How to overlean every wheel

[Chriull]

On 2/19/2023 at 10:00 AM, alcatraz said:

Electric motors often look like this. Is the V13 different?

+1!

I took this graph as chance to once again try to describe how and why overleans happen. This time with some rephrasings by chatgpt to hopefully increase comprehensibility/clarity.

23 hours ago, Uras said:

Our euc motors are powered by batteries - I thought that was the issue with torque for us. Eg how many amps can those batteries supply at a given voltage etc. I think a graph based on battery power wouldn't have straight lines?

The graph shown above represents the typical behavior of DC motors. It makes no real difference how they are powered - cells just decrease voltage with lower charges...

The fundamental formulas that describe the performance of DC motors and lead to this limit graph are based on these first approximation formulas:

The torque produced by the motor is proportional to the current flowing through it, and can be expressed as Torque = Some_Motor_Constant * Motor_Current.

The voltage generated by the motor is proportional to its speed, and can be expressed as Generated_Motor_Voltage = Some_Other_Motor_Constant * speed.

The voltage applied to the motor's armature is the sum of the generated voltage and the voltage drop across the motor's internal coil resistance, and can be expressed as Motor_Armature_Voltage = Generated_Motor_Voltage + Coil_Resistance * Motor_Current.

The relationship between the battery current and the motor current can be expressed as Battery_Voltage / Motor_Armature_Voltage = Motor_Current / Battery_Current.

The maximum current that can be drawn by both the motor and the battery is calculated as Maximum_Current = (Battery_Voltage - Motor_Armature_Voltage) / (Coil_Resistance + Internal_Battery_Resistance).

In reality, there are some practical differences to consider:

- There is a maximum motor current firmware limit. As the motor current is (roughly) proportional to the torque produced by the motor, a firmware-implemented motor current limit will be shown as a horizontal limit line on the graph.

- A battery current limit (as by @Mark_Yu mentioned at ~84A for the V13) starts out higher than the motor current limit line, but decreases with higher speeds until it reaches the limit line shown on the graph. At this point, the torque produced by the motor is limited by the maximum current that can be drawn by both the motor and the battery.

- Although the torque limit lines on the graph are linear, in reality, they are slightly concave due to physical factors. However, for most practical purposes, this slight deviation is negligible.

- The battery voltage to motor armature voltage transformation is not lossless.

- Li Ion cells approximation with just the internal resistance does not show the "two step current capabilities". As cells can provide some higher current from the 'surface charge" for "short" times. More sophisticated equivalent curcuit diagrams use another capacitor to model this behaviour.

TLDR: Every EUC can be overleaned - one just has to accelerate (strong and) long enough. One should not let one irritate by some tiltback inbetween.

It's system-immanent that one reaches the torque limit by prolonged acceleration. The stronger the acceleration, the harder the fall.

The stronger the wheel the higher the speed at which one will faceplant.

If the speed at which one would faceplant is beyond the threshold at which air drag would prevent one from standing, the EUC could be considered safe from faceplants.

Edited February 20, 2023 by Chriull

[Aztek]

3 hours ago, Chriull said:

If the speed at which one would faceplant is beyond the threshold at which air drag would prevent one from standing, the EUC could be considered safe from faceplants.

So, having in mind Master's max speed, it should be considered safe, if charged sufficiently and not pushed suddenly while moving with high speed.

[Chriull]

36 minutes ago, Aztek said:

So, having in mind Master's max speed, it should be considered safe, if charged sufficiently and not pushed suddenly while moving with high speed.

I've seen somewhere from a dyno test (?without real burde) 60mph lift cut off speed? Should be somewhere about 90km/h.

So with some hearty lean (with pads) an overlean around 50 km/h should be not impossible?

So, yes. If staying below this with accelerations should make it quite safe.

Edited February 20, 2023 by Chriull

[Aztek]

16 minutes ago, Chriull said:

with some hearty lean (with pads) an overlean around 50 km/h should be not impossible?

Probably possible. As safe, I meant that one should not be able to make it cut off merely by accelerating smoothly, if the above-mentioned conditions are met.

However, riding any EUC with speed over 50 km/h, especially in traffic, always seemed to me too much risk for the benefit, geared or not. But this is just me...

[Cerbera]

2 hours ago, Aztek said:

But this is just me...

