Observed trials bike designing

General Discussion about large electric scooters and motorcycles and other things with no pedals.
User avatar
DanGT86   100 kW

100 kW
Posts: 1051
Joined: Sep 06 2012 9:18pm
Location: Saint Louis MO

Re: Observed trials bike designing

Post by DanGT86 » Feb 15 2020 8:49am

bikerpete wrote:
Feb 15 2020 12:13am

3. That's true and is what I think I'll need to do. To retain a big rear sprocket to keep chain forces down I'll need to run a big front sprocket - eg to get a 15:1 overall reduction with the clutch primary reduction in the train, with a 90t rear I'd need a 21 tooth front. Looks like I might be making my own sprocket!

Buller makes 219 sprockets all the way up to 25tooth. Might be useful for you.
https://www.buller.net/Category/bully_misc

I have a 22 tooth of theirs I converted to a 3/4" shaft mount with keyway. Buller sprockets are hardened so it took some careful machining but the ID bore is made for a needle roller bearing so its large enough to press fit an adapter sleeve to your motor shaft if you cant cut it. They also have a snap-ring interface to "quick" change them onto their clutches. You could possibly duplicate that design and use the sprocket as is providing you an easy way to change them on the fly.

They also make some awesome multi disk centrifugal clutches. I have often thought about making a device to actuate one by cable. Seems doable but I'm not sure if they would be durable enough.

bikerpete   10 W

10 W
Posts: 92
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Feb 15 2020 7:34pm

DanGT86 wrote:
Feb 15 2020 8:49am
Buller makes 219 sprockets all the way up to 25tooth. Might be useful for you.
https://www.buller.net/Category/bully_misc
That's fabulous, thank you.
Shame they don't do their ring gear series in 219, that would be so easy to make an hub to suit the motor and swap around sprockets whenever. #35 chain is very little used in Oz it seems, so 219 is definitely a better option for me.
It looks like they have a 1" bore, does that sound right?

Thanks for confirming my speed estimates Guy. Makes me more comfortable that I'm working with some reasonable assumptions.

Still struggling to make my final choice of motor from a less than ideal set of options it seems.
I did spend quite some time looking around at less common alternatives, but either they look expensive or I'd be stepping into the great unknown - probably not a good place to go for a first attempt and expect a good result. Found a couple of really nice small axial flux motors - looked expensive but maybe I should ask, you never know. But then I've no idea what controller they might need either. EDIT Got a quote for a nice little 7kg axial flux motor EU4,100. :lol: :lol: :lol:

I should have the clutch in my hands tonight, which will answer a lot of questions I have about how I might fit things together.

User avatar
DanGT86   100 kW

100 kW
Posts: 1051
Joined: Sep 06 2012 9:18pm
Location: Saint Louis MO

Re: Observed trials bike designing

Post by DanGT86 » Feb 15 2020 10:44pm

#35 chain does not have rollers only bushings. I don't know why it is the standout in that regard but I don't want to spend money to find out. With the type of high reduction ratios we all end up with on these bikes 219 is a nice option. I believe that bore on the 22 tooth sprocket I got from Buller was around 1". It had a needle roller pressed into it and 2 drive tangs with a snap ring. I imagine the clutch its made to mount to has the same 2 drive tangs phased 90deg out and the same snap ring groove. It looks like it uses the OD of the needle bearing to align them and the snap ring holds everything together axially.

I do all my speed calcs with this handy RC motor Drive calculator. Its pretty handy.
http://www.recumbents.com/wisil/e-bent/ ... ulator.asp

I know your motor options don't sound ideal to you but I think they are better than you are giving them credit for. I saw another post where you were asking about motors. There are lots of motors that look the same. You can get a similar looking product from cyclone, LMX, lightning rods and alibaba. The difference is that the vendors like LR and LMX who sell this stuff go to the same factories in china that the alibaba stuff comes from and demand particular specs. It might not be easy to see from the description but often times its better bearings or an actually useful output shaft that makes the difference. Sometimes its the stator lamination thickness. So don't go to alibaba and roll the dice when the vendors like LMX and Lightnig rods have cut through the crap for you. It's well worth a few extra bucks to get quality control and an easy to use product compared to 5 weeks shipping on a bargain china mystery box. The last thing you want on your first build is a bad bearing or poorly positioned hall sensor giving you headaches. You'll have enough to figure out even if all the parts work great right out of the box.

As for the axial flux motors in the size and power you are looking for, there are the units from Motenergy that seem to have decent specs but they don't seem to be really common here. I've eyeballed this one a few times for a full moto size trail bike build.
https://www.thunderstruck-ev.com/me1718.html

The problem for me is that most of these have sine/cosine encoders instead of hall sensors so they are commonly paired with big money and hard to program controllers. Ever since the Nucular and the QS3000 hit the market there just isn't a whole lot of reasons for me to want anything else for a build under 20kw. The QS mid drive motors seem to exceed expectations and are so cheap that I could almost see running 2 of them instead of 1 large fancy motor.

bikerpete   10 W

10 W
Posts: 92
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Feb 16 2020 2:08am

Thanks again Dan. I love the wealth of knowledge available on these types of forums.
Can't imagine wading through the single stream of posts looking for similar information on the evil facebook!

