drawings correctly...
1.) I notice your use of 6.5mm stock? I am thinking that it may be a bit (just a tad) overkill and if you reduced it to 4mm but the primary bolts (above and to the right and left of the speed controller) should be
mated on the outside edge of the aluminum lower block with a locking nut... this will apply better structural distribution of torque and other forces.
2.) The straps I assume are to limit range of motion of the engagement arm?
That's actually not a bad idea but I would think a guide of flat angle aluminum with slots cut out for the proper angle of attack (a better way to say this?) and simply delrin or ptfe sleeves where the carriage bolts
pass through the guide slot?
3.) In addition to the above, it has occoured to me that if you had detent position switches or at a minimum an end of travel switch which the roller arm would trigger as it both disengaged and engaged in positive
lock.... well if you had electrical signal for this I believe the HV85 (you are moving up right?) has multiple levels of current limiting and throttle ramp control... since you have far more experience with friction
drive, I may be off base (I've built 1 from an old scooter ripped apart for doner parts and I wasn't too impressed with it) but would it not be a GOOD THING in a friction drive to not only ramp the throttle up
slowly until full engagement which then you could go to a second throttle curve with a much faster ramp up to full current... this would allow the drive to engage the full lock position before engaging any tire
wrecking torque (I am thinking of a 750w nominal with 2100w on reserve friction drive which is why I ask) and should in theory increase the efficiency of the system? What are your thoughts no this
4.) Let's keep with the "Spin the drive spindle so bad that you burst the tire" argument or scenario which is most cited as the non weather related drawback to friction drives... First a question... does the positive
overlock type system combined with travel limits prevent sufficient engagement pressure that if the roller was to slip it would not cause any damage because loss of traction will cause the drive to disengage
from the postive lock position (which in theory if you used the above detents, it would reduce the throttle immediatly allowing for a quick recovery of traction - assuming your positive lock handles this "roller
burn out" condition.
If your design could still "burn through a tire" in the event of a massive slippage under WOT style accelleration, I think I have realized a solution (not the simplest but I would be willing to build you a prototype for
testing if you were willing) which is purely electronics based and rather simple (not too expensive either, maybe 15.00 of parts and 15m of time to assemble, 10m install and calibrate).
When using front hub motors, I ran into the slippage issue and loss of traction all the time (even on dry roads, wet for sure and snow... OMG, lets not go there) which in my situation cost me a good bit of tire
tread and lost power... unlike the situation with friction rollers where if they stay locked and keep spinning and speeding up faster and faster... they would in all likelyhood burst a tire right?
So the solution is a 30 line program (written using BasCom AVR) running on an Atmega32 or better (4.00 give or take) and in crudest form... just a hall sensor or equivelent optical tach on the motor output shaft
and in the most refined version... another 20 or so lines of code and a current shunt (I used the 100A solid state hall effect shunt from a burnt out EagleTree logger but a resistor shunt and voltage divider
network would suffice)...
Basically when that freespin occours - the motor RPM increases far faster than it could possibly under real accelleration and the load current drops... the same as with my hub motors... the mCU watches for this
condition and immediatly interrupts the PWM of the throttle causing the drive to stop spinning... a second later (give or take, it's adjustable) it will begin to ramp the PWM back up until it matches what your
throttle input (it works as a wedge) is and then it disengages... This basically detects within 2 - 4 cycles (interrupt based) the slippage and allows the roller to use the kinetic energy in the rolling tire to rematch
speeds before re-engaging the throttle.
Again this may be totally un-needed with your type of setup, I don't know?
5.) I would think you could replace the 2mm aluminum plate with carbon fiber panel and maintain rigidity while saving a good bit of weight and gaining a "cool factor"
6.) If you mount the speed controller inside the (what do you call it) small covered recess at the top, I would suggest obtaining a PC type (maybe graphics or coprocessor) heat sink which matches up nicely with the
splines in the HV heatsink... an alternative ... you could make one with the proper sized aluminum slats attached to the top cover - the advantage would be simple, better heat distribution which will result in
much longer and more efficient controller operation.
