Understanding the real world considerations and implications of Speed VS Torque models

[Vam Monaco]

On 10/27/2020 at 9:51 AM, LCS said:

Hi all, euc newb here.

Based on a cursory search that I performed, I am sure this topic has been covered a few times (though not exactly in the way that I was expecting, which is why I am starting this new thread). 

There are a few quite deep discussions on how electric motors work, what Torque means, power, etc.  These are very in-depth engineering level technical discussions.  I am starting this thread because I am trying to understand the higher level human aspects of torque vs speed models.

I understand that the gotway c38 vs c30 for example have different magnets, and that at higher speeds, due to the larger magnet sizes, the c38 has more power consumption, i.e. less energy efficiency than the c30, and that this is also where that extra torque comes from.

So, being a new EUC rider, with very little experience, I am trying to understand what some of the longer term considerations are between choosing torque over speed.  I am trying to keep this as generic as possible, and not specific to the c30 vs c38, nevertheless, those are the two models I am trying to decide between.

So, will I hit a wall at some point with the torque model, and want those extra few mph that the c30 offers?  Or will the quick acceleration never get old?

Just curious what your experiences have been regarding long term EUC ownership of either an HS or torque model?

I especially want those who have owned both types of wheels (torque vs HS), but being as how different wheels also have different speeds vs torque ratings, anyone that has owned multiple wheels can chime in with their experience.  When a newb stops being a newb and learns how to ride well, what have you found is most important over time, do you prefer that higher speed rating, or are there different types of rides, some of whom don't care about Top Speed?

Let's try to keep it high level (no engineering level technical details), though objective information is welcome, I'd like to hear about your personal subjective experiences.

Ultimately it would be great if this thread can become a repository of information and experiences that people can read about to decide the ultimate question: 

Torque, or HS?

I thought the same as you and bought the C30, but I’m finding that the time it takes me to get up to 35 mph bores me out of my mind.

If there’s anyone here who is envious of that extra MPH, I will trade you straight across for your C38. My wheel is brand new and has no spills.

[EMA]

On 2/22/2022 at 12:57 AM, I_Must_Bust said:

I'm about 65-68kg and, unless I try to, I don't overpower my tesla. Still not sure if I should go with HS or HT for an RS... I'm looking at getting a new wheel for the summer season coming up. I ride trails but not anything like seen on youtube with mountain biking jumps. More like trails that you could ride on a bicycle comfortably. Not sure if I would ever want to go over the top speed of the HT version given that all of my experience is on a lighter and likely less stable wheel. Wouldn't want to hit RS HS top speed on my tesla, that's for sure.

what your "problem" with the tesla ? speed or range ?

you are not an aggressive offroad driver but you are also not a speed demon, honestly you can get both.
 

if you are going with RS go for the torque

if EXN go for the speed

jm2c

[mrelwood]

1 hour ago, Vam Monaco said:

I thought the same as you and bought the C30, but I’m finding that the time it takes me to get up to 35 mph bores me out of my mind.

I don’t follow. If the amount of torque is not enough for you, you must be fighting an overlean. I would say that it’s the least “boring” situation any rider can be in! 

[techyiam]

1 hour ago, Vam Monaco said:

I thought the same as you and bought the C30, but I’m finding that the time it takes me to get up to 35 mph bores me out of my mind.

Are you using power pads on your RS19 HS?

Under normal riding conditions for you, are you finding that you are accelerating quicker on your RS19 than your T3?

[Bizra6ot]

2 hours ago, Vam Monaco said:

I thought the same as you and bought the C30, but I’m finding that the time it takes me to get up to 35 mph bores me out of my mind.

If there’s anyone here who is envious of that extra MPH, I will trade you straight across for your C38. My wheel is brand new and has no spills.

You don't need c38 my friend if 3 seconds is boring for you, you need a MC or a Tesla car at this point lol

https://youtu.be/HlPwe7TmsbE?t=322

[RagingGrandpa]

On 1/4/2023 at 3:49 PM, RagingGrandpa said:

Cool interactive plot

V14 data from Inmotion below.
Assuming a tire outer diameter of 17 inches:

• 210 Nm = 219 lb pull force
• 750 RPM = 37.9 mph ground speed
• From the slope of the graph, freespin is expected near 900 RPM (45 mph) 

And it's quite interesting that the torque becomes voltage-limited at just 1/3rd of the freespin speed, around 300 RPM in this data. 
We've been speculating for a long time about where that crossover point happens; good to finally see an example. 
It means: the 219 lb pull force cannot be produced for speeds faster than 15 mph. And raising the firmware's current limit cannot improve this, because voltage is creating the limit (not current). 

