Observed trials bike designing

General Discussion about large electric scooters and motorcycles and other things with no pedals.
bikerpete   1 W

1 W
Posts: 58
Joined: Jan 27 2020 9:09pm

Observed trials bike designing

Post by bikerpete » Jan 29 2020 7:46am

I am looking into building a full size moto trials bike, using an existing IC bike rolling chassis.
I'd like it to perform similarly to a modern 250 - 300cc 2-smoke trials bike. I currently ride a 2017 Beta 300 at bottom end of B-grade level.
It will require a mechanical clutch (kind of irrelevant to the electrical design, but included here so there's no misunderstanding.) and a reduction stage. Top speed is largely irrelevant, everything of interest happens between stationary and running pace.

I'm not looking to build this for absolute bare budget, but will save dollars wherever it makes reasonable sense. Looking for something that performs well and is pretty reliable. Not worth spending any money if it doesn't do this.

Finally, I'm in Australia, so bulky/heavy items out of USA usually become prohibitive with freight costs.

My expectation based on the better commercial trials bikes out there is that I need around 5kw continuous and 10-15kw peak (just a couple of seconds every few minutes max duty cycle). Most of the time the bike would be barely using a kilowatt - a lot of time is spent rolling downhill, standing almost stationary, positioning the bike etc. Then brief moments of high power.

Some of the commercial specs are eg:
EM EPure - 6 kW continuous/11 kW peak (Golden Motor HPM5000B 48V I believe). 600 Nm at rear wheel. Mechanical clutch, single speed reduction.
Mecatecno Dragonfly - 15kw, 48.8V/58.8V min/max 360A max. Samsung 25R cells, 25 Ah. Mechanical clutch, 5 speed gearbox.
GasGas TXTe - 15Kw, 50.4V/58.8V min/max, Cells 3000mA. Mechanical clutch. 6 speed gearbox.
Yamaha TYE - can't find specs, mechanical clutch, single speed reduction.

My initial investigation leads me to think the following components might be worth a try:
Motor: QS 3000W 138 70H
Controller: Nucular 24F
Batteries: Samsung 25R possibly, although I'd really like to use prismatic or pouches to get a bit better energy density (space is at a premium in a trials bike). I doubt it's feasible to get batteries delivered out of USA or EU, so it's probably going to be Asian supply.
BMS: haven't looked into it yet until batteries are finalised

Motors:
The QS seems good value and quite a few people are pushing them way above their rated power with seemingly no ill effects.
Obviously the Golden Motor also does the job so perhaps it's a better option?
Any others that spring to mind?

Controllers:
I liked the sound of the ASI BAC4000, but $1,120 plus exchange plus freight seems a bit too much of a premium!
The Nucular seems highly regarded here, but I'm not clear exactly how configurable it is having read through their Wiki.
I'd be more than happy to consider other controllers also, but definitely want something that's reliable.

I'd want to be able to have 2 configurations available by switch so there's a lower power consumption mode for use between sections and for people to try it out with. That's no big deal in many controllers.
Also expect there will be a fair bit of tuning to get the right throttle response. Trials tends to rely on flywheel inertia a fair bit to maintain traction, this bike will have some flywheel effect in the clutch and motor mass, but potentially some effect will need to be created electronically. So throttle ramps/maps might be quite key getting a good response.

Batteries:
At this stage it kind of looks like 18650 cells are maybe the best option simply because they are readily available. I'm kind of reluctant to spend all the time, effort and money on building a battery up with them when really I'd prefer to better fill the available space. I just haven't come across anything that really looks better bang for the buck into Australia.

I still need to design the primary drive to the clutch, but that'll come after I've fixed the motor and the donor bike (probably a GasGas). My feeling is that a 2 speed gearbox would be an advantage for this application, but it's certainly not going to happen in the first incarnation.

Any suggestions gratefully received.

User avatar
DanGT86   10 kW

10 kW
Posts: 730
Joined: Sep 06 2012 9:18pm
Location: Saint Louis MO

Re: Observed trials bike designing

Post by DanGT86 » Jan 29 2020 10:26am

Sounds like a very doable build goal. For batteries I would check with member Jonescg here on ES. He builds nice high output batteries with pouch cells and is in Australia.
viewtopic.php?f=31&t=38252&start=550

As for motor and controller, have you checked out the LMX 60mm motor? It might be enough power for you and is lighter than the QS 3000watt. I think the QS might be a bit heavy for your application and provide more power than you need. As for controller, I think the nucular is a great choice. That't the most compact and configurable controller I can think of at a reasonable price. Vasili is also continuing to update it so you may be able to request specific features related to your flywheel and clutch needs.

There has been much discussion on this forum about trials riders needing the clutch in an electric bike. I can see arguments from both sides making sense but can absolutely see why you would want it since clutch work is so vital to that style. It also wont hurt to have it even if you dont need it.

The flywheel is a harder question. The prevailing wisdom seems to be not to carry any additional weight that doesnt do something for you. Specifically add weight in the form of motor or battery rather than gearbox or flywheel. The unknown is weather the flywheel helps by storing energy to be unloaded quickly or helps by keeping the ICE motor in its powerband. If its the latter and its all about powerband then you probably don't need it. You will have many times the torque on hand with the electric motor compared to gas and the rotor in the electric motor has some mass acting as a flywheel when used with your clutch.

I'm always hoping more pure trials builds happen to answer these questions.

As for batteries, checkout Jonescg and also consider a modular design that's easily to swap. You can get a crazy amount of power out of just a few kg of lipo pouches so you could have a lower power but higher energy density pack in a backpack or in the frame to get you from one zone to the next. Then you could remove most of it for the pure trials riding. I'm sure ditching 5kg of battery will help your riding quite a bit.

