Observed trials bike designing

bikerpete said:
To have a manageable throttle there must be some ramping, but for the clutch we want none.
How about a switch on the clutch lever that when the lever is pulled switched the controller to no-ramp mode. This might make the effect of pulling and releasing the lever more similar to a mechanical.

Still doesn't solve the flywheel issue, but it's possibly achievable today.

Just to be sure we are on the same wavelength, does "ramping" for the throttle refer to either:
Slew rate limiting of the throttle signal, which limits the rate of acceleration of the motor, or:
Non-linearity in the throttle signal, which produces an exponential response curve, resulting in less throttle sensitivity at low openings, and more at higher openings.

Personally, I don't like slew-rate limiting because it kills the small throttle corrections you need to make on rough terrain.
Non-linearity is the thing. Some controllers have a choice of different curves, others allow you to "bend" the curve as you want.
However, if your controller doesn't have that feature, there is a way to accomplish non-linearity if you have a resistive throttle. By adding a trim pot between the wiper and the ground connection of the resistive throttle, it becomes possible to obtain a VARIABLE curve that you can adjust to your liking. I've been wanting to order a Domino throttle for quite a while, to be able to develop that.
It's on my "to do" list, I promise. :wink:

One (small) problem with a non-linear throttle is that you never quite get used to the sensitivity of the throttle, because it's different for each position. But this is still better than a straight throttle.
However, on a low-power bike, the straight throttle won't be much of a problem, while a high-power bike will be difficult to ride at low speed. There's also personal preferences in all of this.
End of rant.
Glad you are back in the thread.

The Domino throttle is the best thing for any ebike. I love mine. Well worth the money.

I didnt realize you could tune the curve with resistors though. I thought you could only remove the deadband at the bottom or change the max point to an earlier position. I assumed it would always stay linear though. Is this not the case? Could it be made exponential/progressive with just resistors?
Nice to have you back Guy.

I was thinking slew rate with the throttle stuff, but really I still think an electronic clutch is a very poor cousin to a "real" clutch.
Yes, the response curve of a pot can be distorted away from linear, by adding a single resistor between the wiper and the ground connection.Depending on the value of the resistor relative to the value of the pot, the curve can be "curved" away from linear.
Check the article here:
The graph for the Log Taper shows how the curve is affected. The resistor could be replaced by a pot, which would then allow to vary the distortion of the curve at will.
Other articles: http://www.geofex.com/article_folders/potsecrets/potscret.htm

There is a problem however, in that the very low end of the resulting curve is distorted in the wrong direction, as seen on the graph. But it would be possible to eliminate this small part by either, limiting the mechanical rotation of the throttle so it cannot go as low. You will then adjust the controller settings so the minimum is now at that new voltage.
If not possible, you could try to just set the controller input to "Hall" instead of "0-5V". By doing so, the controller now expects an input voltage between 1 to 4V. You might now have to limit the maximum mechanical opening of the throttle to stay below 4V.

Those are the things I was planning to experiment with, once I have a Domino throttle. :roll:
Interesting. I find that I like a little less rotation than the domino has anyway so having to limit the bottom of the rotation is not a bad thing. The logarithmic curve in that link might be exactly what I am looking for trying to tame my cheap speed based controller.
I rather abandoned this thread and the time has come to explain why.

I was reasonably well down the track of building my dual-suspension fat-bike based e-trial bike when I happened upon an interesting e-trial that seemed pretty similar to what I was planning.
Now after something more than a year communicating with the manufacturer, testing the bike and putting everything in place I have started importing and selling these bikes in Australia.

So my e-fat bike is returning to it's original role.

This new bike runs a brushed motor and a simple controller, with a 2-speed gearbox. The combination is responsive and effective. The LMX motor and Nucular controller I have will fit nicely into the bike, so I'll fit them at some stage.
Maybe I'll get as far as integrating the clutch and flywheel as I'd planned with the e-fat bike - we'll see. I still think a flywheel and clutch is a key enabler for some trials moves, although I've also learnt that a light, responsive, clutchless e-trials bike has it's own attractions.

If you're interested, http://www.dobbikes.com.au
spinningmagnets said:
Do you have any bnb pics of the 2-speed gearbox internals?

No, I haven't pulled one apart yet. It's on my list of things to do!
It's impressively compact - 10ml is the recommended amount of oil!
I'm guessing it's straight spur gears inside from the sound and feel of it.
There's no syncro, but it changes quite nicely either stationary or moving (not under power).
SlowCo said:
The link you posted doesn't work as it puts the E-S site in front of it.

Nice looking bike and a smart 2 speed gearbox! Any more info/specs on that box and maybe even a possibility to buy those seperately for home builders?

