Tesla & other EV's and hybrids are dangerous!

[PlainName]

I would say one-pedal driving on most EVs is a bit like the kind of braking you get when you put a manual ICE car in 3rd gear and let off the throttle when travelling at say 60 mph.

Thank you - that gives a very good idea. I have to say that's not what was coming across from earlier comments.

Obviously it would not brake at 265kW even if that was technically possible.

As I said, the comments implied it was rather more than ICE enging braking would produce.

Indeed this is a thing designers can control however they wish in software. I thought that would be obvious.

Of course it's obvious! Jeez, you can be a patronising asshole sometimes, you know.

[thm_w]

I wonder, if the regen is set to 'quite heavy' and his foot slips (or he otherwise lets the pedal up), would the driver get nicked for brake testing the vehicle behind?

Unless you cut someone off or have non-functioning brake lights, the fault here is 100% on the person behind, as it means they were following too closely. Full brake pedal at 100km/h is -1g or more (model 3). The max regen is -0.2g. So its quite a difference from hard braking.

There is some good data here, some of it is obsolete as the software has changed since 2020: https://www.autosafety.org/wp-content/uploads/2015/03/Tesla-Regen-Brakes-and-Sudden-Acceleration.pdf

Honestly I would recommend test driving, even if you have zero interest in buying the car. From an "electronics tech" standpoint alone.

[tom66]

Also, brake lights illuminate on regen once over a certain deceleration threshold.

European spec;  USA/elsewhere may be different:

• Below <0.7m/s^2 there may be no brake light signal
  
• Between >0.7m/s^2 and <1.3m/s^2 the brake light signal may be generated
  
• After ≥1.3m/s^2 the brake light signal must be generated

Arguably this is better than conventional cars, where brake lights illuminate even if you hover over the pedal.

[PlainName]

I wonder, if the regen is set to 'quite heavy' and his foot slips (or he otherwise lets the pedal up), would the driver get nicked for brake testing the vehicle behind?

Unless you cut someone off or have non-functioning brake lights, the fault here is 100% on the person behind, as it means they were following too closely.

That only applies if they run up the arse of the brake tester. There doesn't need to be an actual accident to be nicked for it - careless or inconsiderate driving would cover it. Of course, it would need to be witnessed, but with dashcams in practically everything nowadays, braking hard for no reason other than a clear road in front could be tricky to pass off if the car behind has to take evasive action.

[default0.0player]

Also, brake lights illuminate on regen once over a certain deceleration threshold.

European spec;  USA/elsewhere may be different:

• Below <0.7m/s^2 there may be no brake light signal
  
• Between >0.7m/s^2 and <1.3m/s^2 the brake light signal may be generated
  
• After ≥1.3m/s^2 the brake light signal must be generated

Arguably this is better than conventional cars, where brake lights illuminate even if you hover over the pedal.

It's still problematic. When driving uphill, the driver may attempt to accelerate but the gas pedal is not pressed deep enough so the slope is still decelerating the car, the brake light may illuminate, confusing the driver behind. The same applies when driving downhill, the regen torque may not be enough to compensate the slope, the car may be accelerating despite the regen power is significant. In this case the brake lights won't illuminate but the car behind it should brake.

It's better to use torque-to-weight ratio or power-to-weight ratio, when the regen exceeds said ratio, illuminate the brake lights

[Ice-Tea]

It's still problematic. When driving uphill, the driver may attempt to accelerate but the gas pedal is not pressed deep enough so the slope is still decelerating the car, the brake light may illuminate, confusing the driver behind.

If the car in front of you decelerates, I'm not sure if it's a bad thing you're aware of that...

The same applies when driving downhill, the regen torque may not be enough to compensate the slope, the car may be accelerating despite the regen power is significant. In this case the brake lights won't illuminate but the car behind it should brake.

...

Why? The car in fron of you accelerates and then you (the car behind) should brake?? Perhaps you should brake because of the slope but that would have nothing to do with the behaviour or danger posed by the car in front of you...