The Nucular/QS 3000 was certainly looking like the combo to go for in a full size trials bike - no question it seems.

I've definitely been erring toward a known supplier rather than unknown china suppliers. Of course a known good china supplier would be fine too.

It seems so hard to find the exact details of even some of the better regarded supplier motors. eg. I'm talking to LMX about their motors trying to find out the kv - they don't list it and I know they have had the same basic motor in both 37 & 70kv. So what will I get if I order now? Hopefully get an answer to that one tomorrow.

BHT are another supplier which has had a good "big block" for this in the past (37kv), but it disappeared off their list then seemed to reappear more recently, but I don't know if it's the same 37kv one. Trying to find that out. They have a "48-60V 1200W 0.35mm Lamination 60mm magnets" that looks like the older smooth fronted LMX motors but now with a 10mm shaft, and a "48-72V 500-3000W" with a foot on it which is I think the one that came and went.
Very confusing for a newbie!

I'll get there. I think if any of these 3 are 37kv +/- then I'll grab it and be done with trying to decide. They all seem to deliver the sort of power I'm after and aren't too much different in size & weight (if I've got it right).

bikerpete   10 W

10 W
Posts: 92
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Feb 16 2020 6:40am

Well I heard back from BHT bike. "Hi, I haven't stock. All China in stand-by, not any factory can work, virus"

Then my clutch arrived, I measured the drive pinion and guess what - 14mm dia, 4mm keyway. And what's the LMX shaft you ask? Oh 14mm dia, 4mm keyway.

Well that settled it.

So I've now ordered a 2019 model LMX 30mm motor on closeout, and the price just today dropped from 105EU to 98EU. Bonus. Obviously meant to be.

I've started a build thread Fat-e Trials Bike.

bikerpete   10 W

10 W
Posts: 92
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Feb 18 2020 1:57am

DanGT86 wrote:
Feb 15 2020 8:49am
Buller makes 219 sprockets all the way up to 25tooth. Might be useful for you.
https://www.buller.net/Category/bully_misc
I just went to buy a single 22 Tooth sprocket from Buller - US$156.51 shipping UPS for a single sprocket seemed a little rich for me!

I'll wait and see what else I might gainfully buy from the USA then perhaps consolidate it with a shipping agency. Or maybe if I email them they might post it at a more reasonable cost perhaps.

Never ceases to amaze me how much it costs to get stuff out of the USA to Australia.

EDIT: I got a quote for US Post - still US$40 for single $21 sprocket. Better, but I'll probably just laser cut one from Hardox wear plate and polish it up.
Last edited by bikerpete on Feb 22 2020 11:35pm, edited 1 time in total.

User avatar
stan.distortion   100 W

100 W
Posts: 176
Joined: Jun 12 2019 12:53pm

Re: Observed trials bike designing

Post by stan.distortion » Feb 18 2020 7:12am

It could be worth checking out bearing and seal suppliers in your local area, they usually stock belts, pulleys, chain, sprockets, etc. and can often either put you in touch with someone doing basic engineering work for the right price or offer that service themselves. Prices are often surprisingly low, not as low as you'll see on ebay but rarely silly high, the ones that didn't stay competitive are mostly gone out of business.
Projects indirectly sponsored by https://www.dash.org/

User avatar
DanGT86   100 kW

100 kW
Posts: 1051
Joined: Sep 06 2012 9:18pm
Location: Saint Louis MO

Re: Observed trials bike designing

Post by DanGT86 » Feb 18 2020 10:07am

We should start a section on the forum where people post small item shipping requests. There has to be somebody flying from the US to AUS in the next few weeks. A sprocket in a carry on bag wouldn't be a big deal.

User avatar
amberwolf   100 GW

100 GW
Posts: 32163
Joined: Aug 17 2009 6:43am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: Observed trials bike designing

Post by amberwolf » Feb 18 2020 11:59pm

could use items wanted for now

bikerpete   10 W

10 W
Posts: 92
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Feb 22 2020 11:55pm

After a fair bit of searching and reading I've finally found the likely power supply for charging through the Nucular.

A Cisco WS-CAC-3000W AC Power Supply Catalyst 6500.
s-l1600.jpg
s-l1600.jpg (31.28 KiB) Viewed 1876 times
These deliver:
25.0 A @ +3.3 V
5 A @ +5 V
12 A @ +12 V
65.98 A @ +42 V
On the front panel they have a nice on/off switch, an IEC input socket and a Molex that delivers 42V 17A. Main power out on the rear. Don't need any jiggery pokery to get running, nicely enclosed ... Plug and Play 3000W. Should give some headroom for further battery growth!