Good luck. I'll be looking forward to seeing how it goes together for you.

thoroughbred   10 W

10 W
Posts: 99
Joined: Apr 11 2018 6:18pm
Location: virginia

Re: Observed trials bike designing

Post by thoroughbred » Jan 29 2020 6:51pm


bikerpete   1 W

1 W
Posts: 58
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Jan 29 2020 8:29pm

DanGT86 wrote:
Jan 29 2020 10:26am
Sounds like a very doable build goal. For batteries I would check with member Jonescg here on ES. He builds nice high output batteries with pouch cells and is in Australia.
viewtopic.php?f=31&t=38252&start=550

As for motor and controller, have you checked out the LMX 60mm motor? It might be enough power for you and is lighter than the QS 3000watt. I think the QS might be a bit heavy for your application and provide more power than you need. As for controller, I think the nucular is a great choice. That't the most compact and configurable controller I can think of at a reasonable price. Vasili is also continuing to update it so you may be able to request specific features related to your flywheel and clutch needs.

There has been much discussion on this forum about trials riders needing the clutch in an electric bike. I can see arguments from both sides making sense but can absolutely see why you would want it since clutch work is so vital to that style. It also wont hurt to have it even if you dont need it.

The flywheel is a harder question. The prevailing wisdom seems to be not to carry any additional weight that doesnt do something for you. Specifically add weight in the form of motor or battery rather than gearbox or flywheel. The unknown is weather the flywheel helps by storing energy to be unloaded quickly or helps by keeping the ICE motor in its powerband. If its the latter and its all about powerband then you probably don't need it. You will have many times the torque on hand with the electric motor compared to gas and the rotor in the electric motor has some mass acting as a flywheel when used with your clutch.

I'm always hoping more pure trials builds happen to answer these questions.

As for batteries, checkout Jonescg and also consider a modular design that's easily to swap. You can get a crazy amount of power out of just a few kg of lipo pouches so you could have a lower power but higher energy density pack in a backpack or in the frame to get you from one zone to the next. Then you could remove most of it for the pure trials riding. I'm sure ditching 5kg of battery will help your riding quite a bit.

Good luck. I'll be looking forward to seeing how it goes together for you.
Thanks for that input.
I'll definitely follow up Jonescg's batteries. I've also been looking at Samsung 21700 cells as an option. More research.

The LMX 60mm is certainly considerably lighter and could possibly be an option. Given the commercial bikes that are now getting decent reviews are running up to around 15Kw and using 10kg motors, I'm a bit leery of designing around a motor so small. Again, more research to follow and a much appreciated suggestion. If it worked, the 5k weight saving would be great! I'll definitely research that one.

Re flywheel, I'm not intending to add any flywheel mass, what's in the motor and clutch should be enough already I think. Pair that with some careful acceleration / deceleration profiles in the electronics and I'm expecting it'll be good. That's where having a controller with plenty of flexibility in throttle/speed mapping through the whole range might become important. I've no experience in this directly, just thinking it through.

For trials the flywheel is all about storing energy, trials bikes run very heavy flywheels - a lot of moves typically require spinning the bike up (the pros hold the throttle open until the bike starts four stroking, then drop the clutch - that's a huge amount of energy for an instant. Way too scary for me!), then accelerating the bike off the clutch. One of the key things for many of these is that to get the bike to really jump, the throttle must be closed as the clutch is dumped, I'm not completely sure what the mechanics of this is, I think it's probably got something to do with releasing tension on the chain as the suspension extends. This is where flywheel effect is pretty hard to replace - you just can't roll off the throttle fast enough or accurately enough to replicate dropping the clutch on a high revving engine with the throttle closed. I suspect replicating that flywheel effect electronically might be possible, but in reality having a bit of a flywheel might be vastly simpler.

thoroughbred - that Eastgem gearbox/motor looks pretty cool, but I very much doubt it would work for trials for a host of reasons. Might make a nice kids trials bike setup perhaps.

User avatar
DanGT86   10 kW

10 kW
Posts: 730
Joined: Sep 06 2012 9:18pm
Location: Saint Louis MO

Re: Observed trials bike designing

Post by DanGT86 » Jan 30 2020 8:18am

Some controllers allow you to have virtual freewheel effect where the motor gets just enough current sent to it to overcome the natural magnetic drag it would have. Almost every controller has variable regen as well. You could set that up to act opposite your throttle. I think you'll be able to replicate the effect you are looking for with a nucular. I believe the VESC based controllers have the virtual freewheel option.

If you are that used to a standard ICE bike then you probably want to keep a low torque but constantly rotating motor so you can creep along with clutch work with closed throttle.

I have not heard of the clutch dump with the throttle closed. I always thought they revved up and dumped the clutch just letting go of the throttle in the air as its no longer needed. I didnt know there was a deliberate quick closing of the throttle.

bikerpete   1 W

1 W
Posts: 58
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Jan 30 2020 5:02pm

DanGT86 wrote:
Jan 30 2020 8:18am
Some controllers allow you to have virtual freewheel effect where the motor gets just enough current sent to it to overcome the natural magnetic drag it would have. Almost every controller has variable regen as well. You could set that up to act opposite your throttle. I think you'll be able to replicate the effect you are looking for with a nucular. I believe the VESC based controllers have the virtual freewheel option.

If you are that used to a standard ICE bike then you probably want to keep a low torque but constantly rotating motor so you can creep along with clutch work with closed throttle.