[woody]

It's still problematic. When driving uphill, the driver may attempt to accelerate but the gas pedal is not pressed deep enough so the slope is still decelerating the car, the brake light may illuminate, confusing the driver behind.

Not sure how that would confuse the driver behind? After all, whatever the reason the car in font decelerates (braking, hill, massive engine failure, a 16 tons weight suddenly appearing in front) it is nice to have this signaled to the car behind. That is why using a deceleration value coming from a motion sensor is a much better parameter to decide when to light the brake lights than having a switch on the brake pedal. The more so because with a motion sensor much more difficult programmable electronics are involved

[woody]

There is some good data here, some of it is obsolete as the software has changed since 2020: https://www.autosafety.org/wp-content/uploads/2015/03/Tesla-Regen-Brakes-and-Sudden-Acceleration.pdf

Thanks for that!

[Siwastaja]

It's still problematic. When driving uphill, the driver may attempt to accelerate but the gas pedal is not pressed deep enough so the slope is still decelerating the car, the brake light may illuminate, confusing the driver behind.

Not sure how that would confuse the driver behind? After all, whatever the reason the car in font decelerates (braking, hill, massive engine failure, a 16 tons weight suddenly appearing in front) it is nice to have this signaled to the car behind. That is why using a deceleration value coming from a motion sensor is a much better parameter to decide when to light the brake lights than having a switch on the brake pedal. The more so because with a motion sensor much more difficult programmable electronics are involved

Exactly - acceleration-based brake light is purely positive with no downsides - it catches those cases where brake light earlier didn't light up when it should - " the driver in front of me is slowing down significantly - need to pay attention". The reason for slow-down is irrelevant. Driver failing to find their gas pedal in steep uphill is a good reason, because usually all other drivers have no such problems, and as such, slow-down is surprising.

Blink mode for long-sustained heavy braking adds even more safety.

But sure those thresholds need to be carefully set. My observation of KIA EVs is that they blink brake lights totally randomly with no reason whatsoever in smooth road traffic (or even lack thereof). But if what tom66 says its true, EU would regulate also the lower threshold where lights are allowed to come up, and this is based on acceleration, not motor torque (so downhill should not confuse the car). Maybe the cars I have been observing predate this requirement, or maybe they just violate it / are broken. Or maybe the lower "false positive" threshold is too low. Unnecessarily activating brake lights in normal highway / road traffic cause serious concern. That's driving where brake lights should basically never activate - their only meaning should be, "someone is stopping to turn" or "deer jumped in the front". Only 0.1% minority of nutjobs / psychopaths used to randomly press their brakes in the good old days without reason. This was not a significant enough problem. Now every random grandma driving their new shiny KIA EV causes constant random brake light blink noise; others learn not to react to brake lights anymore. And that is bad. Ignoring brake lights adds another 2-3 seconds of delay in cases where someone truly brakes. This is why I always try to routinely react to brake lights by slowing down myself - that propagates the need for attention to the drivers behind me. But lately I can't do that anymore because now I'm responsible of propagating the false signal. Earlier that was only a problem with rare sociopaths on the road. Now it happens all the time.

[tom66]

Also, brake lights illuminate on regen once over a certain deceleration threshold.

European spec;  USA/elsewhere may be different:

• Below <0.7m/s^2 there may be no brake light signal
  
• Between >0.7m/s^2 and <1.3m/s^2 the brake light signal may be generated
  
• After ≥1.3m/s^2 the brake light signal must be generated

Arguably this is better than conventional cars, where brake lights illuminate even if you hover over the pedal.

It's still problematic. When driving uphill, the driver may attempt to accelerate but the gas pedal is not pressed deep enough so the slope is still decelerating the car, the brake light may illuminate, confusing the driver behind. The same applies when driving downhill, the regen torque may not be enough to compensate the slope, the car may be accelerating despite the regen power is significant. In this case the brake lights won't illuminate but the car behind it should brake.