It reminds me a bit of the Oset bikes:https://osetbikes.com/au/bikes/new-20r-mkii-lipo-5200-order-now/
Although I would prefer the Sur Ron Light Bee at that price point:https://www.surronaustralia.com.au/...-ron-australia-light-bee-l1e-road-legal-d4nbl

I've edited the link so it works now.

I don't think the gearboxes will be available separately any time soon - mostly because at present the demand for the bikes in China is so high that they're flat out keeping up.

Yes, it's very similar to the Oset - kind of overlaps both the 20" & 24" Oset. I think it's a nicer build than Oset in several areas, although Oset has it's own advantages.

It's a very different beast to the Sur-Ron - I don't really see them in competition at all. The Sur-Ron is rubbish as a trials bike in comparison, and the dob will never compete with the Sur-Ron as a fast trail bike. Sur-ron weighs 50kg vs dob 35kg.
I've said time and again that what a lightweight Emoto needs is a 2-speed, preferably a deep reduction and a direct gear.
I think a clutch and flywheel would be an asset however.

How does the bicycle sourced air suspension hold up?
I love it! It is small, light & cute.
Do the tires have a good grip on wet rock? The knobs seem a bit small. Are they made of a good compound?
Tubeless or not?
Altair said:
I love it! It is small, light & cute.
Do the tires have a good grip on wet rock? The knobs seem a bit small. Are they made of a good compound?
Tubeless or not?

There are two tyres available (3, but the third is not a great option for any condition so we ignore it):
Kenda Krusade 20 x 4 - quite low profile knobs but largest volume and softest, grippiest rubber. Better for hard surfaces.
CST 20 x 4 - bigger knobs, more ramp on them in one direction, lower volume. Better for softer surfaces.

It's been summer here so wet rocks haven't been on the menu. With winter coming I guess I'll find out soon.

I tend to think riding this bike should be considered closer to riding an electric trials bicycle rather than a trials motorcycle - the technique is closer to a bicycle than a moto, although there's a fair bit of overlap.

Grantmac said:
How does the bicycle sourced air suspension hold up?

As well as any suspension I guess. If you service the forks periodically then they should last a long time. The rear shock is working exactly as designed, and well within it's design pressure/load so it should last many years if kept clean.
Ohbse said:
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 old Kelly KEB is capable of 450A phase, and no throttle lag. And it is significantly smaller than 24f. Same 90V max battery.
It also can instantly put all throttle commanded torque on the brake release.
I have broken the motor shaft when accidently connected lanyard kill switch when throttle was wide open.
Torque mode is magic. I can pull instantly 12-13 kW from battery, and right after crawl like 10 mm/sec consistently on the low throttle without any throttle curves.

I am currently using it on the max battery and phase current settings with QS2000.
At 12-13 kW peaks from battery QS2000 is reaching 120C winding current in 10-15 minutes practicing trials power moves depending on the ambient temp.

My thought on flywheel and clutch. Here is the dyno for the modern GasGas TXT 300:
Gasgas 300 Dyno.jpg
Just 16(11.7kW) WHP and 13.7 (18.6N⋅m) torque, not mutch. It is all in the flywheel & clutch.

The energy that can be stored in typical trials bike flywheel @ 10K rpm is 6-7 KJ (by my rough calculation)
It is enough to pull 150kg from 0 to 30 km/h. How fast it can be dumped?
I think the final limiting factor is tire traction.
How much traction can you get from suspension preloaded and sqashed 5 psi trials tire? 1.5G peak acceleration or maybe even more?
Nice torque curve on the TXT 300. Very strong off the bottom. Looks like it was not yet jetted proper to make power as observed above the 6200 rpm break in the power plot. Overall looks to be a great power plant for the class. Unfortunately, with the limitations of the piston engine, it mostly relies on the clutch when you need a more sudden jolt of energy.

Looking forward to your electric build developments. Electric should be able to add a whole new level of performance to the sport.
speedmd said:
Nice torque curve on the TXT 300. Very strong off the bottom. Looks like it was not yet jetted proper to make power as observed above the 6200 rpm break in the power plot. Overall looks to be a great power plant for the class. Unfortunately, with the limitations of the piston engine, it mostly relies on the clutch when you need a more sudden jolt of energy.

Looking forward to your electric build developments. Electric should be able to add a whole new level of performance to the sport.

Trials ICE engines are so focused on the low end of the power curve that it'll be more than jetting causing that power dip I reckon - long intake tracts, super long exhaust headers, almost no tuned resonance exhaust, tiny carb bodies for throttle response at low RPM ...

I increasingly stick to the belief that there's nothing unfortunate about needing a clutch. A clutch and flywheel are an elegant and efficient solution to a range of problems, particularly relevant to trials. In some ways I think it's the old hammer and nail analogy - "if you have a hammer, everything looks like a nail". Just because you might be able to achieve a similar result electronically doesn't necessarily make it "better". And at the moment no-one has come even close to being able to do electronically all that a flywheel/clutch combo can do.