And my axe

[rcgldr]

It's not that simple because of battery charge, voltage sag, and PWM. The same amount of speed and torque at a lower battery charge will require more current than a fully charged battery. If two battery packs have the same no-load voltage, the pack with more voltage sag will require more current, inducing even more voltage sag. This is why Molicel P42A packs are an option for S18. Hopefully P45B packs will soon be used instead (7% more capacity, and even less voltage sag than P42A).

power := force x speed := torqe x angular speed := volts x amps. If angular speed and torque are constant, then as voltage drops, current increases.

Its also possible to overlean or blow a fuse on a stalled motor (zero speed), like a steep incline or getting stuck in a dip.

 

[mrelwood]

I decided to try and explain overleans and cutouts with a different approach:

 

[alcatraz]

Sweet video! Informative, funny and terrifying. 

[alcatraz]

Is there a log for the V13 crash? It would have been interesting to see what the input voltage was at the moment of the overlean.

If the wheel spins out at 140km/h @ 126V then does that tell us anything about what the "self EMF" voltage to overcome is at 80km/h? It's not as simple as 80/140 * 126 = 72V is it? Like the packs sagged down to near 72v at the point of the overlean.

[Punxatawneyjoe]

58 minutes ago, alcatraz said:

Is there a log for the V13 crash? It would have been interesting to see what the input voltage was at the moment of the overlean.

If the wheel spins out at 140km/h @ 126V then does that tell us anything about what the "self EMF" voltage to overcome is at 80km/h? It's not as simple as 80/140 * 126 = 72V is it? Like the packs sagged down to near 72v at the point of the overlean.

 

[alcatraz]

4 minutes ago, Punxatawneyjoe said:

 

Thank you very much. I don't see a log anywhere in the first post. Hmm...  any ideas?

EDIT! Found it!

Edited February 24, 2023 by alcatraz

[alcatraz]

Ok. The pack voltages didn't dip far at all. They were at 102V at their lowest (only 7V under the highest reading of 109V). 102V is well over the motor's "self-EMF" voltage I think. Which means the wheel's self protection circuit must have limited the current. Without self protection the crash could have been avoided is my guess, looking at this. But, some other kind of failure could have occurred of course. 

[alcatraz]

What if they have no choice but to limit current. Like they have fuses that would otherwise blow if they didn't. 

Like, the only way to stay upright is to remove fuses as well. Possibly... who knows.

[alcatraz]

I have to say that I'm impressed with these packs. They didn't sag much at max power output. Seems the temperatures were 15C on motor and mosfets and 22C on the rest of the controller. Ambient temp 5-10C perhaps?

That should make owners happy that the packs seem good. The safety mechanisms just need some dialing in perhaps?

Edited February 24, 2023 by alcatraz

[mrelwood]

7 minutes ago, alcatraz said:

Ok. The pack voltages didn't dip far at all. They were at 102V at their lowest (only 7V under the highest reading of 109V).

109V was already during a very hard acceleration, so at rest the voltage would’ve been much higher. I think I saw something like 92% or 96% for the battery in the display.

7 minutes ago, alcatraz said:

102V is well over the motor's "self-EMF" voltage I think. Which means the wheel's self protection circuit must have limited the current.

The current peaked at 101A, when the total power was at around 10’400W. But your math is missing the energy required to accelerate. The back-EMF and the battery voltage are (nearly) the same only at the peak of a free-spin test, when there is zero energy left to accelerate or balance. At the time of the overlean the wheel was still pushing around 10kW to the motor. But it wasn’t enough, the rider’s lean would’ve required more.

I don’t know what the exact back-EMF voltage would be at 80km/h, but for this example the 80/140 *126 = 72V is precise enough. So 72V was required to maintain the speed, so there was only 30V available for accelerating. It wasn’t enough for the rider’s lean at that speed. At 70km/h the back-EMF would’ve been at 63V, and the battery voltage at that time was at around 103V, so 40V left for acceleration. The power at 70km/h was at 9’909W.

If the ambient temperature was at 5-10•C, the voltage drop was already much weaker than at room temperature of 21•C.

 

7 minutes ago, alcatraz said:

Without self protection the crash could have been avoided is my guess, looking at this. But, some other kind of failure could have occurred of course. 

The crash could also have been avoided by leaning less...

[alcatraz]

Wow, it's very late in Finland. Thanks for writing so late.

Yes, yes. 

It would be very exciting to compare this log to an overlean on a Master or Master Pro, or even an RS.