7.) Is that motor and pivot part an abstract showing a belt or chain running between the motor and the drive cog/wheel or... is it that you intend to make the carrier plate / mount in a flat panel tear / pear shaped
piece? If it is a solid plate (or plates?) for mounting the motor and actuating the pivot then:
1.) I would suggest patterning after Recumpences mounting beam for the Astro Pro 3210 and the first stage input large pulley.... I can't remember what the part is called but its similar to the rods connecting
an automobiles crankshaft to the pistons - that oblong is only to support a much larger diameter axle on the pulley jackshaft than the motor has and the subsequent bearing.
If you went that route you would reduce the material used by approx 65% which would have weight benefits, then for covers Id go lexan made to look like carbon fiber or actual carbon fiber (or ABS plastic
even) just to hide and protect the drive system.
Can't wait to see the final build.... hope you don't take my input as negative critisim, half of it I am not even sure about it's more questions and observations...
Regards,
Mike
Kepler
10 MW
katou said:
Probably about 4 hours in those drawings. Time well spent as you can get all your dimensions and geometry sorted before you start building which can save you having to make parts twice if you get it wrong. Part of my job as Mechanical Engineer is to draw up various designs for building automation systems so I have a bit of an advantage here
Kepler, I'm going to try to answer question #4 if you don't mind. I'll let you handle the rest.
First off, a traction control system for friction drive sounds very cool, in fact, I thought about using one before I went to a variable pressure system (I call copyright on that name, 8) ). I found that it's just not needed. The sliding action of the mount does all the work, and very well at that.
I have personally never found "burning through a tire" to ever be an issue. In these mounts that pivot or slide the roller gets tighter and tighter as you apply throttle under load. I can pull wheelies from a stand still with my friction drive. I think it would take a really high power system that's very poorly designed to cause that kind of a problem.
I've said it a million times. Most of the rumors about friction drive are just plain wrong except the one about riding in the rain. Hopefully at some point we can talk about friction drive without trying to convince people it works. Btw... It works on gravel, dirty roads, goes over 15 mph, can go up steep hills very well, and it's still very efficient even though it has the word friction in it.
Edit: I just noticed AussieJester dogged friction drive in his thread. No one understands us! lol
First off, a traction control system for friction drive sounds very cool, in fact, I thought about using one before I went to a variable pressure system (I call copyright on that name, 8) ). I found that it's just not needed. The sliding action of the mount does all the work, and very well at that.
I have personally never found "burning through a tire" to ever be an issue. In these mounts that pivot or slide the roller gets tighter and tighter as you apply throttle under load. I can pull wheelies from a stand still with my friction drive. I think it would take a really high power system that's very poorly designed to cause that kind of a problem.
I've said it a million times. Most of the rumors about friction drive are just plain wrong except the one about riding in the rain. Hopefully at some point we can talk about friction drive without trying to convince people it works. Btw... It works on gravel, dirty roads, goes over 15 mph, can go up steep hills very well, and it's still very efficient even though it has the word friction in it.
Edit: I just noticed AussieJester dogged friction drive in his thread. No one understands us!
lol
mwkeefer
1 MW
Todd,
Thanks for answering #4 - I had a feeling that it was a moot point most of the time due to the motion and engagement action of the friction drive swingarm.
With regards to rain, that is my A#1 obstacle to using a friction drive system on my folder... I have to be able to count on it in all weather (even freezing rain and snow) though I would settle for if the friction would handle anything but for snow and freezing rain, and just pop on a front hub motor in those other conditions.
-Mike
Thanks for answering #4 - I had a feeling that it was a moot point most of the time due to the motion and engagement action of the friction drive swingarm.
With regards to rain, that is my A#1 obstacle to using a friction drive system on my folder... I have to be able to count on it in all weather (even freezing rain and snow) though I would settle for if the friction would handle anything but for snow and freezing rain, and just pop on a front hub motor in those other conditions.
-Mike
Kepler
10 MW
Mike,
Thanks very much for you input. I appreciate the interest taken.