It's a significant challenge to electronically create a device that:
  • provides "inertia" to roll up an obstacle with absolutely no chance of the wheel spinning up if it loses traction for a moment
  • enables instant cut to that "inertia"
  • provides instant power with both hands in positions of strength on the handlebars (twisting the throttle down is incredibly difficult to do when you're jumping up with your hips coming right to the bars!)
  • allows for very fine and fast control of power delivery (our index fingers are significantly more sensitive and accurate than moving our whole hand/wrist). This is in conjunction with the item above - try to finely and accurately control a throttle whilst jumping through the bars! ROFL. Way better to build up some stored power, then control it with a nice stable index finger.
  • Smooths out accidental throttle impulses due to rough terrain, but provides instant power response when wanted.

Here's links to two videos. This is a technique fully reliant on instant on/off power.
First is an ICE bike doing a "front on". Carefully watch clutch, brake and throttle and you'll see how precise the timing of all this is (this guy is 2 times Australian Champ, 18 x West Aust champ, used to ride Worlds & Euros etc.) The interval between the clutch lever starting to move out and starting to move back in is 1 second! The interval between clutch starting to move out and the tyre leaving the ground is under 0.75 second.


Second is an electric trials bike with clutch doing the same technique.


I've never seen anyone on an electric bike without clutch do this (could be I just haven't looked in the right places).
The sequence for this is roughly:
Build RPM (inertia) with front brake locked and back wheel held in place (usually with slipping clutch, but could be brake if you're coordinated enough)
Rapidly pull hips/torso forward and up.
As you extend into the "up", but just before full extension, release clutch & brake, close throttle; pretty much all at the same time.
Pull the bike up with the bars as it accelerates, but let it jump ahead of you.
Land on the back wheel, cut power, balance and manage exit off the top.

I cannot imagine how you could do this with just a throttle for most of the reasons in the list above.
Certainly on my unclutched e-trials it is a complete joke trying to do this (I can do it competently on modest obstacles on an ICE bike), the ramp up of power is just waaaay too sloooow, and this bike is pretty quick off the bottom being specifically designed for trials.

Clutch/flywheel combos are pretty good devices I reckon!
Agree, A Clutch will not totally disappear with a electric drive in this app, but rather thinking it could be changed significantly and improved upon possibly. The hydraulic clutch was a big improvement. A programable Electronic clutch (with dialable launch and brake control) may be even bigger improvement. Certainly would need to develop and phase in as/when mature.

The Move in the first vid is one I have practicing a bit recently. Near impossible for me to maintain a good grip with the old cable clutch. Glad those days are gone.

On the motor tune, certainly loads of factors at play here, but trace shows a reasonably bad stumble. Most likely more than a few things going on but place for me to start would be carb settings. It could take some time to work out if important to you. Certainly secondary to a good low end for most.
I haven't ridden a cable clutch in decades, don't think I'm that keen to do so!

Are you talking electrically actuated mechanical clutch? Or an electronic non-mechanical clutch?
I've never thought about an electrically actuated clutch. Could be intriguing, but I'm not sure how much it would really achieve.
I suppose if you put accelerometers on the chassis an electronic clutch could do a bit of the balance control, but that could also be done via throttle. Not sure it would be an advantage or a disadvantage really.
Maybe you could program it to respond to suspension compression - releasing the clutch when suspension is compressed and you have most traction? Sounds a bit scary if the rider isn't quite in the right position as it takes off.

Something the electronics can't do is anticipate the future, which riders do all the time. A splatter might be the classic - rev the bike to the moon, then drop the clutch intentionally looping the bike out backwards knowing that in a fraction of a second the rear tyre is going to impact a big boulder and start rotating the front back down. How do you set an e-clutch to allow that but also provide beneficial launch control? Wheelspin detection might be helpful perhaps, but that starts to get pretty complicated on a vehicle that can at times intentionally be put in radical attitudes (nose wheelies, looping out past vertical as for splatter, etc).

It seems to me that a lot of the automation in road and air vehicles probably don't really apply very well to things like trials bikes that are thrown around all over the place on highly variable surfaces. There isn't really a stable reference to work from. When sometimes you want to be on two wheels, sometimes hopping on the rear and sometimes standing up on the front wheel even gravity isn't an easy reference to work from. Add in the range within which the centre of gravity of rider & machine can move through and it get's pretty complex to automate much.
Anyway, one of the attractions of trials is that it's hard. Who wants to make it easy :)

I think by the time just about anyone mortal is in the 6-9K RPM range on a trials bike it's pretty much academic if it's making best power/torque. Most of the time you get there while winding up the flywheel ready to jet off, the tiny lag in acceleration is virtually irrelevant. Certainly I'm getting well nervous by the time I've got that much energy under the dubious control of my index finger!