I have not heard of the clutch dump with the throttle closed. I always thought they revved up and dumped the clutch just letting go of the throttle in the air as its no longer needed. I didnt know there was a deliberate quick closing of the throttle.
Cheers for that. I hadn't been aware of the virtual flywheel setting.
I'm not sure if it's the same thing, but I was thinking flywheel effect would be achieved by a throttle ramp setting, so when you snap the throttle closed the controller (quickly) tapers the power to the motor.

The other effect of a flywheel that I'm not so sure about replicating electronically is driving traction on slippery surfaces - a big flywheel let's you power up, then roll over slippery bits under inertia only, so there's power on the wheel, but its directly related to wheel speed, so there's much reduced chance of wheelspin. A lot of trials is simply about traction control.
It could be a trailing throttle will achieve the same, or it could be that with a torque based controller that if the wheel looses traction it will spin up, I don't understand them well enough, or have any experience.

It's interesting that Electric Motion's latest & greatest bike (that's getting really good rider reviews) now has "engine flywheel with different set of balancing weights" - seems you can physically add or remove flywheel weight. If they're finally going that route, it sounds like replicating flywheel physics electronically is very hard!

I think the slow creep will be better done by getting used to using throttle instead of clutch, maybe combining them in various situations.

User avatar
DanGT86   10 kW

10 kW
Posts: 730
Joined: Sep 06 2012 9:18pm
Location: Saint Louis MO

Re: Observed trials bike designing

Post by DanGT86 » Jan 30 2020 10:31pm

My low tech old school ebike controllers have the option to have regen any time the throttle is released so it feels like engine braking in a gas motor. There are settings for soft or hard regen force. These are the lowest tech "infineon" controllers on the market from almost 10 years ago. Any of this new generation of controllers like nucular, bac, sevcon, VESC, etc. will likely be configurable for you.

bikerpete   1 W

1 W
Posts: 58
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Jan 30 2020 11:18pm

DanGT86 wrote:
Jan 30 2020 10:31pm
My low tech old school ebike controllers have the option to have regen any time the throttle is released so it feels like engine braking in a gas motor. There are settings for soft or hard regen force. These are the lowest tech "infineon" controllers on the market from almost 10 years ago. Any of this new generation of controllers like nucular, bac, sevcon, VESC, etc. will likely be configurable for you.
I'm guessing it's not so much braking the motor as getting it to drop revs progressively even when the throttle is chopped to 0. And I think key is that even if the wheel looses traction the revs should continue to fall progressively.
That's where I'm not sure how modern torque based controllers operate. If load suddenly drops but the controller is running a throttle downslope, will the motor revs increase or will it continue to drop revs?

A mechanical flywheel at revs but with no throttle will continue to drop revs even if the wheel gets airborne, so when the wheel gets contact again it's still turning somewhere close to ground speed and can regain traction. That's one of the reasons trials moves are often done with high revs and no throttle.
Those with more skill can manage the throttle and clutch more accurately so probably don't rely on the flywheel effect quite as much as we nuftys.

rynhardt   10 mW

10 mW
Posts: 29
Joined: Sep 16 2019 5:26am
Location: Johannesburg, South Africa

Re: Observed trials bike designing

Post by rynhardt » Jan 31 2020 6:51am

I have recently completed a GasGas trials bike electric conversion, you can have a look over here:

viewtopic.php?f=10&t=102413

The plan is to understand the shortcomings and address these as I progress through the year.
For example I have already wired up a boost button on the left handlebar for short-run uphill sections.

I am yet to be convinced that a clutch or flywheel is necessary but will reserve comment on this until I've completed a few more events.

For sure you need a different riding style and much better throttle control if you cannot modulate power with the clutch. Once again the challenge to myself is to discover how to replicate double blips, zaps and splats without a clutch.

I'm looking forward to seeing how your build works out! :thumb:

User avatar
DanGT86   10 kW

10 kW
Posts: 730
Joined: Sep 06 2012 9:18pm
Location: Saint Louis MO

Re: Observed trials bike designing

Post by DanGT86 » Jan 31 2020 10:31am

Nice build thread rynhardt. Nice to see some real world testing and feedback.

One of the thoughts in defense of a clutch is the ability to get the motor barely rotating to apply max torque as opposed to a total dead stop. The initial startup is hard on the motor and controller so even single digit RPM to allow the motor and controller to be synced could be really beneficial for a high current burst.

Regarding the motor ramping down progressively, the cycle analyst allows the throttle down rate to be precisely controlled just like the throttle up rate. It shouldn't be hard to make a motor react like a gas motor in the up and down direction. In a gas motor the throttle is modulating torque. If you have an abundance of torque for the load the rpm rises. If you have too little torque the rpm falls. Its as simple as that. Using amp based control or power based control for the throttle feels very much like an ICE. The ability to dial in a precise ramp should make the electric motor better in every way. You could even theoretically cheat and implement traction control if you had precise enough sensors of wheel speed and load.

I have often thought it would be cool to make the throttle output to the controller the sum total of a rising and falling voltage from the throttle and a fake clutch lever (no actual clutch). In this case full throttle with the clutch in would yield no throttle output. Inversely full throttle with the clutch lever released would be max power. In this case it would feel like a normal bike where you still modulate power with the clutch in precise situations like trials. Its easier to squeeze a lever under rough conditions climbing and hopping than it is to precisely angle the throttle while pulling on the bars.
Just a thought.

An actual mechanical clutch might still be more natural for experienced riders with years of muscle memory.

bikerpete   1 W

1 W
Posts: 58
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Jan 31 2020 4:08pm

DanGT86 wrote:
Jan 31 2020 10:31am
One of the thoughts in defense of a clutch is the ability to get the motor barely rotating to apply max torque as opposed to a total dead stop. The initial startup is hard on the motor and controller so even single digit RPM to allow the motor and controller to be synced could be really beneficial for a high current burst.