I'm fairly sure the lighting rules only apply if some regen is in use, i.e. negative motor power. If driver is going uphill, motor power will be positive unless they are slowing down.

And going downhill with regen on but not decelerating shouldn't illuminate brake lights, because you aren't decelerating.

Anyway, it sounds like you have not driven an EV before.  You don't really have to apply more 'gas' in the same way that you do on an ICE car, the pedal application is roughly constantly unless you wish to accelerate.  So you won't slow down on a hill in the same way that an ICE car does, where the engine needs to put more fuel in under direct command of the user. Obviously the electric motor has to do more work, but this is handled by the drive computer.

[default0.0player]

Exactly - acceleration-based brake light is purely positive with no downsides - it catches those cases where brake light earlier didn't light up when it should - " the driver in front of me is slowing down significantly - need to pay attention". The reason for slow-down is irrelevant. Driver failing to find their gas pedal in steep uphill is a good reason, because usually all other drivers have no such problems, and as such, slow-down is surprising.

Blink mode for long-sustained heavy braking adds even more safety.

Good point, thanks.

The EV makers need to prevent blinking if the deceleration is between >0.7m/s^2 and <1.3m/s^2 to avoid confusing the driver behind as if it's on heavy braking. Although I never personally witnessed that

Anyway, it sounds like you have not driven an EV before.  You don't really have to apply more 'gas' in the same way that you do on an ICE car, the pedal application is roughly constantly unless you wish to accelerate.  So you won't slow down on a hill in the same way that an ICE car does, where the engine needs to put more fuel in under direct command of the user. Obviously the electric motor has to do more work, but this is handled by the drive computer.

Not really, the gas pedal controls the motor torque(at lower speed)/power(higher speed), it does not compensate for the slope. The "coasting point" on one pedal driving mode may vary at different speed but that's a different matter

[Siwastaja]

Anyway, it sounds like you have not driven an EV before.  You don't really have to apply more 'gas' in the same way that you do on an ICE car, the pedal application is roughly constantly unless you wish to accelerate.  So you won't slow down on a hill in the same way that an ICE car does, where the engine needs to put more fuel in under direct command of the user. Obviously the electric motor has to do more work, but this is handled by the drive computer.

You sure? Everything I have read about e.g. Tesla online seems to confirm it's pretty much a torque pedal. LEAF is more like a torque pedal at very low speeds, then power pedal at higher speeds. This is the first time I hear anyone saying that EVs have "speed pedals".

[thm_w]

You sure? Everything I have read about e.g. Tesla online seems to confirm it's pretty much a torque pedal. LEAF is more like a torque pedal at very low speeds, then power pedal at higher speeds. This is the first time I hear anyone saying that EVs have "speed pedals".

Yes torque pedal, with variation based on speed: https://www.reddit.com/r/teslamotors/comments/inzuhz/model_3_factfinding_accelerator_pedal_position_vs/

Leaf should be similar, you can see the mapping here: https://ecutek.atlassian.net/wiki/spaces/SUPPORT/pages/1663500289/Nissan+Leaf+-+Gen2+2013+-2018+Tuning+Guide#Accelerator wonder how hard it is to remap it to be B mode by default..

[tom66]

Anyway, it sounds like you have not driven an EV before.  You don't really have to apply more 'gas' in the same way that you do on an ICE car, the pedal application is roughly constantly unless you wish to accelerate.  So you won't slow down on a hill in the same way that an ICE car does, where the engine needs to put more fuel in under direct command of the user. Obviously the electric motor has to do more work, but this is handled by the drive computer.

You sure? Everything I have read about e.g. Tesla online seems to confirm it's pretty much a torque pedal. LEAF is more like a torque pedal at very low speeds, then power pedal at higher speeds. This is the first time I hear anyone saying that EVs have "speed pedals".

I guess it's complicated, but it's probably best described as speed-dependent torque control.