10 minutes ago, Jason McNeil said:

Just received this V14 dyno motor result from Inmotion. Will try to get a comparative graphic with the V11/V13 motors, coils temps & significance of the readings for context.

 

[mrelwood]

 

1 hour ago, RagingGrandpa said:

• 750 RPM = 37.9 mph ground speed

The max riding speed is 70km/h (~44mph), and I would expect the data to reach that speed as well.

1 hour ago, RagingGrandpa said:

• From the slope of the graph, freespin is expected near 900 RPM (45 mph)

These specs have been announced already. Top riding speed is limited to 70km/h (~44mph) by tilt-back, and free spin speed was close to 105km/h (65.6mph). I’m not sure if your tire size calculations are a bit off, or did you just remember the top speed incorrectly.

1 hour ago, RagingGrandpa said:

It means: the 219 lb pull force cannot be produced for speeds faster than 15 mph.

A third of the V14 free spin speed would be 35km/h (22mph). Whether the rule of the third applies to other wheels is unknown. I would assume that it’s determined largely by the motor’s kv rating as well.

[RagingGrandpa]

Dang, that's odd then...

I was ignoring specsheets and only using the information from their dyno graph, plus an assumed tire OD of 17": 
750 RPM for a 17" OD tire gives 37.9mph by my math; hopefully you agree?

Controller voltage was not shown in the dyno data; perhaps it was not the full 134V. 
If full-voltage freespin is really 65.6, and we saw 45 in the test, the test voltage could have been some ~90V: very very low for such a battery pack.
Perhaps they use a large DC power supply and not a battery?
Hope we hear more details soon; otherwise we're left guessing. 
 

1 hour ago, mrelwood said:

determined largely by the motor’s kv

...which is on display prominently, with a dyno result showing torque vs speed: 

kv (RPM per volt) = 900 RPM / 90(?) Volts = about 10 for this motor
 

Edited September 14, 2023 by RagingGrandpa

[mrelwood]

 

19 hours ago, RagingGrandpa said:

750 RPM for a 17" OD tire gives 37.9mph by my math; hopefully you agree?

It does. Some manufacturers call these beefy 3.0-12 knobbies 18” wheels, so that might be closer to the actual diameter. Though even that  would make 750rpm only 64.6km/h (40.4mph). 750rpm at 70km/h (43.75mph) would make it a 19.5” tire. That doesn’t fit either. Seems that they just didn’t scale the graph all the way to the max riding speed.

19 hours ago, RagingGrandpa said:

Controller voltage was not shown in the dyno data; perhaps it was not the full 134V.

Making torque tests with an empty battery wouldn’t make any sense to me. Sure there’s a largeish voltage drop when making torque tests, but other than that I’m sure the battery must’ve been pretty much full.

19 hours ago, RagingGrandpa said:

If full-voltage freespin is really 65.6, and we saw 45 in the test, the test voltage could have been some ~90V:

EUCs don’t let the voltage go anywhere near that low. That would average 2.8V per cell. I’m sure the V14 tilt-back kicks in at 3.15V at the latest.

19 hours ago, RagingGrandpa said:

...which is on display prominently, with a dyno result showing torque vs speed: 

kv (RPM per volt) = 900 RPM / 90(?) Volts = about 10 for this motor

Oh right, didn’t think it further. Though I still don’t think 90V is the correct value here.

[RagingGrandpa]

To expand on the practical relevance of dyno data... this is long, but I'll leave it below for the patient few that want it. 

Q1) How should we decide what "top speed" to advertise, for a given EUC, without using human crash test dummies? 

"Theoretical top speed": 
To decide this based on objective analysis, we'll need to develop some estimate of the tractive force required to maintain a given groundspeed. This "speed force" graph needs to account for some rolling resistance, and some wind resistance. It assumes a standard-size adult, riding in a standing position, on a level surface, with no wind. It is regardless of EUC type. 
Next, for a specific EUC model, let's assume we have access to a dyno plot showing maximum torque (without pedal dip) at various speeds, like we have for V14. 
The rest is simple: we want to find the intersection of the two graphs, where the motor torque matches the "speed force." To do it, convert all the data into the same physical units: force becomes torque, and RPM becomes mph, both based on the tire size of the specific EUC. 