Regarding the motor ramping down progressively, the cycle analyst allows the throttle down rate to be precisely controlled just like the throttle up rate. It shouldn't be hard to make a motor react like a gas motor in the up and down direction. ...

I have often thought it would be cool to make the throttle output to the controller the sum total of a rising and falling voltage from the throttle and a fake clutch lever (no actual clutch). In this case full throttle with the clutch in would yield no throttle output. Inversely full throttle with the clutch lever released would be max power.
I hadn't realised that "barely rotating" scenario for electric - another + to a clutch.

Also hadn't really looked into Cycle Analyst - I'd seen lots of mentions and had a brief look, but kind of figured it was more of a dashboard than controller manager as it sounds like it actually is. I guess that might have an influence on which controller I choose? I'd started leaning toward the cheap and cheerful Votol that QS (my most likely motor at this stage) supply but I'll need to see if the CA will interface to that nicely. Yet more research.

The reverse throttle style electronic clutch is what Electric Motion have used for years - it works after a manner, but it simply doesn't give the instant snaps of power that a mechanical clutch does. Watching EM riders you see they never ever make the explosive moves that are typical of modern trials. Still very competent bikes, but the proof is in the pudding that EM have now added a mechanical clutch to their top bike in line with GasGas, Tecno and Yamaha.
That's potentially going to be the hardest part of this build - designing and building a 1st stage reduction drive and wet-bath clutch housing - I'd like to ditch the gearbox and just keep the clutch.

Down the track I expect a 2 speed gearbox could be really useful - gear 1 in sections, then gear 2 in the transfers between sections to save power and increase speed. Not this time around though!
rynhardt wrote:
Jan 31 2020 6:51am
I have recently completed a GasGas trials bike electric conversion, you can have a look over here:

viewtopic.php?f=10&t=102413

The plan is to understand the shortcomings and address these as I progress through the year.
For example I have already wired up a boost button on the left handlebar for short-run uphill sections.

I am yet to be convinced that a clutch or flywheel is necessary but will reserve comment on this until I've completed a few more events.

For sure you need a different riding style and much better throttle control if you cannot modulate power with the clutch. Once again the challenge to myself is to discover how to replicate double blips, zaps and splats without a clutch.

I'm looking forward to seeing how your build works out! :thumb:
I've looked over your build a couple of times - good job and achieved quickly (I tend to vacillate for so long that technology moves on and I have to start over :D ).
I'm very confident that motor wouldn't be big enough for what I want, and I really, really doubt you'll be doing zaps of any sort either with that amount of power or without a clutch. Double blips are well achievable though and can get you over lots of stuff that most people would zap.

Looking forward to further reports too.

User avatar
tigcross   100 W

100 W
Posts: 127
Joined: Oct 18 2015 10:47pm
Location: Victoria BC
Contact:

Re: Observed trials bike designing

Post by tigcross » Jan 31 2020 9:05pm

I know this is a bit of a stretch, but have you considered an Axial Flux Motor such as the Lynch? https://lynchmotors.co.uk/vehicle-motors.html

They have a HUGE amount of torque compared to a conventional motor, and are designed to run at lower RPMs which would be an asset for a trials bike. The downside is that the existing Axial Flux Motors are all DC, which involves heavier and more complicated motor controllers.

There is a Belgian Start-up that is working on an AC Axial Flux Motor. https://www.magnax.com/

I used an E-tek, a licensed version by Briggs, in an early version of my project and it was insane, my bike would get from Zero to 80 KPh in about 7 seconds. In the end it was way overpowered for what I wanted to do.

You'll probably dismiss the idea as DC seems pretty old-school now, and there are far fewer choices, but I thought you might like to know about these motors for a future build.

Looking forward to seeing where you build goes.

User avatar
DanGT86   10 kW

10 kW
Posts: 730
Joined: Sep 06 2012 9:18pm
Location: Saint Louis MO

Re: Observed trials bike designing

Post by DanGT86 » Jan 31 2020 11:24pm

bikerpete wrote:
Jan 31 2020 4:08pm

Also hadn't really looked into Cycle Analyst - I'd seen lots of mentions and had a brief look, but kind of figured it was more of a dashboard than controller manager as it sounds like it actually is. I guess that might have an influence on which controller I choose? I'd started leaning toward the cheap and cheerful Votol that QS (my most likely motor at this stage) supply but I'll need to see if the CA will interface to that nicely. Yet more research.

That's potentially going to be the hardest part of this build - designing and building a 1st stage reduction drive and wet-bath clutch housing - I'd like to ditch the gearbox and just keep the clutch.

Down the track I expect a 2 speed gearbox could be really useful - gear 1 in sections, then gear 2 in the transfers between sections to save power and increase speed. Not this time around though!
A brief history on the Cycle analyst and ebike controllers in general: Back in the day most of the hotrodding was done using cheap made in china generic ebike controllers commonly referred to as "infineon" style controllers. These are commonly found on low powered ebikes. As people started to build higher powered bikes the controller manufacturers added more mosfets to handle larger current. The basic control logic was the same but the power handling part of the controller was beefed up.

These infineon style controllers were all speed based throttle. So the throttle angle corresponds to a particular speed and the controller will basically dump as many amps as it can to give you the commanded speed. This was not a problem with low powered bikes under 1000-1500watts. Anything over that and they get really jerky and hard to ride. Just because the controllers could handle the higher power didn't mean they were good at it.

