Observed trials bike designing

[bikerpete]

If folks are curious about the techniques behind the slow / trials sort of riding, this video might be interesting:
https://www.youtube.com/watch?v=7oDyewNtTF0

That's very much the minimum skill level clutch intro. Even in the clutch pop he still talks about using throttle to lift with - that's well and good if you're aim is just to get the front up and roll along, but it's pretty limited.

If enduro is your thing I quite like this Jarvis video with his clutch finger in the corner of the shot. It really shows just how much it's used and just how quick you can go from power-on to power-off and back again, and all the nuance in between - trying to do that on throttle alone would give you RSI pretty quick as you snapped your wrist back and forth like a lunatic, besides being nigh on impossible as you move around over the bike. Notice how the clutch hand is rock steady.
https://yewtu.be/watch?v=En4pLTEUY70

There's a lot of truth to the common refrain in electric drivetrains that you can just put a bigger motor in there ....

No, there's not!
A bigger motor does nothing for accurate, fine control - I'm surprised to hear you say it.
Why oh why do people have this fixation on bigger is better?

 

[DanGT86]

No, there's not!
A bigger motor does nothing for accurate, fine control - I'm surprised to hear you say it.
Why oh why do people have this fixation on bigger is better?

The fixation comes from the math. Many builders have tested this and determined that the weight and drag from gear reductions and multispeed gearboxes ends up being worse than a larger motor. If you are carrying a specific mass with you anyway you might as well have it working for you instead of against you by weighing you down. Companies making roadgoing vehicles and even MX dirtbikes have determined the electric motors have enough torque and rpm to cover the whole range. I get that trials and hard enduro type riding are way different because clutch control and extremely low wheel speed are required.

Electric motors make way more torque per pound than their gasoline counterparts. A gas motor would feel like garbage making no torque and trying to modulate with the clutch in 5th gear. Without the gas motor's picky torque curve and narrow powerband an electric motor should in theory do just fine with it.

Usually the drawback is that there are not an infinite number of different size motors on the market. So if your needs fall right in between two vastly different motors then perhaps gear reductions come out slightly better in your particular unique application. If a manufacturer can only spec a bike with low power to keep the price reasonable then they can extend the capabilities of the bike with a multispeed gearbox. But that only masks the inadequacy of the system. An E motor with 5x the torque of a gas motor probably doesn't need the same flywheel mass.

Your last response to me did make me think about the large motor and clutch situation a little more. Clutches have a particular torque holding capacity. Like you were saying a large clutch capable of handling huge amounts of torque from a larger electric motor has other drawbacks such as pull force and control. So unless we have the R&D budget to design clutches from scratch the design parameters are narrow.

So I think the optimal trials setup might be a single speed bike with a high rpm electric motor. Likely a single speed gear reduction from motor to output sprocket and a clutch placed wherever the torque is most similar to that of the gas motor the clutch was designed for. An appropriately weighted flywheel needs to be in there too. The rotating mass, torque, and rpm of the electric motor would definitely factor into this flywheel mass. I believe having a high rpm motor would smooth the transitions out a little for fine clutch control compared to a giant stump pulling motor at low rpm.

So I kinda agree with you that bigger motor is not necessarily always better but I definitely am not convinced that a multispeed gearbox is needed or a worthwhile weight penalty.

If it turns out that the power, torque, mass ratio of electric motors is just too far away from the gas motors we are used to then complex electronics are going to need to "fool" us into the feel we want. Something like having a controller purposely derating the torque output for clutch dumps to get the torque in line with what you are used to on a gas clutch dump. This would do what you are expecting without destroying your clutch but still be able to perform when you are hill climbing with the clutch lever fully released.

I would absolutely love to be involved in prototyping a setup like this. I don't have the money to complete a full bike build at the moment but I do have a lot of manufacturing and machining experience. If anyone around missouri wants to build a custom gear reduction and clutch setup for a dirt bike I would be interested in helping out. If I were to build a gear reduction and clutch housing with a motorcycle output shaft that easily attaches to electric motor choices on the market what would be the favorites? I'm not trying to make a profit but I don't want to spend a bunch of money for something I don't need. Would anyone be interested in buying something like this?

Just realized when editing some typos here that I am kinda rambling again. Not trying to be an argumentative turd. I just really enjoy thinking and theorizing about this particular topic so I'm probably saying a lot of things that have already been covered in this thread.

 

[bikerpete]

Just realized when editing some typos herethat I am kinda rambling again. Not trying to be an argumentative turd. I just really enjoy thinking and theorizing about this particular topic ...

:lol:
Nothing like a good ramble to get the ideas going.
Certainly I get that you're not being an argumentative ba####ed for the sake of it.
All good.

I don't quite share you're conviction that the lack of gearboxes in production machines is purely due to the fact they aren't required. 40 years as an industrial designer tells me that those sorts of decisions are cost based as much or more than performance based.
The cost-benefit ratio is reduced in an e machine for the majority of users, so in the end the finances dictate the marginal decision.