The inverter is ultimately given a torque command, but the pedal is mapped to have a dead zone around maintaining constant speed with positive and negative torque either side of that.  If you keep constant pressure it's kind of like standard cruise control, with the motor outputting the required torque to maintain speed.  Both the ID.3 and the Polestar I drive do not require any additional "gas" to go up a hill.  You do need more to set off from a hill though, because the motor speed is not constant there.

Pure torque control would be harder to drive.  At low speeds you would need to push the pedal really hard to climb hills because you need that torque.  That would feel wrong, IMO.  Even ICE cars have some kind of mapping there.

I imagine there's a lot of secret sauce around making accelerator pedals feel different too.  I used to drive a Golf GTE and that had a sport mode.  In truth the car produced the same power regardless of whether it was in sport or normal hybrid mode, but the accelerator pedal became a lot more "live" when in this mode, which just gives you the impression of a faster car, even if it isn't actually any faster.  Strange psychology that. 

[jonovid]

looks like form over function leading to poor door release designs.
car safety design engineers been replaced by car design cad artists!
at the end of the day electric or otherwise vehicles are functional machines to do a job
not to become a deathtrap in a minor accident because the sophisticated all electric doors will not open.

[Siwastaja]

Anyway, it sounds like you have not driven an EV before.  You don't really have to apply more 'gas' in the same way that you do on an ICE car, the pedal application is roughly constantly unless you wish to accelerate.  So you won't slow down on a hill in the same way that an ICE car does, where the engine needs to put more fuel in under direct command of the user. Obviously the electric motor has to do more work, but this is handled by the drive computer.

You sure? Everything I have read about e.g. Tesla online seems to confirm it's pretty much a torque pedal. LEAF is more like a torque pedal at very low speeds, then power pedal at higher speeds. This is the first time I hear anyone saying that EVs have "speed pedals".

I guess it's complicated, but it's probably best described as speed-dependent torque control.

But surely the purpose of the speed as a function input is to avoid having to press the pedal deep down at higher speeds, and vice versa, due to the quadratic aerodynamic loss at high speeds, thus increasing torque need per speed. But it means exactly that - more torque at higher speed.

But to act like a speed regulator - to maintain constant speed when more torque is needed due to uphill in relatively constant speed situation, it would need to do the opposite: more torque at lower speed, so that relatively small difference (drop in speed) would drive the torque up to compensate; or use derivative of speed to add torque. This is all the stuff that cruise control does (P,I,D, maybe various feedforwards). But if the speed is constant when uphill starts, then surely the speed input parameter does nothing to the torque. But to maintain the speed uphill, more torque is needed. Ergo, you need to press the pedal further. Matches my experience on LEAF. What I'm missing?

[tom66]

But surely the purpose of the speed as a function input is to avoid having to press the pedal deep down at higher speeds, and vice versa, due to the quadratic aerodynamic loss at high speeds, thus increasing torque need per speed. But it means exactly that - more torque at higher speed.

But to act like a speed regulator - to maintain constant speed when more torque is needed due to uphill in relatively constant speed situation, it would need to do the opposite: more torque at lower speed, so that relatively small difference (drop in speed) would drive the torque up to compensate; or use derivative of speed to add torque. This is all the stuff that cruise control does (P,I,D, maybe various feedforwards). But if the speed is constant when uphill starts, then surely the speed input parameter does nothing to the torque. But to maintain the speed uphill, more torque is needed. Ergo, you need to press the pedal further. Matches my experience on LEAF. What I'm missing?

You're missing that the pedal has a relationship to torque that is dependent on vehicle speed and load, it's not a simple linear relationship like 50% = 150Nm torque.  Even the LEAF has some torque-speed mapping like thm_w mentions... but I suspect there is no consistency between manufacturers, since there is no official standard over what this should look like (other than more pedal = more acceleration, I guess).

[default0.0player]

But surely the purpose of the speed as a function input is to avoid having to press the pedal deep down at higher speeds, and vice versa, due to the quadratic aerodynamic loss at high speeds, thus increasing torque need per speed. But it means exactly that - more torque at higher speed.