The solution to that was the cycle analyst. The CA now in its 3rd generation is not only a dashboard but more of a control center for all the ebike functions like throttle and brakes. When using a cycle analyst you run the throttle and brakes into the CA and using current and speed measuring it sends the appropriate signals out to the controller. So if you set a slow throttle ramp it is pulsing the output so the controller thinks you are slowly twisting the throttle. If you have a 100amp controller but tell the CA to only output 50amps then its modulating the output to keep the controller from exceeding your limits. So by using a dumb cheap speed based controller you can have full programmable control of the throttle ramps, pedal assist power, regen, low voltage cut off, multiple power levels, current limiting, cruise control, motor temp limiting etc.

Its a super handy device and at one time it was the only affordable option to make a nice high powered bike pleasant to ride. Now days the advanced controllers have all of these features in them so the CA is not really needed. It makes a nice dash but would be redundant as a throttle buffer. have 2 bikes using them with cheap controllers around 6kw and 4kw. When those controllers break I'll likely sell the CAs and move on to something like Nucular. Its just hard to justify dealing with the low tech infineon style controllers anymore when ultimately the CA is really just a band-aid for those controllers inferior control logic.

For the gearbox, some users have used 2 stroke engine cases and machined off everything but the transmission. If you put your motor where the crank used to be you can retain the clutch and gears in their original location. Typically only 2 gears are needed so depending on the transmission you can remove the extra gears to save weight. This is a bit of a large and heavy solution for a small trials frame and more suitable on a MX bike but it does save a bit of complex engineering since you really only need to connect the electric motor to the clutch rather than building a whole transmission. Unless you are really needing trials and hwy cruising speed I think the 2 speed isn't worth the complexity. You will have such an abundance of torque with a 10kw electric motor that even geared for 60-70kph you'd still have to limit the current to keep it from looping you out. I wouldn't consider a 2speed unless I was trying to do slow enduro/trials riding and freeway cruising with the same bike. Even then the math says to just buy a larger motor.

It might be worth having a look around at brushed DC motors but brushless stuff is the weapon of choice around here. There will be a lot more tech support from most ES users regarding controllers. Also all the newest controller developments are aimed at brushless motors. Every brushless controller I know of does regen which is nice if you are trying to simulate the engine braking and quick off throttle rpm drop of a gas motor. I rely heavily on engine braking when trail riding on gas bikes and would feel weird without it. Only some brushed DC motors are optimized for regen.

The Votol controller looks like a lot of power for cheap but after following the Votol thread I'm worried about how hard it is to program. Having to figure out software that is half chinese just doesn't seem fun. I have emailed back and forth with QS a few times and its hard to get good clear info from them because of the language barrier. I love the idea of that under rated 3000watt motor but on the controller side you have some pretty specific configuration needs. Maybe email them about the Votol regarding the throttle mapping you want and see if the tech support and communication look promising.

bikerpete   1 W

1 W
Posts: 58
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Feb 01 2020 12:50am

DanGT86 wrote:
Jan 31 2020 11:24pm
A brief history on the Cycle analyst and ebike controllers in general:

For the gearbox, some users have used 2 stroke engine cases and machined off everything but the transmission. You will have such an abundance of torque with a 10kw electric motor that even geared for 60-70kph you'd still have to limit the current to keep it from looping you out. I wouldn't consider a 2speed unless I was trying to do slow enduro/trials riding and freeway cruising with the same bike. Even then the math says to just buy a larger motor.

Every brushless controller I know of does regen which is nice if you are trying to simulate the engine braking and quick off throttle rpm drop of a gas motor. I rely heavily on engine braking when trail riding on gas bikes and would feel weird without it. Only some brushed DC motors are optimized for regen.

The Votol controller looks like a lot of power for cheap but after following the Votol thread I'm worried about how hard it is to program. Having to figure out software that is half chinese just doesn't seem fun. I have emailed back and forth with QS a few times and its hard to get good clear info from them because of the language barrier. I love the idea of that under rated 3000watt motor but on the controller side you have some pretty specific configuration needs. Maybe email them about the Votol regarding the throttle mapping you want and see if the tech support and communication look promising.
That is fantastic info, thanks Dan.

I'll rule the CA out based on this - the less bits the better as far as I'm concerned. Focus on getting a controller that does what I think I need.

Gearbox - I've thought about just using the complete box - it's what GasGas and Mecatecno have both done so obviously has some merit. I'd really prefer to start with a clean slate though to get a bit more freedom to position components and get rid of excess weight. Even on my ICE trials bikes I pretty much just run one or two gears for almost everything, only changing up when moving between sections.
GasGas trials gearboxes are really neat, light and compact designs, but they tend to be the Achilles heal of GasGas (I'm most likely to use a GG frame), another good reason to get rid of it. Part of the appeal of electric is less maintenance, more riding. It would be an easy way to get the clutch in there though.
The 2 speed thing is something I think might deserve future investigation - partly to make best use of a flywheel, and partly for speed/efficiency. Definitely not part of this project though.

Not a huge amount of engine braking in a 2-stroke anyway, and it seems most controllers will allow for some basic regen. which is probably enough. Perhaps an analog regen signal type system would be nice - a control on the bars so you can add more on the steep/long descents or brake with less skidding in the slippery bits etc, but I think (hope) it's not a big deal.

Perhaps foolishly, I'm not too scared about hard to program (perhaps because I haven't experienced that particular frustration yet), I'm pretty patient.

I think I've pretty much talked myself into the QS 3000W and the Votol EM200 (the 150 might be enough, but Chinese watts and amps always seem to be smaller than ours so I'd give it a bit of headroom). Still need to look into the Nucular a bit more.