This shows the theoretical maximum speed where self-balancing is possible, meaning that just a tiny bit of disturbance would cause a crash at this speed. Think: a bump in the road; a touch of headwind; opening your arms and causing more wind resistance. So this theoretical speed is certainly not the "safe top speed." 

There's a great example of determining theoretical top speed for e-bikes, on Grin's page. It uses their database of motor dyno data, and their estimate of speed force, and then you select the tire size. The speed where the black "Load" line intersects the red "Power" line is the answer (47kph for the configuration I linked). 
(Grin plots the speed force in units of power, instead of units of torque, but those are easily converted when you know the tire diameter.) 

"Safe top speed": 
There is no definitive answer here, because there are too many variables. 
Surface type, tire type, tire pressure, wind, clothing, mechanical wear, component variation... just too many things affect the result. 

I think you'll have to choose some standard margin, which I'd suggest is some offset from the "speed force." So we would offset the speed force data by perhaps +20 pounds at all speeds, and then repeat the exercise above to find the speed at which the dyno plot intersects it. This way, "safe" means we'd have 20 lb of excess force available to handle balancing disturbances, at the safe top speed. 
And there's plenty of room for improvement here. Instead of an arbitrary +20lb offset for all EUC types, we could offset by the torque and force of 3 inches of rider CG movement (forward lean), given the tire size of the given EUC... 

I don't recommend calculating safe top speed as a simple percentage of the freespin speed, nor as a percentage of the theoretical top speed from above. A simple percentage assumes a linear motor characteristic; but as you can see from these plots, there are many non-linear areas- so it's unrealistic to assume the limits scale linearly. 
A simple percentage is tempting, but sloppy. Assuming we have dyno data, we can do much better, and with only a tiny bit more math. 

 

Q2) What metric can be used during riding, to tell if the rider is nearing the torque limit? (e.g.: an overlean crash)

The controller knows this directly. Its self-balancing software is commanding the motor torque within an electrical range of -100% (max reverse drive) to +100% (max forward drive). We cheat and call this "PWM duty percentage" as if it were a brushed DC motor... but in reality, it's a commanded effort % coming from the feedback control software, going into a BLDC commutation controller. It's still a normalized percentage, so I'm happy to keep using the name PWM %, as long as we don't lose sight of the fact this is a brushless motor with electronic commutation. 
100% is a hard limit. There is no such thing as 101%. 
100% would give you the maximum torque available. At any speed*. At any voltage. With a headwind; up a hill; always.
Since this PWM % software variable is so reliable, and already exists, I think it's an obvious choice to use it to tell if the rider is nearing an overlean crash. 

* One caveat to the PWM %: At lower speeds, 100% effort is not permitted, because the motor current limit is reached and triggers separate protection limits. It affects torque: at low speeds, the motor current limiter will defeat self-balancing and cause an overlean crash in the same way that reaching 100% PWM at higher speeds would. 
So if we want a good overlean alert metric at lower speeds, it needs to be based on more than just PWM % alone. 
The alert logic could be as simple as:  Warning = (PWM% > Threshold 1) OR (phase current > Threshold 2)

And there is a dynamic aspect not discussed above: the rider will be changing their body position, accelerating and decelerating, impacting obstacles etc, and so the PWM % is not steady during riding. To be useful, we need a warning that occurs in advance of the loss of control, to give the rider sufficient time to react and reduce their acceleration command.
It sounds very challenging to me, to use simple software to predict ahead of time that the human is about to command an impossible acceleration. From personal experience: rate-of-change detection usually leads to nuisance alerts. Until a practical method is demonstrated, we'll need to keep today's level of caution about sudden acceleration pedal dips, and not expect warnings to save us in all situations. 

The remaining decision is "how much margin?" / "what are the thresholds?"
In some situations, I want to keep my riding below 60% PWM, ensuring a large 40% margin available. Think: rough MTB trails, where we strike roots and rocks unexpectedly. 
In other situations, I think people want slim margins and maximum speed. Racing on a smooth asphalt racetrack is exactly this situation- but the racer still wants to be alerted if they're about to crash! 
So I'm suggesting: in order to be useful, the margin should be user-selectable. Otherwise people will quickly discover it's wrong for their riding style, decide to ignore it, and we'll be back to the "gotta ride the beeps to win the race" situation which we're stuck in today. 

 

Q3) When the rider is nearing that torque limit (indicated by the metric above), how should we alert them? 

This is a human-interface question, and inherently subjective. 
Some people like beeps; other people can't hear them.
Haptic feedback like tiltback or intentional motor vibrations seem tricky to make reliable and clearly perceptible.
It becomes a discussion of human opinion more than objective analysis, and I don't have any suggestions beyond "experiment with everything practical and choose what people prefer." 