But to act like a speed regulator - to maintain constant speed when more torque is needed due to uphill in relatively constant speed situation, it would need to do the opposite: more torque at lower speed, so that relatively small difference (drop in speed) would drive the torque up to compensate; or use derivative of speed to add torque. This is all the stuff that cruise control does (P,I,D, maybe various feedforwards). But if the speed is constant when uphill starts, then surely the speed input parameter does nothing to the torque. But to maintain the speed uphill, more torque is needed. Ergo, you need to press the pedal further. Matches my experience on LEAF. What I'm missing?

As I explained before in this replyhttps://www.eevblog.com/forum/chat/tesla-other-evs-and-hybrids-are-dangerous!/msg6236591/#msg6236591, the motor controller + motor can be seen as a synchronous buck plus a DC motor. To control the motor speed, you adjust the duty cycle of the "synchronous buck". Since the inverter output/motor input voltage is battery voltage X duty cycle. The theoretical, unloaded speed is the vehicle's top speed X gas pedal position. Control the voltage to control the speed.

However, this sort of control is not suitable for EV and only used in low-cost e-bikes or e-scooters. An EV's motor has very high efficiency, to get that, the resistance of the motor winding has to be made very small. If the throttle is a little bit higher the torque will be very high, because the torque is proportional to the electric current, and electric current is (battery voltage X duty cycle - back-EMF) /  resistance. In fact for most high power traction motor application. The battery voltage X duty has to be very close to the back-EMF to prevent too much current from damaging the motor.

Therefore the control logic is use current control to control torque. When the vehicle is coasting, the battery voltage X  duty is exactly equals to the motor's back-EMF. We can describe it using a linear function. Assume the motor's back-EMF per speed is A, the throttle/target current is B. The duty cycle is A X speed + B, regen is when B is negative. When the speed is very high, |B| is limited, the duty cycle is controlled even closer to the speed, to prevent the power from going over the limit.

Cruise control or any other of hill assist that use the speed as the target, the outer loop is speed, but the inner loop is still the current/power

[NiHaoMike]

Since I learned to drive a gas car (as I expect pretty much everyone did except possibly a few very young ones here), I would allow the car to slow down going uphill to avoid it going well above the speed limit when going back downhill. (That's for the usual up and downhills in most areas, not mountain driving which requires downshifting.) Going to a hybrid, that remained mostly unchanged except Toyota hybrids have indicators so I use that instead of simply going by feel.

Haven't switched to full EV yet but I don't see why it would be any different. Is it simply that the "unlimited" regen capacity of an EV means there's not much downside to maintaining speed going up hills and using regen to stay within the speed limit going down the hill?

[Someone]

Haven't switched to full EV yet but I don't see why it would be any different. Is it simply that the "unlimited" regen capacity of an EV means there's not much downside to maintaining speed going up hills and using regen to stay within the speed limit going down the hill?

Many many different balances in play here, the additional speed up the hill adds V³ losses...   but in ICE engines those additional losses can be offset or overtaken by increased engine efficiency (https://en.wikipedia.org/wiki/Brake-specific_fuel_consumption) at the higher load, particularly uphill where the potential energy gain is also adding in. Synthetic testing for fuel efficiency standards have made those engines tuned for 80-100km/h operation. Many petrol/diesel cars are going to be more efficient keeping to the speed limit up the hill and then engine braking on the way down (rather than gaining speed coasting from a lower starting speed).

EV is likely to be much less variable in efficiency vs load, I'd guess least energy would be keeping as slow as practical at all times.

[Siwastaja]

But surely the purpose of the speed as a function input is to avoid having to press the pedal deep down at higher speeds, and vice versa, due to the quadratic aerodynamic loss at high speeds, thus increasing torque need per speed. But it means exactly that - more torque at higher speed.