User avatar
DanGT86   10 kW

10 kW
Posts: 730
Joined: Sep 06 2012 9:18pm
Location: Saint Louis MO

Re: Observed trials bike designing

Post by DanGT86 » Feb 01 2020 12:08pm

How about using variable regen and using the left side gear lever to trigger different strengths of regen? Basically this would feel somewhat like gears in that 1st engine brakes harder than 2nd etc.

Again, its quite a bit of engineering to simulate a gas motor feel so it really makes you analyze which gas motor behaviors are features and which are hindrances we've learned to put up with.

For me personally I like the idea of a clutch and an electric motor that always idles just a tiny bit. At least for pure trials. The fact that so much of trials is from truly zero rpm kinda sells the clutch. Any kind of trail riding where you are rolling at any speed probably negates the clutch.

I'd hate to end up becoming so accustomed to specifically electric behavior that I couldnt still ride a gas bike so keeping the rear brake as a foot lever and the left hand as a clutch is a preference for me on my theoretical moto build I cant afford. :wink:

rynhardt   10 mW

10 mW
Posts: 29
Joined: Sep 16 2019 5:26am
Location: Johannesburg, South Africa

Re: Observed trials bike designing

Post by rynhardt » Feb 02 2020 1:27am

When I did my build I considered having regen, but apparently it is variable depending on battery charge level. This would introduce another aspect that would needed managing so I skipped it.

As for becoming used to riding with different control setups, no worries there. I still ride my ICE bike and there is no confusion. It's a bit like switching between bicycle riding and motorcycle riding - my brain seems to treat these as completely separate activities.

rynhardt   10 mW

10 mW
Posts: 29
Joined: Sep 16 2019 5:26am
Location: Johannesburg, South Africa

Re: Observed trials bike designing

Post by rynhardt » Feb 02 2020 1:41am

bikerpete wrote:
Jan 31 2020 4:08pm
Down the track I expect a 2 speed gearbox could be really useful - gear 1 in sections, then gear 2 in the transfers between sections to save power and increase speed. Not this time around though!
I suppose this depends on how the event is set up, but over here the sections are generally no more than a few hundred meters apart so top speed is academic. Anyway you use approximately the same amount of energy to get from point A to point B - well almost, depending on your motor's efficiency curve which generally depend on load, not speed.

User avatar
stan.distortion   100 W

100 W
Posts: 139
Joined: Jun 12 2019 12:53pm

Re: Observed trials bike designing

Post by stan.distortion » Feb 02 2020 8:32am

DanGT86 wrote:
Feb 01 2020 12:08pm
How about using variable regen and using the left side gear lever to trigger different strengths of regen? Basically this would feel somewhat like gears in that 1st engine brakes harder than 2nd etc.

Again, its quite a bit of engineering to simulate a gas motor feel so it really makes you analyze which gas motor behaviors are features and which are hindrances we've learned to put up with.

For me personally I like the idea of a clutch and an electric motor that always idles just a tiny bit. At least for pure trials. The fact that so much of trials is from truly zero rpm kinda sells the clutch. Any kind of trail riding where you are rolling at any speed probably negates the clutch.

I'd hate to end up becoming so accustomed to specifically electric behavior that I couldnt still ride a gas bike so keeping the rear brake as a foot lever and the left hand as a clutch is a preference for me on my theoretical moto build I cant afford. :wink:
A clutch would be very hard to live without due to the stored flywheel energy. Trials bikes are usually around 20kw but the clutch allows a momentary burst several times that. Removing the flywheel also has other effects, deceleration, vertical stability, axial rotation (blipping throttle effects front/rear weight distribution) and probably more I've never thought of. Take away those effects and an E bike will never feel right to an IC rider, a very high speed declutchable flywheel should be able to keep or even improve those characteristics without sacrificing weight too much. Lol, and please ignore thoughts of spinning battery packs, it couldn't possibly end well! ;)

User avatar
DanGT86   10 kW

10 kW
Posts: 730
Joined: Sep 06 2012 9:18pm
Location: Saint Louis MO

Re: Observed trials bike designing

Post by DanGT86 » Feb 02 2020 9:26am

rynhardt wrote:
Feb 02 2020 1:27am
As for becoming used to riding with different control setups, no worries there. I still ride my ICE bike and there is no confusion. It's a bit like switching between bicycle riding and motorcycle riding - my brain seems to treat these as completely separate activities.
The muscle memory thing is interesting in this regard. Not having a clutch frees up the left hand lever. The obvious choice seems to be to put one of the brakes there. That then leads to the question of whether to run the rear on the right or left. I've often wanted to swap my trials bicycle brakes over to moto style (rh front lh rear). Seems logical to standardize but its so ingrained at this point I have a hard time on the bicycle with them swapped.

Here is another flywheel question. If the max RPM of gas trials bikes is 10-12k rpm but most of the electric motors available for this rev to 6krpm max, do you need to use a heavier flywheel to get the same stored energy and benefit? Does trials riding on a gas bike ever use that max rpm?

Side note, you guys inspired me with all of this talk so I got the trials bicycle out yesterday for the first time in a while. My everything hurts today. :thumb:

bikerpete   1 W

1 W
Posts: 58
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Feb 05 2020 6:08am

DanGT86 wrote:
Feb 02 2020 9:26am
I've often wanted to swap my trials bicycle brakes over to moto style (rh front lh rear). Seems logical to standardize but its so ingrained at this point I have a hard time on the bicycle with them swapped.

Here is another flywheel question. If the max RPM of gas trials bikes is 10-12k rpm but most of the electric motors available for this rev to 6krpm max, do you need to use a heavier flywheel to get the same stored energy and benefit? Does trials riding on a gas bike ever use that max rpm?