42V should be about right for charging my 16S pack even from absolute worst case LV, leaving a few volts headroom for the Nucular to do it's stuff.
Oh, and they're CHEAP. Listed down to AU$25 on ebay!

Hopefully it does the trick.

john61ct   100 GW

100 GW
Posts: 7480
Joined: Dec 18 2018 2:06pm

Re: Observed trials bike designing

Post by john61ct » Feb 23 2020 3:10am

Yes excellent value, but not compact not light not quiet.

Need 220 Vac to get 3kW, 110V is 1400W, both are at 16A input.

Seen $12 for good used, $30 new open box.

Don't pay over $80 shipped

_______
Big sister WS-CAC-8700W
has three inputs can be separate circuits

total input determines output power

**will not power up with 110V single input**

2800W from @ 2x 110Vac, or 1x 220Vac

______
Middle sis 6000W version

dual inputs, 16A each, also
**will not power up with 110V single input**

6000W if both inputs are 220 VAC

2900W if both inputs are active but at least 1 is at 110 VAC or a single input is active and at 220 VAC

_______
red-headed stepchild WS-CAC-4000W

flat-out requires 240V

bikerpete   10 W

10 W
Posts: 92
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Feb 23 2020 4:04am

john61ct wrote:
Feb 23 2020 3:10am
Yes excellent value, but not compact not light not quiet.
Need 220 Vac to get 3kW, 110V is 1400W, both are at 16A input.
Seen $12 for good used, $30 new open box.
Don't pay over $80 shipped
I've got a big shed, away from the house with 15amp 240V outlets - don't care if this one's big, heavy and loud.
I might find a small, light and quiet something down the track. Oh and if I come close to pulling 3000w then something is going seriously wrong and it's time to call the fire brigade!

They'll be US$ you're talking there aren't they?
Double it, take away the number you first thought of, square the result and add the product of that and your first two numbers and you get A$ :lol:

A$58 to my door. That's US$38.
Fingers crossed it does what I ask.

john61ct   100 GW

100 GW
Posts: 7480
Joined: Dec 18 2018 2:06pm

Re: Observed trials bike designing

Post by john61ct » Feb 23 2020 12:26pm

Still getting it for about the scrap metal value, amazing, bonza eh?

Altair   1 kW

1 kW
Posts: 330
Joined: May 03 2014 10:35am
Location: Quebec, Canada

Re: Observed trials bike designing

Post by Altair » Mar 28 2020 10:43am

I've decided to remove the AstroFlight motor and install the BHT that was in my first bike, in its place. That first bike has now been reconverted back to its former use, a standard pedal fatbike, and it saw good use in the last month, helping me regain some strength in the old legs.

The motor substitution will solve the overheating problem on the new bike, but the weight went up to 59lbs, from 54½ originally. New motor mounts had to be machined and I also raised the ground clearance a bit, to 12" under the skid plate. The top of the pegs is now also at 12".
P1070244r.jpg
P1070244r.jpg (481.22 KiB) Viewed 1347 times
P1070240r.jpg
P1070240r.jpg (492.22 KiB) Viewed 1347 times
The BHT, not needing a gearbox like the Astro, makes less noise, but the Kelly KBS-E controller adds a bit because it is a trapezoidal controller, while the Astro was running on a sinusoidal KLS-S. Anyway, the KBS-E in the pics is only temporary, awaiting for a Nucular.
Cheers!

SlowCo   1 MW

1 MW
Posts: 1703
Joined: Jan 05 2015 5:43pm
Location: The Netherlands

Re: Observed trials bike designing

Post by SlowCo » Mar 28 2020 2:08pm

That looks very well built!
For what top speed is it geared?

Altair   1 kW

1 kW
Posts: 330
Joined: May 03 2014 10:35am
Location: Quebec, Canada

Re: Observed trials bike designing

Post by Altair » Mar 28 2020 3:32pm

Around 40Km/H maximum. The gearing is 96/11= 8.72:1
I wanted it to climb like a goat :)

john61ct   100 GW

100 GW
Posts: 7480
Joined: Dec 18 2018 2:06pm

Re: Observed trials bike designing

Post by john61ct » Mar 28 2020 7:02pm

lust. . .

bikerpete   10 W

10 W
Posts: 92
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Mar 29 2020 12:53am

That is, as usual, very nicely done!

I'm a little envious of your access to CNC machinery (or considerably superior machining skills). Mine are all manual, at least I have a DRO I move between the lathe and mill.

gbeals   1 W

1 W
Posts: 57
Joined: Apr 15 2020 11:43pm
Location: Honolulu, Hawaii
Contact:

Re: Observed trials bike designing

Post by gbeals » Apr 25 2020 2:45am

Hello Gents - I hope you don't mind me jumping into this thread. I'm behind you in progress on the Trials bike, but very deep into my research. I had the pleasure of demoing the 2020 Electric Motion bike here in the Jungle's of Hawaii. (My back yard) I rode both the escape and the e-pure Race Trials bike. I first will say the thing is a BEAST! BEAST!!!! I will say.... I'm an experienced motorcycle rider, super bikes, harley, bobbers, dirt, supermoto - but new to Trials bike. So I can't come close to a trials rider but I have something to offer this conversation regarding the electric vs hydraulic clutch.