And kudos to @Freestyler for taking initiative, creating prototype firmwares, and getting Gotway's attention with PWM %-based tiltback features!

Comments appreciated
 

Edited October 4, 2023 by RagingGrandpa

[mrelwood]

1 hour ago, RagingGrandpa said:

Q1) How should we decide what "top speed" to advertise, for a given EUC

What people easily lose sight of is exactly this. The MSX 84V for example is advertised as a 60-65km/h wheel, while the similarly (or slightly better) powered V11 with the same free spin speed is electronically limited to 50/55km/h. Then people choose the MSX because they think it’ll have a larger safety margin at their riding speeds. But no. They’d be better off with the V11.

 And the OG Sherman has a 104km/h free spin speed, and advertised riding speed I think was 80km/h? The V14 has practically the same free spin speed, but the top speed is limited to 70km/h by tilt-back. Again, not comparable for safety margin.

 If you compare wheels based on a certain safety margin, be sure to compare the same parameters for different wheels. Limited top speed of an Inmotion or KS is not comparable to the unlimited top speeds of Begodes and the likes.

[alcatraz]

It's super dangerous to look at advertised speed specs equally. 

The question to answer: Does this wheel have enough torque left FOR ME to remain upright at this certain speed, wind resistance, weight, road inclination, rolling resistance?

Big brand wheels can also come up short under certain conditions. That's why I recommend for everyone to do a quick calculation for every wheel they ride, before they set off. 

Even a battery that's done 5000km might perform differently and as such doesn't provide the same torque as when it was new. Someone that relies on the wheel's safety measures alone might be sorry. Add a greater margin for certain circumstances. The average chinese test rider is wayy lighter than most westeners. 

Edited September 29, 2023 by alcatraz

[timmytool]

As a T4v1 user that now has a v2+ motor while they're both c30 the v2+ motor has a lower kv (I'm also sure the v2+ motors' efficancy is also lower also). A lower kv while not the same as magnetics changing from c30-c38, in practice it is here.

My observation is: more kv(thinner magnets) better efficancy at higher speeds and a higher top speed. Lower kv(wider magnet) will give more torque but it will disappear sooner, and have a lower top speed. Note how the motor is loaded and driven affects how "kv changes" effects its use. Multirotors flyers talk about kv changers differently then euc users because 

An analogy is release angle while throwing a ball, if you throw it to hi(low kv) it will get good hight(torque) but low distance(top speed), to low(hi kv) and it wont get much hight(torque) but will get better distance(top speed). Just like throwing a ball to much kv will have no hight and ground out with little distance aka not enough torque to maintain any speed. To little kv and it will out torque god, but a walker will have a beter top speed. If you take this to the extremes and map the balls trajectory there is a optimal point based on the needs, is it top speed, torque at 30km/h for trails, efficancy at 40km/h to street cruise, etc etc. But takeaways are more kv will have lower bottom end torque but it will extend out longer before dropping off. A higher voltage battery likewise will be like a stronger throw, it will hold its torque longer before dropping off and will have a hight top speed. More amps does not transfer as easy, its like you throw it at a steeper angle and faster but still get the same distance, aka more torque along the whole power range but the same free spin speed(you'll get a marginally better safe top speed).

But lastly motor construction like iron core flux capacity, windings, heat tolerances etc etc will affect motor performance. Each needs to be optimised to get the best out of the magnets used. Going to c40 from a motor optimised for c30 will necessitate other changes to remain optimal, but with no euc motor being optimal(in my view) the euc companies just use the magnets as a cheap way to change the motors kv without needing to alter the stator windings.

[daniel1234]

On 1/3/2024 at 7:38 AM, Jason McNeil said:

This is some insightful data on the efficiency curve of the V14's motor

just few insights, theoretical graph with real numbers (those numbers from inmotion seems off, but distribution is similar) :

 

- 1 turn = 2pi radian -  at low speed you can get to 30 percent efficiency - so 70 percent of input power will end in waste heat. Problem will scale with size of the motor, rotor mass, and its power(back-EMF). 

- I would love to see different motor types on our EUC. Adding voltage / power is not yielding desired outcome. Wheels getting heavier and range is not increasing. 

So companies are pushing axial flux motors, some are trying to make motors without magnets -> lets see what future will bring.

 

Edited January 8 by RagingGrandpa
(bugfix quote)