I totally agree with "If you are carrying a specific mass with you anyway you might as well have it working for you instead of against you by weighing you down."
But my conclusion is "Therefore use a smaller motor and put in a 2-speed gearbox which adds quality rather than simply quantity".

The enemy of clutches is heat, so reducing rpm before the clutch can be beneficial. Yes, speed reduction is a trade off between rpm & torque, you still transfer the same power, but real-world inaccuracies in clutch control, heat dissipation, oil movement etc probably favour a slower slip ratio.

An E motor with 5x the torque of a gas motor probably doesn't need the same flywheel mass.

No. It possibly needs a bigger one.
Remember "an accelerating motor is totally different from a decelerating flywheel", or whatever the quote was.

One of the worst characteristics of e bikes I've ridden, from my narrow perspective, is that when you close the throttle its like throwing it in neutral. No inertia carrying you on, no inertia resisting further acceleration - it's horrible really.
Closing the throttle pretty much turns off all power control and throws the outcome to the wind.
If you've setup perfectly it's fine. If you've been a little off then too bad, there's no nice gradual transition during which you can respond.
A fast gearing just makes it worse as what little inertia the motor etc have is more easily overcome by the back wheel. A low gearing amplifies the wheel speed changes & makes better use of the available motor/flywheel inertia to smooth things out.

I wonder if you aren't onto something with the small, higher rpm motor with a decent flywheel.
Most trials bikes have headed toward bigger diameter " stump pullers" because that's the sort of power you want at the back wheel. Probably also that it simplifies the reduction drive too.
But what you're saying makes sense - stores more energy in a lighter flywheel and tones down the light switch response a bit.
That would be really interesting to experiment with.

It's actually a bit what the Dragonfly has done, & for which I hear it being criticised for, but maybe it will prove to be a very sound decision.

I wish I was in Missouri & we could collaborate.

 

[bikerpete]

I've been about this idea of a small diameter high RPM motor.

The logic behind a large high torque motor seems to be along the lines of, "we need high torque, low RPM at the wheel, so use a high torque, low RPM motor".
It makes sense on a simplistic level, and it also makes it possible to use very simple & cheap reduction drives.
But that thinking is far too simple and totally ignores the use of clutch & flywheel. It is 'baggage' from the first e-trials designs where there was no clutch and no flywheel and reduction drives were either single-stage or used standard off-the-shelf industrial components, eg HTD belt or chain.
Torque was all you had.

But know I'm thinking, how about a smaller diameter higher kV motor with a flywheel that wrapped around the outside of the motor.
Now you've packaged up a more optimised flywheel design & motor into a nice compact unit.
As long as the motor has enough torque to drive the bike up whatever you need, that's enough, don't give it more. More is definitely not always better!

Give it a clutch and 2-speed 'box and you've now got something pretty interesting.
Nice smooth control, fast torque control via clutch, overall power creation via a less responsive throttle.

I don't remember if this has been said before, but it's important to recognise that a clutch is solely a direct torque control device.
Unlike a throttle it does not alter the power source RPM, it directly changes torque delivered.
That's why a clutch cannot be substituted with a throttle.
Combine that with a flywheel that decelerates as power output increases and you've got a very unique control system.

Imagine a controller that reduced RPM as it increased torque - that is so far from what any controller I know about does, but really that's what it boils down to I think.

I'm really intrigued by this small diameter motor idea.

 

[ecstatic_subjectivity]

DanGT86, a modern multiplate clutch in a supersport will easily take ~200+hp for tens of thousands of miles, including aggressive launches and the like - wet bath multiplate clutches on modern motorcycles are pretty absurd in terms of the abuse they'll take. MotoGP bikes only have a single launch, but are easily taking over 300hp out of a dry clutch assembly. So there's no need to build a custom clutch assembly. The clutch on the EM is tiny relative to what a gas dirtbike runs, and the overall output of the QS138 motor is about 50% of the output of a 300 2t motor, even if you burst to ~25kw.

My long term plan, if the concept plays out nicely, is going to be to build a gearbox based around a KTM 6 speed assembly. The first prototype will just have the electric motor drive the crank, to be as close to the 2t design as possible and to see how that plays out. The second version, I'll get a set of cases and remove the crank assembly and figure out how to drive the clutch basket directly, with a flywheel weight on the output shaft of the motor (or potentially on the clutch basket itself).

Third version, once the setup is proven, I'll explore custom designed assemblies. Until then, too much money for an unknown outcome - it's very easy to find a KTM woods bike and weld a sprocket to the end of the crank. If you wanted a very simple thing to prototype with, you could buy a motorcycle clutch, primary drive gear, attach a flywheel weight to the end of the motor, and attach the clutch basket to a jackshaft, and have a very straightforward drive system with a clutch. You'd probably want to put it in an enclosure, as they like to be bathed in oil.

