Trials Motorcycle EV Conversion

Sine

1 mW
Joined
Mar 10, 2019
Messages
16
I'm planning on doing an EV conversion on a 1998 Gas Gas TX 320 Trials motorcycle. I'll be using the factory transmission, clutch and crank. The plan is to mount the motor in the area where the ICE top end would normally go and couple the motor to the crank with a sprocket/chain drive set or gears running a ~1:2 or ~1:3 ratio to mimic similar effective redline to the factory 2 stroke and mechanically widen/dampen torque application. I'll be removing the factory flywheel and crank connecting rod, then filling the crank weight area with lead to act as a flywheel to free up space on the side of the motor.

I'm liking this kit from thunderstruck due to the sealed design and peak output, but also feel it might be a bit too much for the application. The peak watt output is similar to the factory gas motors in modern bikes but realistically the upper end of the power ranges are not much used in trials except for sections that have more enduro-like emphasis.

https://www.thunderstruck-ev.com/curtis-brushless-16kw.html?fbclid=IwAR33lTOXdKRAArPlxOxL55m2CCCmgLvhFRauQPQDQr1oVvUq8UtDNWRcxew

I've kludged together a few traction electric applications before, but the Curtis controller seems to be a bit daunting.

Some of the concerns I have:

- I want to be able to have an on demand 'idle' RPM for the gyroscopic balance benefits at low-no ground speed and the fact that the crank flywheel will already have energy stored. Revs and clutch dumps are common in trials riding. I'm not sure if this is a possibility with programming of the controller? Can I fake this with a switch activated resistor on the 0-5k throttle? Is there a way for it to accept a tachometer input and have it self adjust to meet the target at 0 throttle input?

- ICE Trials bikes have a very early Torque peak that falls off rapidly as RPM increases, this works with the flywheel. The smoothness of throttle on/off or partial application is going to be crucial for grip. I'm assuming these controllers can be programmed to have non-linear torque/power curve relative to throttle position? Something similar to modern drive by wire systems that

- I'm in IT and have done my share of programming in the past, but no idea how any of this works for Curtis controllers. Do I need an OEM level programmer for this? How difficult is this for someone who has intermediate skills with java and python?

- (in the far far far distant future) is it possible to have another potentiometer that measures swingarm position and alters the power map based on these input with the tach to act like a traction control system?
 
Sine said:
- I want to be able to have an on demand 'idle' RPM for the gyroscopic balance benefits at low-no ground speed and the fact that the crank flywheel will already have energy stored. Revs and clutch dumps are common in trials riding. I'm not sure if this is a possibility with programming of the controller? Can I fake this with a switch activated resistor on the 0-5k throttle? Is there a way for it to accept a tachometer input and have it self adjust to meet the target at 0 throttle input?
I doubt the Curtis has those functions (you'd have to check with Curtis directly), but you could fake it with the switched resistor network. (or other voltage input).

You may also be able to use the Cycle ANalyst from Grin Tech http://ebikes.ca to do this sort of thing, by using it's speedo input as the tach input, and it's throttle setup and speed limiting functions. Not sure of the exact setting syou'd hvae to use, but if you check with Grin they may be able to help you work out if it would owrk and exsactly what to set.


- ICE Trials bikes have a very early Torque peak that falls off rapidly as RPM increases, this works with the flywheel. The smoothness of throttle on/off or partial application is going to be crucial for grip. I'm assuming these controllers can be programmed to have non-linear torque/power curve relative to throttle position? Something similar to modern drive by wire systems that
The Cycle Analyst can be used to translate a throttle input to a torque (current) control, rather than the more common "speed" (PWM) control most controllers use. Basically what it does is monitor battery current, and throtlte input voltage, and modulate throttle output based on those two.

YOu amy also be able to use the torque sensor input (intended for pedal assist sensing) as a voltage input to modulate power as well.

- I'm in IT and have done my share of programming in the past, but no idea how any of this works for Curtis controllers. Do I need an OEM level programmer for this? How difficult is this for someone who has intermediate skills with java and python?
If the Curtis is programmable at all, they should have software that just conects via serial. If they don't or won't provide that, you likely have no way to program it, as you would have no way of knowing what values to send, syntax, bit count, baud rate, etc. There are no standards for this stuff, and even within a brand of controllers they aren't all inter-compatible.