Side note, you guys inspired me with all of this talk so I got the trials bicycle out yesterday for the first time in a while. My everything hurts today. :thumb:
In Oz all our bicycles are setup moto style RH front brake - makes for amusing scenes when an American or European hops on them for the first time!

I don't think many moto trials bikes will be hitting 12k rpm. They have seriously choked down carbys (a 300 runs the same carby throat as many 65-80cc motocross bikes), big flywheels and are built for low down torque - I'd guess that mine's never been past 6K absolute tops.

If all this thread achieves is that it got you out on your trials bike, I consider it a winner!

Altair   1 kW

1 kW
Posts: 318
Joined: May 03 2014 10:35am
Location: Quebec, Canada

Re: Observed trials bike designing

Post by Altair » Feb 05 2020 6:49pm

Very interesting thread, guys! My post here will be a little off-topic from real trials bikes, but it will surely interest many of those who would like a bike suitable for "light-trials" riding.
I've ridden trials bikes since I was 18, and sold my last one about 5 years ago when I passed the sixties. So at this "advanced" age :wink: I needed something that was lighter, smoother and quieter. I had bought a fatbike, which was fun, but was badly lacking in the engine department...
That's when I decided to convert it to electric. Being used to ride on very rough terrain, I didn't want pedals in the way, so that helped define what type of bike I was going to build. I built a battery pack from 16 prismatic LiFePo4 flat cells, converted a MAC hub motor to a mid-drive, and bought an Adaptto mini-e controller. After a winter of riding this setup, the Mac turned-out to be underpowered. So I replaced it with a BHT motor and a direct chain drive of 8.7:1 and that helped tremendously. The bike was now a very good light-trials mount. I've been riding this bike for the last 4 years, all year long here in Canada. Here is what is looks like.
P1060658-1.jpg
P1060658-1.jpg (506.82 KiB) Viewed 1237 times
The performance is quite impressive for such a simple bike. Sure it doesn't have a suspension, but the light weight of about 57lbs/29.9Kg makes it very agile. With the BHT motor of 1500W, it climbs like a mountain goat. Note that I made footpegs mounts that keep the rider weight on the rear wheel. I think I had more fun on this bike than I had with my real trials bikes, even though I couldn't anymore do jumps and many things that require a suspension.

Altair   1 kW

1 kW
Posts: 318
Joined: May 03 2014 10:35am
Location: Quebec, Canada

Re: Observed trials bike designing

Post by Altair » Feb 05 2020 6:57pm

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.
P1070219-1.jpg
P1070219-1.jpg (450.38 KiB) Viewed 1235 times
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.
P1070215-1.jpg
P1070215-1.jpg (507.86 KiB) Viewed 1235 times
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.

Altair   1 kW

1 kW
Posts: 318
Joined: May 03 2014 10:35am
Location: Quebec, Canada

Re: Observed trials bike designing

Post by Altair » Feb 05 2020 6:58pm

Now I would like to address many points that have been asked in this thread, or elsewhere. Regarding the use of a clutch, I agree that it is indispensable for a real trials bike, either I.C. or electric. However on a fatbike, I don't think it's necessary at all, if the controller allows full rated current into the motor at zero RPM. The Adaptto wasn't able to do that, the current had a rising slope with RPM, which has always pissed me off to no end. I'm hoping that the Nucular will be OK in this regard.
There was also mention of the ability to creep forward very slowly without use of a clutch. In my experience, this is not a problem if your controller has a good control of the motor. For example, with the bike here I can get stuck going uphill and I can feather the throttle to keep some torque on the rear wheel and stay in place. I can even go backwards while keeping a small forward drive. What helps a lot is to set an exponential response for the throttle curve. It gives you a finer control of low throttle settings.
On the topic of simulating the engine rotating inertia, I agree that this is needed because when I chop off the throttle, the bike instantly has zero drive and is basically in neutral. This is useful in many cases but of course not always. What I had thought of doing was to set the throttle closing on a slow decay, and install micro switches on the brake levers; upon actuation they would instantly bring the throttle signal to ground (or to its lowest voltage), which would allow the simulation of engine inertia. But this doesn't work because the closing ramp is generated inside the controller, so not modifyable on the fly. An alternative would be possible by using a resistive throttle (with a potentiometer instead of a Hall effect sensor) and adding a diode/capacitor/resistor at the output so the signal is always ramped-down. The controller would be set on instant-off throttle response. Then the brake switch idea would work. (The switches might need to be adjusted so they engage only when the brakes are applied strongly, so as not to interfere when riding casually.)
On the throttle modes, I absolutely prefer the torque mode because it is the one that gives the best control of the motor, regardless of the RPM. Each position of the throttle handle always commands the same torque, so you get used to it very quickly, and it doesn't vary with the situation. Other modes are speed-dependant. Lastly, Regen. I never use it because it prevents you from coasting, decelerates the light bike too much, and basically gets in the way of smooth riding. Also when you're getting down a very steep & rough bank, from time to time you need to let the bike go foward a bit by itself, just to regain your equilibrium, in a case like that you don't want interference from anything, other than gravity...