Certainly like most friction drives, mine doesn't work in wet weather and a rim drive arrangement would possibly overcome this. That being said, a rim drive is a far more complicated arrangement and to be honest, if going to that sort of complexity, I would go down the path of a Recumbence style drive. I think the whole point of a friction drive is lack of complexity. My bike weighs 11kg with the drive and battery so if it rains, I pedal That being said, it would still be interesting to see the design attempted especially out of wood 
I agree that 6.5mm stock is overkill strength wise but the main reason for using this thickness is so I have sufficient material on each side when boring a 3mm hole through the blocks. The actual top blocks are not meant to provide rigidity at all. They are effectively only a weather cover for the ESC however keeping them 6.5mm makes a good stable surface for the bearing block bolts to tighten against. You may have noted that there are no nuts (except the main mounting bolt). All these plates are drilled and tapped. Another reason why 6.5mm stock has been used.
The sub assembly photos may explain things better.
I might have a look at doing something like mainly to give the design a little more elegance. Functionality wise though, the cable setup is difficult to beat.
You have me a bit confused with how limit switches would work with this type of system. Basically, switch it on and the drive activate moving all the way to the end stop under high torque loads and then disengages when powered off. I'm confused how limits switches could fit in here.
In relation slow ramp, the Castle Creations ESC has very nice slow ramp and allows the drive to smoothly climb the tire to the lock position. That being said, I do need to be moving before engaging the drive. For the low setting, I don’t engage the drive until I am over 15 kph. For the high setting, 30kph. Power wise. my low setting peaks at about 600W and my high setting, around 1200W.
The drive is very positive when locked but has the potential to spin on the tire if I am going too slow when I engage the high speed setting. This something I always need to aware of hence the speed rules I set for myself.
I like it. Could be a very useful addition to the drive and would definitely work very with my 2 speed switch arrangement.
I certainly have thought about making the flat plates out of carbon fiber but in this design, they become heat sinks for ESC. Carbon looks sooo cool but once the side plates are anodized black, they will look quite good too. However, the two plates holding the motor swing arm assembly would work in say 2.5mm carbon.
In this design, the ESC is a very tight fit with no room for anything else. Also air flow means the ESC is open to the weather, something I am trying to lessen. The side plates, I think will make very effective heat sinks for the ESC.
Thanks Mike, I see nothing negative in your post and appreciate your input. (Just maybe a bit shorter next time )
Thanks very much for you input. I appreciate the interest taken.
mwkeefer said:
Certainly like most friction drives, mine doesn't work in wet weather and a rim drive arrangement would possibly overcome this. That being said, a rim drive is a far more complicated arrangement and to be honest, if going to that sort of complexity, I would go down the path of a Recumbence style drive. I think the whole point of a friction drive is lack of complexity. My bike weighs 11kg with the drive and battery so if it rains, I pedal
That being said, it would still be interesting to see the design attempted especially out of wood mwkeefer said:
I agree that 6.5mm stock is overkill strength wise but the main reason for using this thickness is so I have sufficient material on each side when boring a 3mm hole through the blocks. The actual top blocks are not meant to provide rigidity at all. They are effectively only a weather cover for the ESC however keeping them 6.5mm makes a good stable surface for the bearing block bolts to tighten against. You may have noted that there are no nuts (except the main mounting bolt). All these plates are drilled and tapped. Another reason why 6.5mm stock has been used.
The sub assembly photos may explain things better.
mwkeefer said:
I might have a look at doing something like mainly to give the design a little more elegance. Functionality wise though, the cable setup is difficult to beat.
mwkeefer said:
You have me a bit confused with how limit switches would work with this type of system. Basically, switch it on and the drive activate moving all the way to the end stop under high torque loads and then disengages when powered off. I'm confused how limits switches could fit in here.
In relation slow ramp, the Castle Creations ESC has very nice slow ramp and allows the drive to smoothly climb the tire to the lock position. That being said, I do need to be moving before engaging the drive. For the low setting, I don’t engage the drive until I am over 15 kph. For the high setting, 30kph. Power wise. my low setting peaks at about 600W and my high setting, around 1200W.
mwkeefer said:
The drive is very positive when locked but has the potential to spin on the tire if I am going too slow when I engage the high speed setting. This something I always need to aware of hence the speed rules I set for myself.
mwkeefer said:
I like it.