Here's a good video of a 2 stroke bottom end disassembly, if you're curious about the design on modern motorcycles:
https://www.youtube.com/watch?v=gp8CHmG6nw8

That's very much the minimum skill level clutch intro. Even in the clutch pop he still talks about using throttle to lift with - that's well and good if you're aim is just to get the front up and roll along, but it's pretty limited.

If enduro is your thing I quite like this Jarvis video with his clutch finger in the corner of the shot. It really shows just how much it's used and just how quick you can go from power-on to power-off and back again, and all the nuance in between - trying to do that on throttle alone would give you RSI pretty quick as you snapped your wrist back and forth like a lunatic, besides being nigh on impossible as you move around over the bike. Notice how the clutch hand is rock steady.
https://yewtu.be/watch?v=En4pLTEUY70

Yeah, Jarvis is great - I figured I'd give the simpler intro for folks who didn't have experience with the style of riding, though

No, there's not!
A bigger motor does nothing for accurate, fine control - I'm surprised to hear you say it.
Why oh why do people have this fixation on bigger is better?

I think if you're comfortable scope constraining the usecase, it's a 6 of one / half a dozen of another argument. If you assume no one needs to go over 90mph, moderate (for a motorcycle) acceleration is fine, you can avoid a large number of complex systems, by running it as a single speed. If you accept certain scope constraints, the lack of a transmission makes a lot of sense. Simpler, less drag, less stuff to maintain, cheaper to assemble.

The Brammo, IMO, has way too many gears, and it felt pointless to have to shift so much. You could easily ride it around in 4th all the time, and it was fine, although it was a lot more fun around town in first or second, just belligerently fast off the line - if I was on my KTM SuperDuke 1290, I had to basically do race starts to keep with it up until about 30mph (at which point, he'd be shifting to second, and I'd finally have the clutch out and blow by him). For your average rider, I think a single speed is fine, and the simplicity and low cost of maintenance is appealing (no oil changes!). When you start to get in to the more specialized use cases, I think everything changes, and the flexibility of the gearbox allows the bike to excel in significantly more situations than you would with a single speed alone. I definitely got lazy and just rode the Brammo around in 4th gear sometimes, it wasn't inspiring but it also wasn't designed for that.

If your curious, here's my hot lap at Refuel, back in 2013:
https://www.youtube.com/watch?v=_H6FRplenUg

You'll note I never use any gear below 3rd, and the bike can't really pull much over 100mph even with a fairing on it so first and second are just along for the ride. The zeros, with the single speed setups, were quite fast, although they had overheating issues given the amount of time you spend at full throttle around a track. They also saved time by not having to shift. If I had the choice of a 3 speed (3rd, something between 4th and 5th, and 6th) and 10-20% more power, I'd absolutely have taken that, for this specific use case.

Been a long day, so apologies for my rambling as well Nice to participate in this conversation! Can't wait for my parts to show up!

 

[DanGT86]

That's a great video to really get a detailed look at things.

I have seen people on here machine the crank portion of the motorcycle case off and put an electric motor in that void. The reason I was thinking custom is that half of the case volume seems to be taken up by transmission that I didn't plan on needing. I'd like to explore a very small housing that just has the clutch on it. Although by the time one factors in the mounting points to the frame and the benefit of having the output sprocket exactly where the bike engineers put it it might just be easier to use the factory case even if the transmission gears are all removed. Saves a lot of engineering and is extremely light weight.

My talk of custom clutches was from the standpoint of wanting to keep the exact feel of a trials bike clutch with an electric motor that was throwing way more torque at it than it was designed for. I was just trying to say that it would be nice if the gear reduction between the motor and clutch basket was a ratio that got the torque closer to its design parameters. I suppose if you do that then its back to needing gears.

I'm looking forward to seeing your builds with these gearboxes.

Nice that you have some seat time racing these electric motorcycles to have a real world feel for this stuff.

 

[speedmd]

As much as I like the more direct route of using existing clutch designs, I am holding out some hope of someone testing out a newer approach to it in some sort of eddy current setup.

 

[DanGT86]

Thats a cool thought but wouldnt the conductors have to be lighter than a comparable mechanical clutch for it to be worth it?

 

[bikerpete]

As much as I like the more direct route of using existing clutch designs, I am holding out some hope of someone testing out a newer approach to it in some sort of eddy current setup.

I've looked at using eddy current clutches in other industrial machines.
That are really heavy and don't much like spending a lot of time in the transition (slip) zone. They build up heat which is hard to dissipate.

I dare say that's partly due to their current popular use case & the resultant design parameters, but I suspect some of it is inherent.

Also springs consume zero energy. In a battery powered machine, dumping power into a clutch when springs can do the same job seems completely counterproductive.

I liked the idea but certainly moved on from them.