I TRIED to use it and then gave up quite quickly. It is definitely not needed and felt silly from a typical stop. Rolling on the throttle lifted the bike in the lightest map (there are 3 maps). I then thought let's try to dump the clutch some and see what happens - Like in trials. It made a cool noise and then BOOM - bike flew up as I expected it to. It's light and I'm no trials rider, so I'd expect you could do some pretty sizable hops with all that torque. My next test - Crack the throttle violently. Almost felt the same. The power delivery was faster of course when you dump - I enjoyed the electronic clutch just as much - but a little more as it served a dual purpose. One setting was for Electronic clutch the other was for regeneration. I like the regen on steep hills - used as engine braking and generated plenty of recharge.

MAPS - I cranked the bike up the the 3rd - supposed to simulate a 300cc bike. YUP. I think so. It hit the mark for me. I did find it was just too much power in the woods. Maybe for a PRO Trials rider, but not for me. I found 1 and 2 maps climbed anything i tossed at it, and was the most responsive (sometime too responsive) map. I design applications and mobile apps so looking forward to diving into the controller and seeing what I can do there.

I would buy one - however I know how much they cost. I am a fabricator and love to build things and can do this at a fraction of the costs. So I will build my own. I dug into the bike a little and can see the bulk of the parts and you folks have already mentioned them within this thread. I know the 2019 EM used 5KW BLDC Motor - HPM5000B from Goldenmotor- I think the 2020 uses the same. I'm not sure if it is liquid or air cooled - but I will find out soon.

I have already learned quite a bit from reading this thread and look forward to more. I'm new to electricity so that is where my learning curve is. I haven't quite figured out all the math; which is why I started to figure out what EM has, and just go from there.

Image

bikerpete   10 W

10 W
Posts: 92
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Apr 25 2020 6:15am

gbeals wrote:
Apr 25 2020 2:45am
Hello Gents - I hope you don't mind me jumping into this thread. I'm behind you in progress on the Trials bike, but very deep into my research. I had the pleasure of demoing the 2020 Electric Motion bike here in the Jungle's of Hawaii. (My back yard) I rode both the escape and the e-pure Race Trials bike. I first will say the thing is a BEAST! BEAST!!!! I will say.... I'm an experienced motorcycle rider, super bikes, harley, bobbers, dirt, supermoto - but new to Trials bike. So I can't come close to a trials rider but I have something to offer this conversation regarding the electric vs hydraulic clutch.

I TRIED to use it and then gave up quite quickly. It is definitely not needed and felt silly from a typical stop. Rolling on the throttle lifted the bike in the lightest map (there are 3 maps). I then thought let's try to dump the clutch some and see what happens - Like in trials. It made a cool noise and then BOOM - bike flew up as I expected it to. It's light and I'm no trials rider, so I'd expect you could do some pretty sizable hops with all that torque. My next test - Crack the throttle violently. Almost felt the same. The power delivery was faster of course when you dump - I enjoyed the electronic clutch just as much - but a little more as it served a dual purpose. One setting was for Electronic clutch the other was for regeneration. I like the regen on steep hills - used as engine braking and generated plenty of recharge.

MAPS - I cranked the bike up the the 3rd - supposed to simulate a 300cc bike. YUP. I think so. It hit the mark for me. I did find it was just too much power in the woods. Maybe for a PRO Trials rider, but not for me. I found 1 and 2 maps climbed anything i tossed at it, and was the most responsive (sometime too responsive) map. I design applications and mobile apps so looking forward to diving into the controller and seeing what I can do there.

I would buy one - however I know how much they cost. I am a fabricator and love to build things and can do this at a fraction of the costs. So I will build my own. I dug into the bike a little and can see the bulk of the parts and you folks have already mentioned them within this thread. I know the 2019 EM used 5KW BLDC Motor - HPM5000B from Goldenmotor- I think the 2020 uses the same. I'm not sure if it is liquid or air cooled - but I will find out soon.

I have already learned quite a bit from reading this thread and look forward to more. I'm new to electricity so that is where my learning curve is. I haven't quite figured out all the math; which is why I started to figure out what EM has, and just go from there.
The more the merrier!
The cost of the EM was what also sent me down this path, I then took the left fork down the Fat-E trials bike path.