One of the things many people mention with e-trials bikes is that it can be hard to judge the amount of power built up in the flywheel because there's so little audio feedback. Something I guess you get used to with time, but it's certainly more subtle than a 2-stroke exhaust noise! Fortunately.
Cable clutch is all there was when I started. Lots to consider on the tech side of the human interface for a trials setup. Thinking some sort of Launch control can be setup simple or as involved as you can dream up. Motor ramp up- Rev limiter and clutch engagement timing- intensity set to what ever impulse via a tested out delay. Traction control is a pretty well established technology in road cars now. It would certainly need to be thought through a bit deeper for this app. As you say, ones body movements is all important and ones dynamics may be something to add to the telemetry mix.

As far as clutch's go, the existing mechanical ones are relatively bullet proof. Ideally, some sort of eddy current type setup would be my thoughts for a future goal. For starters, most likely a solenoid actuator would be a option. Something that has some very good feel (variable pressure).

Lots can be done on the balancing side via gyros, but that is pushing things a bit too far from a sports standpoint IMO. The technology is Certainly worth exploring- having on the grocery getter with the young ones onboard if can be done simply and cost effectively. Interesting times.
balancing stick!

small bike
I started in the '70's, so yep cable clutch only back then. Glad not to have one now.

I guess a lot of this comes down to the name of the game.
For the most part Observed Trials is simply about who has the most skill at riding a motorbike over difficult terrain without putting their feet down. Adding electronic geegaws and jimjams to compensate for lack of ability makes it another sport.
Traction control - that's the rider's job & a skill to develop.
Launch Control - also rider's job.
Balance control - that's the fundamental rule of the game. Allocate that to an electronic controller and you might as well sit at a desk & play a computer simulation!

I'm all for electric e-trials bikes - less maintenance, little noise, low emissions, smooth power delivery etc. But I can't see any reason to move everything to electric - clutches, brakes, suspension ... they all work pretty darn well using mechanical systems.
Even F1 cars still use hydraulically actuated multi-plate friction clutches (a 1kg clutch can handle 1000hp! How about a tiny carbon plate trials bike clutch?).

Those self balancing gizmos are certainly pretty cool, and things like one-wheel scooters are excellent devices using self balancing tech.
But I agree with you that they're a bit too far - they fall into the same category as 2-wheel drive, caterpillar tracks and lifting rotors - no place in Observed Trials as far as I'm concerned.

Change for change sake can be a bit of fun and can provide learning, but can also be a big waste of time.
Agree on the sport being a test of rider skills. Still, progress toward what makes a better ride within some limitation of active systems important also.

Electric motor can be made with a relatively light or heavy rotor. A separate powered "flywheel-motor-brake" can be clutched in as needed for bursts. May need to be a twin, running in opposite directions so rider balancing is not made obsolete. :shock:

Lots of topics to work out.
Something I have been wondering about with this is the current ramp up time. Controllers limit the rate current is allowed to raise. So whether you went full throttle or dumped a mechanical clutch the current ramp up time and thus torque ramp would be very similar. Obviously the mechanical energy in rotating mass would make the clutch dump a bit more powerful but I think an electric motor would feel like it bogged down after the initial hit.

If the controller was set for a super fast ramp to avoid the bogging down after a clutch dump then the bike would ramp too violently when using regular throttle.

So to get a good riding experience you need a very wide range of tuning on the controller end and a lot of very precise data about the behavior of gas bikes you are trying to duplicate. Right?

Seems to me that a lot of the behaviors that define ICE motorcycle are unintended results of gasoline and airflow and the very specific powerband they create.
Interesting concept moving the flywheel off of the direct drive or clutched main motor. One could minimize the moment of inertia (MI) of the main motors rotor and it would spin up very quickly compared to spinning up the flywheel at the same time. It would also stop faster. Possibly not needing to be clutched at all on the main motor side. The flywheel-motor (kinetic drive) would be somewhat independent and free to add a significant bump when the main motor is moving slowly or barely moving. Riders Speed control of the kinetic side would be an interesting analysis. The combination may yield a much more responsive and drivable setup than a full time connected flywheel. The separate kinetic drive could also be connected and not spun up during low traction conditions and could calm down a overly peppy motor controller setup.
Traction control - that's the rider's job & a skill to develop.
Launch Control - also rider's job.

Looking a new bike previews, this topic is no longer in the distant future for most all riders. KTM announced both traction and launch control on 2023 race bikes. A bunch of other big improvements. https://www.youtube.com/watch?v=AXLroHf6z8M