DanGT66 - an electric motor doesn't necessarily produce more torque than the ice bikes, the clutch design is totally a non-issue. You just alter the number of plates, diameter and springing until it suits. There's innumerable options out there that are cheap, reliable and more than capable of the puny outputs we're talking.

 

[speedmd]

Eddy's are used to stop high speed trains so heat is most likely more a design issue. Same with its efficiency compared to mechanical closures. Lends itself to easy flywheel mods like plate stacking for much flywheel mass adjustment potential. Again, I am just looking with interest for some deeper diving into the potential.

 

[bikerpete]

Eddy's are used to stop high speed trains so heat is most likely more a design issue. Same with its efficiency compared to mechanical closures. Lends itself to easy flywheel mods like plate stacking for much flywheel mass adjustment potential. Again, I am just looking with interest for some deeper diving into the potential.

I remembered that the clutches I'd previously looked at were powder clutches, not eddy clutches.

I know I'm full of projects just trying to cobble together readily available parts, let alone develop something totally new.

 

[speedmd]

No harm. For quick-direct steady improvement, I agree, its best to do what you can do, today. Great to look at what can be done years down the road also and incorporate tech as it becomes avail. Always interested in what makes sense in performance improvements and the dreamers input needs to be somewhat limited to not get in the way of the next rev.

 

[bikerpete]

Recently I made some significant changes to my e-trials bike with the QS138-70 and Fardriver that have been really successful and illuminating.

I reconfigured the HTD belt primary drive to a 1/8" bicycle chain so I could squeeze a big (160mm dia x 20mm thick) flywheel in next to the motor. I also modified the dry 4-plate clutch to a wet plate.

I'm running both the original flywheel and the new flywheel, so 3x the energy storage now.
What a difference!
The acceleration is smoother and slower so you've got a little time to react with the clutch as the motor builds RPM. ie I can snap the throttle wide open and pick the moment to dump the clutch as the RPM builds to the point I want.
The bike no longer acts as if it's on a switch when you cut the throttle, the flywheel really carries the bike nicely letting you adjust the instant power with the clutch.
Acceleration off the clutch is super sharp at last. None of this terrible ramping up of power if using throttle, and a really good reaction from all that energy in the flywheel. Way, way faster response when you want it.

One of the things that this change really drove home is the way a decent flywheel lets you make gross power adjustments with the throttle but really fast, fine adjustments with the clutch. It is just such a huge difference riding the same bike with just a change to the flywheel.

The other thing that it really made clear is the stuff DanGT86 was talking about in a previous post here (https://endless-sphere.com/forums/viewtopic.php?f=10&t=104472&start=150#p1713286). This was about the way ICE make power in response to load etc.
Changing to a big flywheel introduces some of the effects Dan was talking about with ICE. The big one is that the power delivered automatically responds to load. This is exactly what the flywheel does over short durations. If there's no load it just sits there gradually spinning down. But apply a load and there is instant power delivered totally automatically, like magic.
That is really a game changer.
The other aspect of this that was made really apparent was that there are plenty of times when you want more power, but you don't want acceleration. That just doesn't happen if you've got bucket loads of power in an electric motor - twist the throttle and there's acceleration. But how about when you've just punched the bike more or less vertical so the back wheel hits a face and climbs, you realise that you've undercooked it and you need to give it a bit of power to get to the top. On an ICE trials bike with their relatively low torque and big flywheels and the nature of ICE that Dan was talking about, when you give it a handful of throttle the bike doesn't accelerate, it just doesn't slow down. The power comes on strong but without any excess acceleration. More like a big truck engine - they deliver huge torque, but they don't set any records for acceleration.
This behaviour is critical - the last thing you want when the front wheel's pointing to the sky and and you're high off the ground on a big rock is for the bike to want to accelerate. That's going to end really badly as it loops out over the top of you.

The oil bath clutch was also a big winner - immediately smoother, more predictable and all round nicer. But then I needed to do some serious hackery to fit a 5th plate in the space designed for 4 because it slipped and wasn't fast enough to grab with all that flywheel mass. It worked though. Hooray.

 

[DanGT86]

This is really cool that you have a real world test mule to try all of this on. Do you have a build thread with pics of this stuff?

How do you like the fardriver? How tunable is it?

An obstacle for me to starting another build has been the lack of user friendly (in my opinion) controllers. I am really happy with my Nuc 6FET but the Nuc controllers are too hard to get right now to base a build around one. Seeing other people suffering through poorly translated unsupported software to tune controller settings makes me reluctant to do a full size motorcycle build.

Glad to hear you are making progress with your trials bike research.

 

[speedmd]

Great developments Bikerpete.

Potentially having the added flywheel mass incorporated in the motor side of the wet clutch has me thinking.

https://emrax.com/wp-content/uploads/2022/11/EMRAX_188_datasheet_A00.pdf
The 188 (sub motor of choice) with a oil tight clutch housing cover directly mounted, with the clutch's mass and oil simultaneously robbing heat from the motor would be trick.