I'll point you to the "Trials riders need their clutch?" thread, it might help shed some light on the pros and cons of clutches for trials. You can certainly ride trials and have plenty of fun without a clutch (see Altair's comments about having more fun on his clutchless fat bike than he possibly ever had on his clutch equipped ICE bikes), but I'm absolutely convinced there are very common trials maneuvers that you simply cannot really do without a clutch. Just depends if you want to do those things. If you can, get a bit of time on trials bikes (ideally with someone who knows how to ride trials) so you get clear what it is you want to do with it - single track weapon? Technique development? Trials comps? Mucking about silently in the back yard?

I'd also say that most motorcycle riders coming into trials soon discover that the skills they thought they had are considerably short of what's required for even quite modest trials. Be prepared for a pretty steep (but very rewarding) learning curve if you really want to ride trials.

As for the regen thing, that's no big deal to solve - I'm planning on fitting a left hand thumb throttle for regen braking. Just one way of tackling the problem.

I can't comment on the electric experience vs ICE as I'm still slowly working on my build and don't expect it to be ready for some months yet. I'd love to have a ride on the ePure but I've never seen one anywhere near my neck of the woods, only the earlier model.

I've been an industrial designer of over 30 years, and worked for a micro hydro company for some years developing turbines and aspects of one of their alternators, so learnt a bit about electrics. I've also worked a bit with PLC's and some very basic electronic circuits. Despite that the amount I've had to pickup for this project has been huge. Just the myriad little (and big) things like which connectors to use, how to reasonably safely construct and manage a LiPo battery, which cells to use, where to source a suitable power source for charging ... the list just goes on! Certainly drawing from the EM components list will reduce a lot of those issues. I expect your battery will be the most effort to sort through and get built.

I'd suggest buying a trials bike from about 2005 on, and grafting an electric into it. 125, 250, 300 - makes no difference, they all share the same geometry and components. All the current brands are good, each has it's strengths and weaknesses, but there aren't any duds and it'll probably be hundreds of hours before you out-ride any of them. That's a reasonably well trodden path that works pretty well from what I've seen. Some people keep the gearbox, others don't.

I'm not sure which controller the EM is using now, they used to use Kelly but I seem to recall the latest ones use a different controller? Kelly don't seem to be highly thought of here, but they obviously work well enough. Probably a good choice to go with whatever EM use, although I'd hazard a guess that reliable, production grade supply and support is possibly a bigger decider for them than absolute best performance. The Nucular controllers seem on balance to be the current cream of the crop for a variety of reasons, although you pay a signifcant premium over the more common Chinese controllers.

When I was looking at building an electrified moto trials bike it looked pretty hard to go past the QS motors for cost and performance - people are easily driving the 3kw mid motor to 15-18kw! How their low speed torque compares to the Golden is a bit of an unknown, to me at least. Torque is king on a trials bike, outright power is almost irrelevant. Much of what you'll read on ES forums talks about power and speed - from what I've learnt here, that's got only marginal relevance to what we need. A motor that's happy putting out big power at high speed can overheat in seconds when being asked to put out a load of torque close to stall. I know the GasGas electric uses a liquid cooled motor - can't recall for the EM or the Dragonfly.

I'd be interested to see the full component list for the EM if you put that together.

User avatar
DanGT86   100 kW

100 kW
Posts: 1051
Joined: Sep 06 2012 9:18pm
Location: Saint Louis MO

Re: Observed trials bike designing

Post by DanGT86 » Apr 25 2020 12:42pm

bikerpete wrote:
Apr 25 2020 6:15am
When I was looking at building an electrified moto trials bike it looked pretty hard to go past the QS motors for cost and performance - people are easily driving the 3kw mid motor to 15-18kw! How their low speed torque compares to the Golden is a bit of an unknown, to me at least. Torque is king on a trials bike, outright power is almost irrelevant. Much of what you'll read on ES forums talks about power and speed - from what I've learnt here, that's got only marginal relevance to what we need. A motor that's happy putting out big power at high speed can overheat in seconds when being asked to put out a load of torque close to stall. I know the GasGas electric uses a liquid cooled motor - can't recall for the EM or the Dragonfly.

I'd be interested to see the full component list for the EM if you put that together.
Following along with your thought process has helped me analyze my own lessons learned building EVs. I think you will see things a bit differently once you have logged a few miles/hours on your new ride. I think you are a bit stuck in the mindset of gas engines. Here is why I say that:

Torque and power:
Torque is important but it can be created through gearing. Power cant. Any of the electric motors you are comparing QS and Golden will have ample torque. Their torque curves are functionally flat. The reason people care about comparing torque on gas engines is that its a narrow window. Piston speed and air/exhaust flow create an ideal volumetric efficiency range. If you are out of that range it is crap. The clutch and gearing is used to keep the engine at peak torque. If you have more torque than load the rpm speeds up making more power torque x rpm. A trials gas engine is designed for low rpm torque since that's where you spend lots of time. An electric motor will provide torque abundantly from most rpms so you don't have to keep it in any range. Unless something is drastically wrong with the motor design you don't really need to worry about comparing torque from one brand of motor to the other. Its a product of diameter, copper fill, magnet strength. Unless a company is building garbage the torque of 2 motors with similar power and rpm will be similar enough to not really concern yourself with. Selecting one with the rpm/volt (kv) that matches the rpm you expect to spend most of your time in will give you the best efficiency.