80/140 = 57% sounds like a low number to hit an actual physical limit. It was a software limit that took the rider down.

I'm interested in the column ID(A). Right after he hit 100A on bus current, ID(A) switched from around 0 to -25.

What's the point of that 200A value then if 100A is the real limit? Hmm...  

Edited February 24, 2023 by alcatraz

[soulson]

3 hours ago, mrelwood said:

The back-EMF and the battery voltage are (nearly) the same only at the peak of a free-spin test, when there is zero energy left to accelerate or balance.

How certain are we of this? I don't think I would rely on the assumption that the "140 km/h free spin speed" is limited by rotor flux or back-EMF. Those limits definitely do exist, but I would want to see more motor parameters first before assuming that 140 km/h is not just some arbitrary firmware emergency halt limit.

Edited February 24, 2023 by soulson

[mrelwood]

On 2/24/2023 at 9:52 AM, soulson said:

How certain are we of this? I don't think I would rely on the assumption that the "140 km/h free spin speed" is limited by rotor flux or back-EMF. Those limits definitely do exist, but I would want to see more motor parameters first before assuming that 140 km/h is not just some arbitrary firmware emergency halt limit.

That is of course technically possible to do, but to my knowledge so far the free spin speed hasn’t been limited by software on any wheel. I also don’t quite see the point in doing so.

[techyiam]

49 minutes ago, mrelwood said:

That is of course technically possible to do, but to my knowledge so far the free spin speed hasn’t been limited by software on any wheel. I also don’t quite see the point in doing so.

+1

I can see the usefulness of knowing what free spin speed is for an electric wheel.

For BLDC motors, peak power occurs at slightly less than the half of the free spin speed if the current is not limited by the controller in any way.

But I fail to see the benefit of limiting the free spin speed by firmware. 

Edited February 26, 2023 by techyiam

[soulson]

The usefulness of limiting the maximum angular velocity in software would be to limit danger when the system is already out of control. If the motor is rotating at 1333rpm / 140kph, something is wrong. We didn't get here by leaning forward, so there's no value in continuing to accelerate forward to maintain balance. However, there is value in halting torque generation, because the wheel might currently be an unguided missile screaming toward a pedestrian/vehicle/dog/structure. It could also by tumbling about or laying on its side, in either case where continued acceleration is unwelcome.

[techyiam]

On 2/23/2023 at 11:52 PM, soulson said:

How certain are we of this? I don't think I would rely on the assumption that the "140 km/h free spin speed" is limited by rotor flux or back-EMF.

27 minutes ago, soulson said:

The usefulness of limiting the maximum angular velocity in software would be to limit danger when the system is already out of control.

You are changing context here, and you are no longer talking about free spin speed. People don't ride anywhere close to the free spin speed. The torque monotonically decreases to zero as the wheel reaches the free spin speed. 

Edited February 26, 2023 by techyiam

[soulson]

30 minutes ago, techyiam said:

People don't ride anywhere close to the free spin speed.

Correct, people cannot ride at that speed. That's why I said that if the motor is rotating at that speed, something is wrong. Either no one is riding the wheel, or the sensors or controller have critically malfunctioned. Therefore, the safest thing for it to do is stop.

[mrelwood]

40 minutes ago, soulson said:

Correct, people cannot ride at that speed. That's why I said that if the motor is rotating at that speed, something is wrong. Either no one is riding the wheel, or the sensors or controller have critically malfunctioned. Therefore, the safest thing for it to do is stop.

And that’s what they do. All wheels stop within a few seconds of reaching the free spin speed, unless something’s wrong with the wheel.

[soulson]

Agreed. My point is just that safety halting point need not necessarily be the maximum rated motor speed. They can be different, and the safety concerns I mentioned are reasons to limit the motor speed by software to less than its maximum rated speed (where the back-EMF would be the limit).

[techyiam]

9 minutes ago, soulson said:

Agreed. My point is just that safety halting point need not necessarily be the maximum rated motor speed. They can be different, and the safety concerns I mentioned are reasons to limit the motor speed by software to less than its maximum rated speed (where the back-EMF would be the limit).

No need.

Maximum Free Spin Speed implies no load, tire off the ground speed test.

Tire on ground, with load, the maximum spin speed is already below the maximum free spin speed. And it is not by choice. There simply isn't enough torque available (not enough current) for the wheel to self-balance any further, nevermind to spin to the maximum free spin speed.

 

 

Edited February 26, 2023 by techyiam