- (in the far far far distant future) is it possible to have another potentiometer that measures swingarm position and alters the power map based on these input with the tach to act like a traction control system?
YOu could use the Cyclce analyst's analog aux input to change which preset(s) are used, or to change current limiting or power limiting.

Or use the CA's temperature sensor input similarly.
 
I really like the functionality of the CA. It seems like it does a lot of the things the Curtis can do internally, but overall just easier to setup and manage. The Curtis itself has its own programming language, my understanding is that almost every aspect of its function can be software defined, but this seems a bit laborious. The Curtis controller can operate in Torque or Speed mode as well.

Might include the CA just for the display and speedo alone even if it's not ultimately incorporated for motor control.
 
You might find that is a bit too much motor. 100a continuous is a lot of power. You can likely use something smaller since trials only has short peak power requirements.
26lbs relatively high in the bike, plus the room you lose for batteries will be noticeable.
 
Grantmac said:
You might find that is a bit too much motor. 100a continuous is a lot of power. You can likely use something smaller since trials only has short peak power requirements.
26lbs relatively high in the bike, plus the room you lose for batteries will be noticeable.

I agree, the problem is they don't seem to make sealed motors smaller and outdoor trials is a messy sport. I think i might just have to make some shrouding and air filters (and hope to not cook the motor) so I can use a smaller unit.
 
Grantmac said:
You might look at the QS mid drives.

Those are cool! But I would think the unsprung weight increase would have profound impacts on the ability to jump the bike. Sprung vehicle weight + rider weight is used to compress the suspension, unsprung weight would absorb energy when the suspension is unloaded and make it harder to jump.

Also, I need to be able to maintain the clutch. Clutch is used to store energy in the flywheel (or rotor mass) so it can be dumped to launch the bike from low speeds. It's also used to prevent any remaining inertia in the powertrain from forcing the vehicle forward after after landing so the brake can be applied more precisely.
 
I don't mean the swingarm unit, they sell just the motor and also the controller separately.

I'm looking at using the 1kw for a bicycle and it will easily take 4kw efficiently. The 3kw is known to be happy at 15kw.
 
Grantmac said:
I don't mean the swingarm unit, they sell just the motor and also the controller separately.

I'm looking at using the 1kw for a bicycle and it will easily take 4kw efficiently. The 3kw is known to be happy at 15kw.

Hmmm, depending on the course the motor behavior will be barely moving at all, but a lot of rev ups to clutch dumps, to scaling small hills over a long period. What's the expected duty cycle on that kind of motor at 15kw?
 
SlowCo said:
Look at this conversion from jonnydrive where he uses the QS mid drive motor with a chain to the gearbox:
https://endless-sphere.com/forums/viewtopic.php?f=28&t=94666&start=125

You might also want to check out Eastgem and their kits. The D7500 would be a great fit at a very nice price: http://www.eastgem.net/d7500-k-1.html
More products:
http://www.eastgem.net/accsesories.html

I'm now considering this kit.

https://www.electricmotorsport.com/pmac-g8018-72-84v-180a-electric-motorcycle-drive-system.html

Obviously substantially more expensive than the one you posted. However, I think the primary differences will be Sine/Cosine encoding. I think this will be more appropriate for the application as low-no speed operation dominates so low RPM smoothness is a huge boon. Not sure if the one you supported supports regen, but having something akin to engine braking is also ideal.

Not sure if there's any difference in build quality etc. otherwise. Though, if the differences won't be too severe I can be talked down into wallet saving options.
 
The motor in the kit you linked to is only 4.5kW (6hp) continuous rated. Then the Eastgem d9000 rated at twice (9kW) the continuous power seems a better choice: http://www.eastgem.net/d9000.html
Not sure how much import taxes you would have to pay though sent from China?

For a trails motorcycle throttle response seems to be the most important factor. So get a good programmable controller preferably with Bluetooth connectivity for changes during a ride (when stopped...) and a serious throttle like a Magura or Domino.
 
SlowCo said:
The motor in the kit you linked to is only 4.5kW (6hp) continuous rated. Then the Eastgem d9000 rated at twice (9kW) the continuous power seems a better choice: http://www.eastgem.net/d9000.html
Not sure how much import taxes you would have to pay though sent from China?

For a trails motorcycle throttle response seems to be the most important factor. So get a good programmable controller preferably with Bluetooth connectivity for changes during a ride (when stopped...) and a serious throttle like a Magura or Domino.