Need for a clutch:
I think a clutch is needed for true aggressive trials riding but in my opinion its not why you think. There has been much discussion about the amount of energy stored in a flywheel during a clutch dump. On the clutch thread someone posted a graph of dyno data showing a very brief impulse of power from the clutch dump but it was a very small blip. Again with the gas engines narrow torque curve means you get junk power unless you can store enough energy in the flywheel such that your gas engines rpm wont drop out of power during the dump. When you clutch dump the crank speed would drop to nothing if you didn't have a moment of slippage that the motor can power through. This keeps the piston speed high enough to not upset the airflow and to get a good combustion event to make adequate torque to keep the motor running. Power from an ICE is discrete pulses that depend on smooth transitions of load. The clutch is a load smoothing device like a torque converter is in an automatic trans car. The energy stored in the flywheel itself is super tiny with regards to moving the bike. It is hugely beneficial to keeping the crank moving when you surprise the motor with instant load. Imagine if the energy stored in the flywheel was equal to someone throwing a brick at you. A 100mph brick shot at your bike would not shoot your bike up onto a rock. You'd barely move forward. It's a lot of energy and you would notice but its not going to launch you.

Possible clutch benefit to an EV:
Electric motors make a ton of torque everywhere. That being said torque x Zero rpm is still zero power. The lower the RPM the higher the current needed to achieve the same power. So on launch from a truly dead stop there is a practical limit to the amount of current most controllers can provide. I believe some small slippage from a clutch could be beneficial. It could be centrifugal or just an overload clutch based on torque but something to get you off zero rpm makes a big difference in the current required.

Clutch benefits for trials and aggressive riding:
Modulating a long pull lever at different angles and body position is way easier than a delicate twist throttle especially when the throttle is the entire grip you are holding. Trials and off road riders make constant use of the clutch as an instant power off switch as well as a way to load and unload the chassis and chain. I agree with you that its probably most helpful if not completely necessary to most trials riders especially those with years of muscle memory.

Ramp up time or instant impulse needed for trials moves:
This kinda ties all the above points together. Basically the piston speed, torque curve, and power band of an ICE all determine the time it takes to ramp up to higher rpm. The ramp up is nothing more than the throttle requesting more torque by upsetting the equilibrium of airflow in and power out. The fastest launch is provided by making the motor happy then using the clutch to grab onto as much of this power as you need. The electric motor's version of this is simply the time it takes for the current to energize the wires. This does not happen instantly but I would theorize that it does happen exponentially faster than a gas engine takes to increase power. The motor is an electromagnet that can energize from zero torque to max torque within a single revolution. A gas engine cant come close to this. If the piston is down and you open the throttle all the way it takes many combustion events for the torque to ramp up. I wish I had some data on how long it takes for current to magnetize wire. I'm sure someone on here has that data. We could calculate the similar change in power of a gas motor by back-figuring the tire size, gearing, transmission ratios, etc. I may start a thread about it over on the motor section because I am genuinely curious.

This is why I think your comparison of wheelies from full throttle vs wheelies from dumped clutch on gas bikes is not a fair comparison to electric vehicles. I am almost positive that an electric motor with no clutch will have more than enough torque to zap you onto any obstacle as well as possibly brake your frame or just overpower the tire. EV torque is insane and in most cases detuned to protect the drivetrain. The only caveat being the need to help the controller overcome the startup issues of a truly zero rpm motor. There is some non zero amount of time for it to properly sync.

I had an electric car with a clutch and 5 speed. Shifting was unexciting and dumping the clutch did almost no good at all other than terrible noise.

I'm not sure if you are familiar with recumpence's davinci drive. He uses a torque limiting clutch on the primary drive. His setups use RC motor controllers which are notorious for violent startup. Incorporating a torque limiting clutch drops the current required at startup. From a dead stop the controller looks at the motor almost like a dead short. Allowing just milliseconds of slip can lower the current required by 2 or 3x. If you already have infinity billion times more torque than you need then limiting it to a consistent value with a slipping clutch would give you a consistent max throttle launch every time as well as keep your motor and controller from overloading. Again this is all under my assumption that your motor and gearing provide more torque than you could possibly need. I believe a QS motor on a 150lb trials bike geared for 40mph would meet this requirement.

I hope I don't come off as a know it all. I am far from an expert and hopefully have called out where I am speculating and lack data. Like I said your planning on this build has made me dig into my understanding a bit and its helpful for me to try and articulate my thoughts as part of my own learning process. Hopefully we get some motor experts in here.

bikerpete   10 W

10 W
Posts: 92
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Apr 25 2020 7:51pm

DanGT86 wrote:
Apr 25 2020 12:42pm
Following along with your thought process has helped me analyze my own lessons learned building EVs. I think you will see things a bit differently once you have logged a few miles/hours on your new ride. I think you are a bit stuck in the mindset of gas engines. Here is why I say that:
That's a good read Dan, thanks.