 

[bikerpete]

This is really cool that you have a real world test mule to try all of this on. Do you have a build thread with pics of this stuff?

How do you like the fardriver? How tunable is it?

An obstacle for me to starting another build has been the lack of user friendly (in my opinion) controllers. I am really happy with my Nuc 6FET but the Nuc controllers are too hard to get right now to base a build around one. Seeing other people suffering through poorly translated unsupported software to tune controller settings makes me reluctant to do a full size motorcycle build.

Glad to hear you are making progress with your trials bike research.

I'm doing this on a dob ""Bad Sheep", which I import and distribute.

Unfortunately I don't have a the Bluetooth dongle for the Fardriver yet (generic ones don't work) so I haven't been able to get into the Fardriver. From what I've gleaned it's not too bad to adjust, but has some quirks. I really wish it had a speed&torque mode - I ran my Nuc 12F in the bike for a while and found that mode pretty nice.
It seems that the Fardriver only has an On/Off regen function that gives some people a bit of grief to get working which is a shame - I'd like to be able to have variable regen, not a big deal though.
It handles tickover or idle well, just sits there at whatever RPM is dialled in, although I'm not clear if that's standard or a custom feature - the model number on the controller doesn't seem to be one of the standard ones I see elsewhere.
There is one slightly annoying "feature" - when the motor is stalled the driver shuts down the output. No doubt a protection feature but sometimes when you jump to an obstacle the sudden impact stalls the motor for an instant - then it's freewheel. Not very pleasant. The big flywheel has improved this considerably as it's a lot harder to suddenly stall.

Great developments Bikerpete.

Potentially having the added flywheel mass incorporated in the motor side of the wet clutch has me thinking.

https://emrax.com/wp-content/uploads/2022/11/EMRAX_188_datasheet_A00.pdf
The 188 (sub motor of choice) with a oil tight clutch housing cover directly mounted, with the clutch's mass and oil simultaneously robbing heat from the motor would be trick.

The Emrax looks a beast of a motor, I imagine your choice of controllers is somewhat limited if you were to try to use all it's power?

I'm not too convinced about a clutch direct mount to the motor. As slip speed increases heat increases, you may well end up just pumping heat into the motor, not out of it. My prototype wet clutch is quite limited in the volume of oil I can fit in there, but I've had to stop a few times to wait for the smoke to stop coming out! In trials applications the motor spends an insignificant amount of time producing high power - take a look at the size of a 300cc trials radiator - they're tiny and work fine even in 40C+ heat, so motor heat is largely a non-issue. I've never manged to get my QS to more than just detectably warmer than ambient riding trials.
I also wonder if having a high speed differential between plates will make it slower to lock up.
It's certainly something that would be interesting to experiment with and find out exactly how it all acts.

Talking about motors ...
I've had a discussion elsewhere with someone who is convinced that a trials bike needs a large diameter motor in order to produce high torque at low RPM. That seems to be the conventional wisdom when you look at most e-trials bikes that have been made.

I've begun wondering if that's the wrong way of looking at it, mostly due to the importance of flywheel inertia. I note that none of the e-trials bikes I'm aware of have large, high energy flywheels, they are all pretty small compared to the the total flywheel energy in an ICE bike (flywheel and crankshaft plates combined mass). Even including the rotor mass I suspect they are all well down on energy storage.
My current thinking is that maybe it makes sense to use a much higher revving motor with a decent flywheel on it and a bigger primary reduction. This would give much more efficient energy storage into the flywheel, which could then be lighter.
The QS motor at 48V with field weakening on the Fardriver barely makes 5,000rpm, so the flywheel needs to be relatively large & heavy compared to say my 300cc that revs to around 10K rpm.

I've been imagining a relatively small diameter motor with a cup shaped flywheel that wraps around the outside of the motor at one end (to keep the whole package nice and short..

The motor does need to provide some solid power for driving up steep, grippy obstacles and accelerating into an obstacle on a steep hill, but I'm thinking it might be more versatile to have higher rpm spinning the flywheel and rely on a bigger reduction ratio to get the torque to the back wheel.
With a decent clutch/flywheel a slight lack of low down torque can be compensated for by slipping the clutch, although ideally that wouldn't be needed.

Any suggestions for a motor that might spin to say 10k rpm on 48V and that can put out around 7-8kW peak? Preferably light and cheap :lol:

 

[speedmd]

Running Sensorless may be possible with the addition of a clutch. This would reduce the erpm limits to some degree. This would open up the larger rc motors to the app also. Out runners would lend themselves perfectly to a drum flywheel that could incorporate a eddy clutch.