But to act like a speed regulator - to maintain constant speed when more torque is needed due to uphill in relatively constant speed situation, it would need to do the opposite: more torque at lower speed, so that relatively small difference (drop in speed) would drive the torque up to compensate; or use derivative of speed to add torque. This is all the stuff that cruise control does (P,I,D, maybe various feedforwards). But if the speed is constant when uphill starts, then surely the speed input parameter does nothing to the torque. But to maintain the speed uphill, more torque is needed. Ergo, you need to press the pedal further. Matches my experience on LEAF. What I'm missing?

You're missing that the pedal has a relationship to torque that is dependent on vehicle speed and load,

No, read again. I said there is relationship, but for what you say to be true about uphills, the relationship should be the opposite way - given constant pedal position (e.g. 50%), more torque at lower speed. Read my message again and then explain what I am missing?

Your central claims are:
* ~Constant pedal position
* ~Constant speed

The only logical conclusion is that uphill, torque needs to increase. That can only happen via sensing the slope (unlikely?), or detecting small error in speed (the small drop) increasing torque as a response - but that would be exactly the inverse mapping than online sources say about speed-torque relationship in the pedal mapping (which is, as many sources say, more torque at higher speed, not less). That would be a pure speed pedal - pedal position effectively setting cruise control setpoint, with fixed torque limits you can't control. That would be an undrivable car.

That leaves my claim, driver pressing the pedal down to maintain speed like always before, but that's exactly what you are saying is not the case.

So what I am missing? My theory is that the cars you mention have just relatively lot of power and torque available, and snappy pedal response, thus, the adjustments you need to do in moderate uphills are small enough you just do it subconsciously. Or, you just drive with cruise control mostly so do not notice. And when driving without cruise control, you would do it in urban environments e.g. with intersections which means you would usually also want to slow down going uphill (e.g., visibility reasons) and that happens naturally.

[Marco]

Target speed and speed error are two different variables.

Also they probably take into account derivatives of the pedal position.

[Siwastaja]

Target speed and speed error are two different variables.

Also they probably take into account derivatives of the pedal position.

Target speed and derivative of setpoint are concepts of cruise control. I don't think any car manufacturer has a concept of "target speed" in manual throttle mode (drive by pedal), but maybe I'm wrong. If there is, then tom66's observations would make sense.

[tom66]

All I can say is both my e-cars drive that way.  I have not driven a Leaf, so could not comment on anything other than what is publicly provided about that car.  I have not tried to climb mountains, I'm talking normal slopes like you find on highways and regional roads.

For what it's worth my hybrid ICE did not behave in the same way - you had to apply more throttle to go up hills.  I suspect this was a fuel economy decision, since obviously increasing engine load means you burn more fuel, so by forcing you to press the pedal harder they game the economy slightly.

[Siwastaja]

All I can say is both my e-cars drive that way.  I have not driven a Leaf, so could not comment on anything other than what is publicly provided about that car.  I have not tried to climb mountains, I'm talking normal slopes like you find on highways and regional roads.

For what it's worth my hybrid ICE did not behave in the same way - you had to apply more throttle to go up hills.  I suspect this was a fuel economy decision, since obviously increasing engine load means you burn more fuel, so by forcing you to press the pedal harder they game the economy slightly.

So, I really suspect just: having a lot of power, sensitive pedal, subconscious unnoticed adjustment of the pedal uphill or slight slowdown on slopes you don't notice because you are concentrated on traffic, because you really can't support your observation with hard facts. Surely if I compare LEAF to older gasoline cars I have driven, going uphill needs much less extra throttle, and the reason is simply better availability of torque. Especially some normal 1.6-liter family car with manual stick in 5th gear has very little extra torque available to begin with, requiring significant extra movement on the pedal, like, from 30% position easily to 70% position (going even further is likely diminishing returns on torque and could push efficiency down, prompting shifting to 4th gear instead; I don't really know, just assuming). With an EV running at 1/6th of the available torque, the same effect could be from 30% to 35% position. I'm probably noticing it more because I use LEAF's ECO mode all the time, I don't like modern day super sensitive snappy pedals. The "default" mode tries to pretend being sporty, but just leaves a dead zone, because while 160kW is decent power, it isn't a sports car.