I'm definitely aware that I'm making dangerous assumptions based on zero experience, but really the bottom line for me is that when I watch the riders on the clutchless electric trials bikes, including riders in the top handful in the world, they just can't do some of the things that a clutched ICE bike can. As soon as those same riders get on a clutched bike, be it electric or ICE then they can make these missing moves. And although I'm rubbish at them, these are the moves I really like playing around with.
Sure, it could be that there just isn't anyone who hasn't learnt their techinques on an ICE bike, so we just don't yet know how to best use an electric, I think that's only part of the story though.

I also don't think it's a difference in absolute power or torque that's significant, even little 125cc ICE bikes can do maneuvers I've never seen a clutchless electric do, it's the timing of the delivery. I know you've looked at the "Trials riders need their clutch" thread, so I wont re-iterate my thoughts here, but I really think the clutch is largely a means to enable the use of a big flywheel. Not to smooth out the engines power delivery (this is almost certainly why we started using big flywheels, and they're definitely required for that) but to smooth out the rider's ability to control it. As with so many machines, the weak link is the operator.

It's almost certain that a flexible enough controller could achieve the same or even a better result, but the software in that controller would be quite specific to the task and I don't see anything even remotely up to the task in common production. One day.

My thoughts around the torque capability of eg QS vs Golden is not so much around outright quantity, I take it on faith from everyone's comments that there'll be ample, more around the ability of the motors to produce torque at very low rpm without cooking themselves. From what I've read it seems like there's two paths - big diameter rotor to reduce the current demand for a given torque, or copious heat shedding ability, or both. I haven't compared, but I think the Golden is a bigger diameter rotor so potentially it might deal with low RPM/high torque more easily than the QS? Certainly I'd expect the liquid cooled Golden to be better at dealing with this sort of mal-treatment.
In my kind of trials riding I spend a fair bit of time slipping the clutch on steep hills whilst dead stationary, while I figure out what the heck I'm going to do next, or building up the courage to actually do it.
I love the idea that Altair has talked about of rolling backwards downhill while applying forward drive with throttle alone - that would be awesome, but I'm guessing it would be brutally hard on the motor?

In absence of other practical evidence I'm backing my guesses with the fact that the biggest and oldest electric moto trials manufacturer in the world has now fitted their top bike with both clutch and additional flywheel. I'm going to do the same. If I discover I didn't need to, well I'll just pull off the flywheel and ignore the clutch.

If someone wants to program a Nucular so I can have two independentally controlled ramp up & ramp down speed curves, with the ramp down having a setting that prevents the motor increasing rpm irrespective of applied load .... :wink:

User avatar
DanGT86   100 kW

100 kW
Posts: 1051
Joined: Sep 06 2012 9:18pm
Location: Saint Louis MO

Re: Observed trials bike designing

Post by DanGT86 » Apr 25 2020 9:15pm

As I was typing my last post I realized just how many justifications for a clutch I agree with. So even if I was completely correct that it wasn't actually needed for the torque impulse hit I'd still want it for power modulation and panic button instant power off switch. I also agree that the getting everything correct would require much better than average programming of throttle control. The last few weekends I have been out tuning my throttle ramp via CA3 on an old school speed based controller. It's been a real pain and I wish I would have pre ordered a nuc months ago.

Despite the point it sounds like I was trying to make in my last post I would probably employ a clutch in a production trials bike if I were a major manufacturer. Asking riders to adapt to an entirely new riding style after decades of muscle memory seems like a bad idea. That and the fact that I'd hate to get good at trials on my electric bike and then be completely useless on a gas bike. Some things might be better left alone even if they are holdovers from old technology.

As for electric clutch, I would imagine there would have to be some programming that basically allows a way steeper throttle ramp up speed via clutch lever release than is allowed through throttle twist. If the throttle ramp was similar to a clutch dump without the lever then it would be terribly aggressive all the time. I can easily imagine tying the lever and the twist throttle together opposing each other such that the total voltage output to the controller is the product of the two. But even in that scenario the speed at which you release the lever would have to vary how aggressive the ramp up is or you would have to modulate them both all the time for smooth operation. Someone crafty could probably make a throttle buffer device with an arduino where a user could manipulate the values based on the throttle and lever inputs. That would require a ton of tuning though. You would have a table based on absolute position of both levers as well as modifiers/multipliers based on the rate of change in lever position. I'm willing to bet the electric clutches on the market are a version of that. Even with the right software a person sick of tuning it would probably think having a real clutch would save me a lot of trouble. Sometimes analog mechanical stuff is just the best way.