Specs are sparse on the 9000, the length isn't listed but it looks longer than I have to fit in my frame without protruding. It also says it takes a square wave controller and uses hall effect for sensing; my understanding is that this means low speed throttle will be 'jittery.' I'm planning on a heavy flywheel so it may mechanically dampen some of this effect, but unsure how perceptible it will be in practice.
 
If you are going to have the motor "idle" at a certain RPM then low speed commutation is far less important.
Getting accurate throttle control would be a higher priority for me.

I'd also not worry about continous ratings in a trials environment. How often are you pinned at full throttle, under load for more than a minute per section.
 
Grantmac said:
If you are going to have the motor "idle" at a certain RPM then low speed commutation is far less important.
Getting accurate throttle control would be a higher priority for me.

I'd also not worry about continous ratings in a trials environment. How often are you pinned at full throttle, under load for more than a minute per section.

I think the Sine/Cosine encoding and Sinusoidal control would still be perceptible at lower mid-upper throttle ranges wouldn't it? Idling with the clutch in I can see my heavy flywheel making things imperceptible, much int he way the flywheel dampens the power pulse acceleration/deceleration in ICE engines. However, clutch out at low RPM the flywheel does less I would I think.

In a section, never. Outside a section in happy-fun freestyle land, frequently. =D
 
SlowCo said:
Length is in the text: 170mm motor + 35mm axle = 205mm (just over 8 inch).

Yeah, might be a little too chunky for the application. The frame itself is 9 inches wide at its widest point (not the handlebars). The motor will need to be recessed in the frame as falling over is part of the game.
 
From what I've heard the QS motors have better quality than the Danzel. I've looked at both for my next build and the plan is QS. Also I believe their controller is FOC with torque control.

Designing around a 100% duty cycle at 15kw is going to take more motor than you need IMHO. The Emotion trials is 6kw/12kw burst with a single speed.
 
Grantmac said:
From what I've heard the QS motors have better quality than the Danzel. I've looked at both for my next build and the plan is QS. Also I believe their controller is FOC with torque control.

Designing around a 100% duty cycle at 15kw is going to take more motor than you need IMHO. The Emotion trials is 6kw/12kw burst with a single speed.

I'd like to ditch the factory motor entirely, I think there's just too much value in a clutch and flywheel; I don't think I'd lose too much ditching the gearbox itself. I'd love a CVT Transmission with a grip-shifter though, but I'd have to make that myself (would be a long term goal).

Also, I reconsidered the kit in the OP, I posted a link to another kit with 6hp continuous and 19hp peak that I'm leaning more towards that has sine/cosine encoding PMAC motor. It's still a bit huge for the application using the factory trans, but the output/layout will be more similar to a competitive commercial e-trials example bikes like the Gas Gas TXE and Yamaha TY-E, whom both make use of clutches, 6 speed gearboxes and flywheels.
 
I just got the D9000 in the mail and the thing is huge! I would have to agree that there are much better designed and more powerful choices out there for its size, but for the price and how quick Alex at Eastgem ships this stuff to you, I just have to try the stuff out

package said from china 8kg pretty beat up but motor looks ok

im planning a full size Honda 250 dirt bike build probably this winter
 

Attachments

  • IMG_5406.jpg
    IMG_5406.jpg
    321.9 KB · Views: 958
  • IMG_5405.jpg
    IMG_5405.jpg
    277.7 KB · Views: 958
  • IMG_5404.jpg
    IMG_5404.jpg
    319.1 KB · Views: 958
skeetab5780 said:
I just got the D9000 in the mail and the thing is huge!

im planning a full size Honda 250 dirt bike build probably this winter

Serious lump of motor :thumb:
What controller are you going to use on it?
 
SlowCo said:
skeetab5780 said:
I just got the D9000 in the mail and the thing is huge!

im planning a full size Honda 250 dirt bike build probably this winter

Serious lump of motor :thumb:
What controller are you going to use on it?

did not mean to high-jack the thread, just no plans to start my own yet until winter and wanted people to see the new Denzel potted 9000 motor

I think i'm going to try the Votol-100 or 150
 
SlowCo said:
skeetab5780 said:
I think i'm going to try the Votol-100 or 150

Doesn't Eastgem/Denzel offer a plug&play controller for their motor?

Yes but they do not offer field weakening to go faster than your voltage which is becoming pretty important these days, not for a trials bike so it may not pertain to all builds
 
Back
Top