Dreaming again, sorry. This one caught my eye.

https://www.ebay.com/itm/255829972392

or this one
https://www.ebay.com/itm/255829966778?_trkparms=amclksrc%3DITM%26aid%3D1110006%26algo%3DHOMESPLICE.SIM%26ao%3D1%26asc%3D243355%26meid%3D79367ce4a5f340bebc755f484c65f0a1%26pid%3D101113%26rk%3D1%26rkt%3D12%26sd%3D255829972392%26itm%3D255829966778%26pmt%3D0%26noa%3D1%26pg%3D2563228%26algv%3DDefaultOrganicWebWithV5RefreshRanker&_trksid=p2563228.c101113.m2109

or this one
https://www.ebay.com/itm/174284612938?hash=item28942c254a:g:cCoAAOSw1gxev4KR&amdata=enc%3AAQAHAAAA4H8ibcUjv71DQlYoIpZZWVGtga9XiZ2TnLQ%2F7dq%2BWof9Yyhit7nRy6yrEgyL%2BSuWJIYsoBgULAq8rm%2BbwE7nCxwZ%2B4Zd0hY8Tp0BH2wVwFnkFTpdCbXFT2UcVoAP%2FiQ%2BPaJcfYMljJLBBE30%2BHOVkYULlCy07YwSTCz%2F805NS1HhX1pNeLBHSLjqRJIwizitEBhFjQUQyOs9RQMOp8HQVWbGTZQgvcg%2F0BAYnyANIL42jxthtM66WxT8jcAWg%2Be18VQ3y8PkTcPOuIF4Ewb8H1aKcWlHoU7w%2FH1PxQRpSn1N%7Ctkp%3ABk9SR9bqnoySYQ

or this one https://www.ebay.com/itm/163413707674?hash=item260c374b9a:g:zVoAAOSwbG5idhkb&amdata=enc%3AAQAHAAAAsA5WhpnA4AQl%2F8XPpvGWQTGhQFKvS%2Fm01ZAHIN9kxc3ZAakXw09iQ6Wdp59wJjmIstcq4OgU%2F4Pj%2BMHYe2rRxLSLZe%2F26WoGLh%2Bgl2awwwhflqRQ7Rav698HFeaIjxc3302U4%2F9eWABtCieT%2FLUkTd4spwIleS2VfgCLF75hy34hlfG7xoR7ml7%2FN9PKT4yqJryNwzQpyAyg%2BIbHPwPRSI1UsHwsVqx0oJG7GVHmy%2F9u%7Ctkp%3ABk9SR8DynI2SYQ

potentially for super light ebike https://www.ebay.com/itm/164126480339

Beauty in combining the heat soak of the clutch and motor is you have only one cooling loop (as large as needed) to handle the heat. The spinning lends itself to passive pumping to a external pipe loop style cooler with excellent heat path conductivity of the oil. If you can get a dry clutch setup to work, air (fan) from the motor side with properly places fins could work to cool both sides.

Progression of power and torque and flywheel optimization is a interesting study. The larger diameter rotors would allow more broad relative torque band. You can have and be able to use high levels of torque at near zero rpms with them eliminating the need for super high rpms. 5 rpm to 5000 rpms is 1000x; while 500 to 15000 rpms is only 30x. Lots to ponder.

 

[bikerpete]

Running Sensorless may be possible with the addition of a clutch. This would reduce the erpm limits to some degree. This would open up the larger rc motors to the app also. Out runners would lend themselves perfectly to a drum flywheel that could incorporate a eddy clutch.

Dreaming again, sorry. This one caught my eye.

Beauty in combining the heat soak of the clutch and motor is you have only one cooling loop (as large as needed) to handle the heat. The spinning lends itself to passive pumping to a external pipe loop style cooler with excellent heat path conductivity of the oil. If you can get a dry clutch setup to work, air (fan) from the motor side with properly places fins could work to cool both sides.

Progression of power and torque and flywheel optimization is a interesting study. The larger diameter rotors would allow more broad relative torque band. You can have and be able to use high levels of torque at near zero rpms with them eliminating the need for super high rpms. 5 rpm to 5000 rpms is 1000x; while 500 to 15000 rpms is only 30x. Lots to ponder.

Interesting idea using an outrunner - ditch the aluminium shell and replace it with a whacking great copper flywheel mass (higher SG than steel and good heat transfer).
Trouble is it looks like these outrunners are fairly low RPM. There are only 3 factors at play in flywheels - RPM, Mass, Radius.
Energy stored goes up by the square of RPM, but proportionally to mass and radius. EDIT: I've been quite rightly corrected. Inertia goes up by the square of RPM and Radius, but directly proportionally to mass. So if you reduce the RPM you either need a large or a heavy flywheel, or both. That's why I was thinking of a higher RPM motor with a relatively small & light flywheel.

I'm becoming increasingly convinced that chasing torque is totally the wrong path. Enough is all you want. More is worse.
In trials you typically only need big torque for very short moments, which a flywheel provides nicely - the rest of the time you need "enough" torque delivered linearly and smoothly.

Oh, and my understanding is that dropping 5000 rpm provides the same amount of power out of a flywheel irrespective of the starting RPM. So 5-500rpm is a fraction of the energy return of 500-15000.