Even with a mechanical clutch things might still be more difficult than they seem. For example, lets say your controller was programmed to ramp up current to the motor at a specific rate. If you pull in the clutch and open your throttle all the way there wont be any resistance on the motor. It will spin up to full speed using almost no amps and then hit its programmed rpm limit. From there when you dump the clutch the motor will feel the load causing the controller to begin ramping current up. A BLDC motor at max rpm doesnt produce its max torque because the rotor is already spinning so fast that there is nothing to push on. Kinda like how when you pedal your bike to your max cadence down a hill there is nothing left to push off of because you can't make your leg go faster. I'm not actually sure your max impulse from a clutch dump would be at max rpm. I'm thinking it might not be and that you would want to program your clutch in RPM to correspond with max motor torque. Its a lot to think about.

A little anecdotal experience from my recent throttle ramp adventures that might be relevant or interesting to you: My controller is speed based. So throttle angle % equals desired road speed. The controller will dump max phase amps to achieve the commanded speed. The cycle analyst V3 is between them. I have it set in current mode. So the CA3 attempts to translate my throttle angle into desired amp draw by manipulating the voltage signal out to the controller. This basically fools an old fashioned controller into acting like a current based control instead of speed based. After years of internal combustion engines we are all used to torque based throttle. Its intuitive. When you go up hill you give more gas. In speed based throttle at 125phase amps my LMX motor is violent and un-ridable. It loops out without warning and my neighbor now has a broken wrist to attest to that fact. There are gain settings that control how fast the CA3 responds to current over the commanded limit. With these set high it over reacts and cuts power too early then will reapply it then cut it again then reapply causing a horrible oscillation. With the gain settings low enough to not oscillate it restricts the ramp up speed too slow to be fun. After much trying I gave up on simulated current control and switched to a straight ramp based on time. I just command the volts/sec that the throttle is allowed to rise. This smooths the accel but once you are rolling along it leaves you wanting a faster ramp up speed because a faster ramp from a roll is a whole different animal than ramp up from a dead stop. There are some other programmable features that help work around these issues but so far I havent achieved a throttle that feels right. I have 15-20 hours in tuning and it's just not right. I wish the curve was more adjustable. I believe I need a logarithmic or exponential throttle ramp or one that is more precisely based on road speed and rate of throttle change rather than just throttle angle. Or I could just give up on expecting too much from a $200 controller on a high powered bike.

Long story long here, there is a need for precise control of throttle ramps and there is a lot of habit and expectation we all take for granted after 100 years of very mechanical analog vehicles

bikerpete   10 W

10 W
Posts: 92
Joined: Jan 27 2020 9:09pm

Re: Observed trials bike designing

Post by bikerpete » Apr 25 2020 9:49pm

DanGT86 wrote:
Apr 25 2020 9:15pm
As I was typing my last post I realized just how many justifications for a clutch I agree with. So even if I was completely correct that it wasn't actually needed for the torque impulse hit I'd still want it for power modulation and panic button instant power off switch. I also agree that the getting everything correct would require much better than average programming of throttle control. The last few weekends I have been out tuning my throttle ramp via CA3 on an old school speed based controller. It's been a real pain and I wish I would have pre ordered a nuc months ago.
......
I have 15-20 hours in tuning and it's just not right. I wish the curve was more adjustable. I believe I need a logarithmic or exponential throttle ramp or one that is more precisely based on road speed and rate of throttle change rather than just throttle angle. Or I could just give up on expecting too much from a $200 controller on a high powered bike.

Long story long here, there is a need for precise control of throttle ramps and there is a lot of habit and expectation we all take for granted after 100 years of very mechanical analog vehicles
It's that old nut - what's achievable versus what's practical & expedient.

My guess is that two independent ramps, one for the throttle and one for "clutch" lever would work reasonably well. Throttle to something that you can control smoothly, clutch almost instantaneous. The speed at which you release the clutch would give the modulation, much like a mechanical clutch. How the motor torque profile fits with that is well outside anything I can even begin to contemplate.
My guess is that the human might be the best integrator for that - as long as we get some good analog feedback on the motor RPM it's probably just a matter of gaining experience at judging the right mix. We humans are pretty good at some of those analog integrations. Much like catching a ball at full tilt run - that's an awful lot of digital calculation, but a kid can do it unconsciously after a decent amount of practice. Besides, where's the fun if the machine does everything for you!

You're a persistent man with that controller. That almost reflects what I anticipate with trying to create a full electronic clutch/flywheel effect - just way more time and experimentation than I could tolerate. Good luck.

john61ct   100 GW

100 GW
Posts: 7480
Joined: Dec 18 2018 2:06pm

Re: Observed trials bike designing

Post by john61ct » Apr 26 2020 9:06am

> recumpence's davinci drive. He uses a torque limiting clutch on the primary drive

Anybody got a link to that?

Very interested in anything that would tame a crazy-torque standing start

to protect a "fragile" drivetrain.

An actual physical slipping clutch would be very interesting, but I'm a complete noob there.


Post Reply