Could be a very useful addition to the drive and would definitely work very with my 2 speed switch arrangement.mwkeefer said:
I certainly have thought about making the flat plates out of carbon fiber but in this design, they become heat sinks for ESC. Carbon looks sooo cool but once the side plates are anodized black, they will look quite good too. However, the two plates holding the motor swing arm assembly would work in say 2.5mm carbon.
mwkeefer said:
In this design, the ESC is a very tight fit with no room for anything else. Also air flow means the ESC is open to the weather, something I am trying to lessen. The side plates, I think will make very effective heat sinks for the ESC.
The pivot supports the motor and the motor casing drives directly against the tire. The photo below should make it a little clearer.mwkeefer said:
mwkeefer said:
Thanks Mike, I see nothing negative in your post and appreciate your input. (Just maybe a bit shorter next time
)
Kepler
10 MW
EVTodd said:Edit: I just noticed AussieJester dogged friction drive in his thread. No one understands us!
At least he rates it above a frock motor :? I think.
Kepler
10 MW
The new drive is coming along very nicely with most of the metal work now complete. Now working on getting all the electronics squeezed into the tiny compartments which is proving to be a challenge but all doable. I am very happy with the rigidity of the unit with it not even showing the slightest amount of flex under load. This will make sure good solid pressure is maintained on the tire when in operation. I am quite surprised how much lighter this unit then the original unit. Total weight excluding the battery is only 1350 grams. Not bad for a drive capable of an output of 2kW. Hopefully will have it ready for testing in the next few days.
Hopefully will have it ready for testing in the next few days.A mech eng whose job includes making drawings. Okay, I don't feel so bad now. I thought I was getting drawing-whipped by some dude who learned sketchup a few months ago!
May I ask how you tapped those holes? I hate using a tap wrench and doing the old 3/4 turn forward, 1/2 turn back, 3/4 turn forward thing. I keep thinking that there must be a way to use the drill press to just power it in, but I haven't found anything like that.
I'm with you, thicker is better, leaves room for oopsies. And drilling a hole into 4mm plate? No thanks. Personally I'm working with 12 mm plate for my design. 2 lbs of extra aluminum I don't think I'll really notice.
Looks great, I'm excited to see it on the bike!
Katou
May I ask how you tapped those holes? I hate using a tap wrench and doing the old 3/4 turn forward, 1/2 turn back, 3/4 turn forward thing. I keep thinking that there must be a way to use the drill press to just power it in, but I haven't found anything like that.
I'm with you, thicker is better, leaves room for oopsies. And drilling a hole into 4mm plate? No thanks. Personally I'm working with 12 mm plate for my design. 2 lbs of extra aluminum I don't think I'll really notice.
Looks great, I'm excited to see it on the bike!
Katou
AussieJester
1 TW
EVTodd said:Edit: I just noticed AussieJester dogged friction drive in his thread. No one understands us!
I didn't say anything against them just said i wouldn't use one, IMO there are far better options. Have nothing against
those that use a friction drive or frock motors for that matter. Eah to their own
That said, i have never seen a friction drive that comes even close to the quality of the one Kepler has produced it is outstanding
design and brilliantly fabricated....
KiM
p.s and yes Kepler i rate them over a frock motor haha...just
keep up the great work :wink:
Kepler
10 MW
katou said:
Tapping long threads into metal blocks are difficult to do in a drill press because the swauf builds up and jams the thread. Very easy to break the tap off especially when only using 3mm taps. The way I do it is I set the block up in a drill press vise nice and square and put the tap in the drill chuck. I then bring the tap down to the job holding light pressure on the drill press handle and hand turn the chuck until the tap has picked up nicely. I then undo the chuck and remove the job with the started tap and proceed with the old 3/4 turn forward, 1/2 turn back, 3/4 turn forward thing. It takes time to make a nice sharp thread but a little patiance goes a long way. Always use plenty of lube also. CRC or WD40 works really well with aluminium parts.
kfong
100 kW
Looking good Kepler.
My tapping method if I have to use the drill press is to spin the drill with the tap. Shut off and swing down before the drill spins to a halt. This usually gets you started correctly. Lock in place, then remove and do the old hand tapping routine. If you have a tapmatic tapper, then the drill press does all the work. You just ease it in and out to remove swarf. You can find them on ebay. They are pricey, but worth it if you have to do a lot of taps.