The Astroflight 3220 came up as a possible higher RPM motor - possibly a little low on power? I haven't done any real research on them though.

 

[speedmd]

A good basic flywheel calculator for comparing some size ideas. http://www.calculatoredge.com/mech/flywheel.htm

Most Out runner rotors will explode if over revved too far. Outer clutch rings could double as supports or support rings added to the rotor as clutch mounts and double as cooling fins. The latter would decouple the motor and clutch heat paths and allow the clutch rings to expand and move as needed without inducing stresses on rotor other than those transmitted from the fixturing bushings. Much as a floating brake rotor does now. 8) These high pole- tooth count motors are known to challenge many controllers erpm limits well below your target RPM range. On In runners, the IPM designs typically yield greater rpm extension with field weakening. Only hits to the higher KV-rpm setups is that the lower rpms you go the sooner you will be pushing the motor into a lower efficiency range as well as possibly require a greater (2 stage) gear reduction. Much discussion on the topics years back. I recall comparing the Joby JM1 to the JM2 and many of the trade offs involved.

Miles has some great resources on motors page. https://endless-sphere.com/forums/viewtopic.php?f=30&t=16056&start=25

 

[speedmd]

Mecatecno Trials E moto build (with clutch). 54 kg wet!

 

[jkrienert]

Then last year I began building another one. It is based on an RSD frame, with a carbon fiber fork, also from RSD. This time I decided to go with an Astro Flight 3210 motor, coupled to a planetary gearbox of 5:1. Then a chain ratio of 8:1 to the rear wheel. The controller is a Kelly KLS-S. This setup has much more power than the BHT motor, but I have problems with overheating and I don't like the noise of the gearbox, even if minimal. I'm spoiled now, I might have to go back to the BHT. Anyway, it has Carbonfan rims like on the earlier bike, and the battery is made from 18650 cells in a 14S7P arrangement. It has 1KW/Hr of capacity. The front tire is a 4.8" like the previous bike, but the rear one is a 5.2" this time for maximum flotation on snow. BTW, fatbike tires are larger in diameter than regular trials tires and 5.2 is about 31" high! People always comment on this tire when they see it.
View attachment 266456
Pegs are way back, handlebars are as rearward as possible thanks to a short stem. Stance is perfect on the bike with knees slightly bent. It's very easy to lift the front end over obstacles, and there is good clearance under the frame, with a plastic skidplate. The seat is kept low so I can put my ass right over the rear tire when needed. The bike is narrow so it retains its "bicycle" qualities. The battery box is only 3" wide, compared to 6" for the previous bike. There is now a chain tensioner which is hung from the derailleur bracket. Even though the bike has a standard rear hub that comes with an integrated frewheel, I've disabled it and the sprocket carrier is now firmly locked into the hub, making a direct-drive assembly. The reason for the disabling is that I was constantly breaking pawls inside the freewheel, because of the free play. As there are only 28 points of engagement around the freewheel, that gives up to 13 degrees of free rotation of the sprocket before the pawls engage. If you've opened the throttle quickly, the motor has had time to accelerate quite a bit before the pawls engage, and then they engage with a bang. So direct-drive is much better, it gives instant throttle response to the wheel, and the drag of the un-powered motor is not high enough to cause a problem.
View attachment 266457
I machined all the parts myself, including the battery box which is made of fiberglass. The bike weights 54.4 lbs (24.7Kg). Ideally the controller should be placed inside the battery box, or at least in an enclosure, but this was impossible here as the frame triangle was too small to enclose everything. And yes the controller looks enormous because it has an integrated heatsink. I should have cut it off as there is no need for it in this low-power application. BTW if you wonder what's hanging from the handlebar stem, it's a softcase that houses my phone and displays Speed, Voltage, Amps, Watts, temperature. It communicates with a TinyBMS located inside the battery box. Front brake is on the right, like all motorcycles.

Wow, those are some clean 'horses'. It looks like that controller could maybe fit in the frame above the top tube and in front of the seat post, or even in the front of the head tube? Though, both those mount points would be challenging to maintain stability. All in all... nice builds. Production level quality. Kudos.

 

[bikerpete]

Mecatecno Trials E moto build (with clutch). 54 kg wet!

Yep, the Dragonfly is a nice looking machine, but after looking at every video I can find of it I'm convinced it's achilles heels aee once again - lack of any significant flywheel mass & too high gearing. Such a shame they dropped the prototype's gearbox.

Video 8 in this lot demonstrates to me the shortcomings. In the gaps you can really see way the power just disappears as soon as the clutch is popped. When he rides the flat straight after the gap you also see the pretty brutal acceleration on throttle.

Definitely a significant addition to the e-trials world & I'd love to try one, but no way it's as capable as current ICE bikes.

I should add, for most real-world ordinary riders the EM or the Dragonfly are way more capable than the rider. But they're missing the icing on the cake that ICE bikes have.