bikerpete   1 W

1 W
Posts: 58
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Feb 06 2020 12:23am

Altair wrote:
Feb 05 2020 6:58pm
Now I would like to address many points that have been asked in this thread, or elsewhere. Regarding the use of a clutch, I agree that it is indispensable for a real trials bike, either I.C. or electric. However on a fatbike, I don't think it's necessary at all, if the controller allows full rated current into the motor at zero RPM. The Adaptto wasn't able to do that, the current had a rising slope with RPM, which has always pissed me off to no end. I'm hoping that the Nucular will be OK in this regard.
There was also mention of the ability to creep forward very slowly without use of a clutch. In my experience, this is not a problem if your controller has a good control of the motor. For example, with the bike here I can get stuck going uphill and I can feather the throttle to keep some torque on the rear wheel and stay in place. I can even go backwards while keeping a small forward drive. What helps a lot is to set an exponential response for the throttle curve. It gives you a finer control of low throttle settings.
On the topic of simulating the engine rotating inertia, I agree that this is needed because when I chop off the throttle, the bike instantly has zero drive and is basically in neutral. This is useful in many cases but of course not always. What I had thought of doing was to set the throttle closing on a slow decay, and install micro switches on the brake levers; upon actuation they would instantly bring the throttle signal to ground (or to its lowest voltage), which would allow the simulation of engine inertia. But this doesn't work because the closing ramp is generated inside the controller, so not modifyable on the fly. An alternative would be possible by using a resistive throttle (with a potentiometer instead of a Hall effect sensor) and adding a diode/capacitor/resistor at the output so the signal is always ramped-down. The controller would be set on instant-off throttle response. Then the brake switch idea would work. (The switches might need to be adjusted so they engage only when the brakes are applied strongly, so as not to interfere when riding casually.)
On the throttle modes, I absolutely prefer the torque mode because it is the one that gives the best control of the motor, regardless of the RPM. Each position of the throttle handle always commands the same torque, so you get used to it very quickly, and it doesn't vary with the situation. Other modes are speed-dependant. Lastly, Regen. I never use it because it prevents you from coasting, decelerates the light bike too much, and basically gets in the way of smooth riding. Also when you're getting down a very steep & rough bank, from time to time you need to let the bike go foward a bit by itself, just to regain your equilibrium, in a case like that you don't want interference from anything, other than gravity...
Thank you so much for joining this thread, a wealth of knowledge gained from experience!
Your current build looks fabulous! Sooo tempting to go down that path and just keep the Beta 300 for what it does well.

Not sure where to start with my questions really, so in no particular order:
1. You've moved to a Kelly KLS, but mention the Nucular - are you thinking the Nucular might be a better option, or was that just because I'd mentioned it as an option earlier? Does the Kelly also suffer from a similar problem you experienced with the Adaptto? Most recently I've been vacillating between a Kelly for the Torque/Speed mode and the Nucular (that also has that mode, but without any ability to tune it, and some bugs that are waiting to be ironed out), now it sounds like the torque/speed mode might not be of any benefit.
2. Torque mode - it sounds like you've spent time with the other modes, and what you say makes sense to me. My concern with torque mode is that if the wheel slips then the motor will accelerate pretty wildly until you can get off the throttle. An ICE bike does that too, but with their lower torque and more inertia it's possibly not as dramatic.
3. What turn count is the Astroflight out of interest? It looks a lovely light & compact motor, but I'm not really surprised it gets hot - light, compact aero style motors just aren't built for being abused at low rpm/high current I think. It's such a hard juggle when you really try to minimise weight!
4. Exponential throttle curve - this is something I've been thinking I'd like, but I haven't worked out how to achieve it with most controllers. Nearly everything I've looked into just has a setting for rise/fall rate and limits, but nothing I've seen that makes the throttle non-linear except the ASI I think it was. There's some talk about the Nucular getting throttle curves, but if it does't exist now I'm not going to assume it will ever come into existence. Perhaps this has to be done in the throttle itself. I'd quite like to keep a standard Domino trials throttle for it's quick action and run a cable to a throttle box (they seem hard to find). I've seen people adjust the cable groove in the Domino so it's a spiral to give them a non-linear pull. That might be a simple solution.
5. Rear sprocket carrier - how did you fix the sprocket to the hub? Is the carrier threaded on where the freewheel usually goes?

I like the thinking around modifying the throttle to get ramp-down and adding brake switches. Maybe a pressure transducer/switch in the brake line so it doesn't vary with pad wear etc.?
Regen - thanks for confirming that it's not desirable. That's what I figured. I was planning having a separate thumb throttle for regen so I could still access it on long descents.

More thinking.

Ohbse   10 kW

10 kW
Posts: 876
Joined: Dec 26 2013 5:15pm
Location: Auckland, New Zealand

Re: Observed trials bike designing

Post by Ohbse » Feb 06 2020 1:18am

It's impossible to replicate the functionality of a clutch, having the ability to dump that stored energy near instantly is pretty killer - but I think you could get pretty close with enough of a reduction ratio.

I've got a Nucular 24f, I think it's absolutely fantastic. Votol/Kelly/etc are absolute trash in comparison. Specifically its ability to 500 phase amps into a dead stalled motor reliably is difficult to replicate in any other controller that I'm aware of. You can precisely configure the ramp rate of current increase. By using a brake cutoff switch in place of a clutch, you can pre-position your throttle, release your brake and instantly burst power. It's how I get the front wheel on my Surron up - works a treat. My bike is capable of ~110km/h on 16s. Were it geared down to half that, it would wheelie on command with ease.

If you've got some detailed motor specifications on the QS you can determine what 500 phase amps is going to give you, customise a sprocket configuration to give you the wheel torque required. With the ability to field weaken quite substantially above base speed of the motor, you'll still be able to go a reasonable speed if you need to move around.

Regarding batteries - I think you're way up the wrong tree looking at 18650's. You're looking for power density and instantaneous response, range not a factor. You should really be looking for power dense Lipo packs - you can buy these from the Aussie Hobbyking warehouse, so supply shouldn't be an issue. Voltage is up to you, anything from 16-20s would be perfect when combined with the Nucular 24f. If you're somewhat close to a support vehicle, building a couple of smaller packs and swapping between them would give you a decent weight saving relative to filling it with enough 18650's to get your desired power output.

Post Reply