My tapping method if I have to use the drill press is to spin the drill with the tap. Shut off and swing down before the drill spins to a halt. This usually gets you started correctly. Lock in place, then remove and do the old hand tapping routine. If you have a tapmatic tapper, then the drill press does all the work. You just ease it in and out to remove swarf. You can find them on ebay. They are pricey, but worth it if you have to do a lot of taps.
AussieJester said:I didn't say anything against them just said i wouldn't use one, IMO there are far better options. Have nothing against
those that use a friction drive or frock motors for that matter. Eah to their own
I know, I was just joking around. I'm used to friction drive being the redheaded stepchild of the ebike world. :lol:
AussieJester said:
Now you're just being mean.
Btw Kepler, post your bike on my blog!
Kepler
10 MW
AussieJester said:EVTodd said:Edit: I just noticed AussieJester dogged friction drive in his thread. No one understands us!
I didn't say anything against them just said i wouldn't use one, IMO there are far better options. Have nothing against
those that use a friction drive or frock motors for that matter. Eah to their own
That said, i have never seen a friction drive that comes even close to the quality of the one Kepler has produced it is outstanding
design and brilliantly fabricated....
KiM
p.s and yes Kepler i rate them over a frock motor haha...just
Cheers Kim, much appreciated.
So when are you going to come up for a good name for friction drive? I was thinking a spandex drive or something like that
AussieJester
1 TW
Kepler said:So when are you going to come up for a good name for friction drive? I was thinking a spandex drive or something like that
hehehe @ spandex drive
I am thinking there might be a market for the downhill mountain bikers if the friction drive was good enough to work on the dirt? i saw a post i think by Miles yesterday linking to a quick release ICE motor that fastened on the back of the downhill mountain bike, it is used to get the rider up the hill again for the run down instead of using the chair lifts that are at the downhill courses. Your friction drive and a few lipos in a back pack would be alot lighter and less 'cumbersome' while in the back back when your coming down the hill...Might be a market for your unit in this area. If anyone recalls the link to these ICE motors please post i don't remember the thread it was in.KiM
Kepler
10 MW
Might need a bit of a re design to work on dirt especially with a direct drive motor. I think Todd's design could work in that application though. Must try and find that thread and see what sort of roller they are using.
AussieJester
1 TW
Cheers Miles i thought it was your L337 Autobotness skillzorz at work again hehe..
Kepler, it wouldnt need to be perfect and get super high speed i don't think, if it provided assistance to make the trup back up the hill faster and less strenuous so the downhill run could be tackled in less drained state it would be good enough IMO..
Its light and could easily be configured for quick release setup and stored in back pack better than the ICE version... I should of included Todds drive in there when i commented also sorry Todd yu have also excelled buddy apologies if i offended you
KiM
Kepler, it wouldnt need to be perfect and get super high speed i don't think, if it provided assistance to make the trup back up the hill faster and less strenuous so the downhill run could be tackled in less drained state it would be good enough IMO..
Its light and could easily be configured for quick release setup and stored in back pack better than the ICE version... I should of included Todds drive in there when i commented also sorry Todd yu have also excelled buddy apologies if i offended you
KiM
gtadmin
10 kW
When I was toolmaking (er.. almost 40years ago), we used tapping heads in drills, lathes etc, and kerosene as the lubricant for aluminium. Things must have progressed since then!katou said:
Cheers,
GT
Solcar
10 kW
Kepler,
Great work. Now we can use a cheap fuel cell that runs off of a safe type of hydrogen compound as a lightweight power source. It'd be nice if it could be regenerated with electricity. That would give us lightweight drive and power supply.
Great work. Now we can use a cheap fuel cell that runs off of a safe type of hydrogen compound as a lightweight power source. It'd be nice if it could be regenerated with electricity. That would give us lightweight drive and power supply.
Solcar
10 kW
Miles, that's what I mean. A clean, light weight power source would be great. :wink: I was hoping for the Millennium Cell, but it apparently turned out to be vaporware. It would have gotten closer to that goal. http://www.campist.com/archives/millennium-cell-and-horizon-fuel-cell-hydropak-pow.html