GasGas have also returned to development of their e-trials bike
Unfortunately it also appears to lack a flywheel & the 6 speed box is probably complete overkill that will never get used in full.
But it's a better looking layout than the original txt-e. And now they're owned by KTM their might be some extra minds brought to the project.

 

[DingusMcGee]

bikerpete,

This virtual ?flywheel edirtbike idea has some merits for modulated short bursts of power (and/or momentum) in wheelie assisted maneuvers.

For those finite derivative short time span momentum shock waves needed from a rotating metal flywheel something the size of a QS 2000 motor with 20 Ah LiPo cells at 24C discharge rate (= 480 amp production very quickly) and additional burst controller at set at 480 amp with a thumb button when actuated would pitch your ass backwards off the edirtbike by a very quick wheelie much faster backwards than you were creeping forward with the rear tire on the ground.

How do I know this? I modified a Hall sensor throttle so it had no rotational lag from the zero position to sending a signal to the controller. A little whiskey throttle would send my ass backwards and off the bike so fast I would be out of the way of bike doing a wheelie over and it landing on the seat. I quickly threw that throttle away.

I do have the physics skills to do the math but I do not know any frequency response times like amperage rise rate that a LiPo cell can generate.

To test this thumb button burst switch you will need some very high discharge cells, 2 controllers and a thumb button on the left. You will use a left bar rear brake to modulate a burst. For the burst controller get something like the 9 series Fardriver race controller with many programmable features. Share the Hall sensors and wire the phase wires for forward direction.

You might be able to wire a CA onto the second controller throttle as they have a throttle setting for instantly going to full throttle.

If powering the same active phase wires creates some problem with the full time controller, add three 2ms cutout relays to its phase wires and have the thumb switch actuate them.

Likely you will not need a metal freewheel and clutch after you learn controller burst pulse skills?

 

[speedmd]

Not sure it is better to add flywheel mass and gear box -ratio options vs just adding the weight into a larger motor (higher torque) with significantly more rpm - power range.

It is some 30% lighter than my gasser. That alone makes me want one to play around with.

The clutch looks like a good design , and spinning side clutch plates could be possibly made more massive to play around with a bit. Field weakening would allow it to spin much faster. Certainly reduction ratios would need to be dialed a bit.

 

[bikerpete]

Likely you will not need a metal freewheel and clutch after you learn controller burst pulse skills?

Somewhere back in this thread (or maybe it was another - I had quick look but didn't immediately see it) there's some discussion about the very large difference between rising and falling power delivery.
What you're suggesting is rising power, what a flywheel delivers is falling power. They are in no way interchangeable for trials use.
Two key points are that the power needs to fade as it is delivered and if traction is lost the system must not accelerate further. A flywheel does this totally automatically, throwing a burst of power out of a controller does neither and actually does the opposite in both cases.

Nice idea, but it would be a disaster for trials riding. Trials is largely about creating and using traction - you have to be able to very accurately modulate power to match available traction - when there's bucket loads of traction you can supply bucket loads of power, but in a few hundred milliseconds when that traction halves you need to be able to supply half the power. Often once the wheel spins, it's all over.

Not sure it is better to add flywheel mass and gear box -ratio options vs just adding the weight into a larger motor (higher torque) with significantly more rpm - power range.

It is some 30% lighter than my gasser. That alone makes me want one to play around with.

The clutch looks like a good design , and spinning side clutch plates could be possibly made more massive to play around with a bit. Field weakening would allow it to spin much faster. Certainly reduction ratios would need to be dialed a bit.

I know we differ on this, but it's that difference between the nature of the power delivery once again, not just the amount of power. The Dragonfly has plenty of power at lower speeds (I've heard a couple of reports saying it lacks power at higher speeds), more wont be better.
The clutch is definitely good - probably the same suppliers that several of the other brands use by the look of it.
Adding mass to the clutch plates would be completely pointless - the clutch is geared down from motor RPM considerably so any mass added there is pretty ineffectual, and the amount of mass you could add would be so small as to be un-noticeable. Those clutches are often setup with varying thickness pressure plates to get the exact feel that's needed. The overall pack thickness is varied by 10ths of mm to create quite noticeably different characteristics. Adding a few tenths of a mm of steel to the plates is totally insignificant. Maybe if you machined a steel basket to replace the aluminium one ... but it's going to be grams of difference at low RPM, not the kg's at high RPM I think would make a real difference.

What makes you think they aren't using field weakening already? I've no idea if they are or not, but given they've been working on this thing for several years and FW isn't exactly secret rocket science I fully expect that if they could usefully use it, they probably are.
Perhaps the reports that the power falls away significantly at higher speeds suggests they are in fact using FW?

This drop in power at higher rpm is a problem on the EM too.
I wonder if a higher kV motor with bigger reduction would help to solve that? That's something I'm currently trying to learn about - torque/power curves of motors as their kV is changed.