Does using the accelerator to hill-hold harm the motor?

[Ulmo]

I like driving very smoothly when there are passengers in my S, especially when stopping at a light. I obsess on bringing the car to a perfectly smooth stop without feeling the suspension jerking back. It's really hard to do it when stopping uphill though so I sometimes use the accelerator to keep the car at 0 speed. I know that with a small toy-sized motor if you hook it up with a battery but don't allow it to spin, it'll burn out, so that leads me to wonder if doing so with a Tesla motor also harms it. What do you guys think?

One thing I liked about the 2005 Mercedes Benz E500 (with SBC) that I used to drive was that the computer helped me come to a smooth stop without even having to adjust my pressure on the pedal. It's as if the car knew where I wanted to stop, and just stopped there, very quietly and precisely (one poster here called it the Chauffeur's Stop; I think in the MB, it was extremely close to a Chauffeur's Stop, but not quite, and actually the computer made it impossible to do the Chauffeur's Stop (it would just go haywire if you tried)). I miss that crisp yet smooth German precision. American cars are squishy and completely undisciplined, feeling awful and going all over the place, very difficult to control the precise stop without that jerk and usually stopping in the wrong location, and the Tesla is somewhere in between that and the MB. I've taken to giving up in the Tesla, preferring to just go full speed until the stop and just slamming on the brakes then. Might as well: it feels the same, if not better. The only three options in the MB were an SBC antilock brake stop, a tire-squealing stop (you'd have to have your windows open to hear the squealing -- I've been pulled over for stopping like that in my MB before I knew it made a noise, because I thought no one would notice since it was so quiet inside), or the precise smooth stop the computer did when driving normally, none of this jerkiness we get from normal Tesla stops (or worse, American cars, as well as the Japanese cars I've driven).

It's the type of thing an unwealthy person like me assumes the Bentley would have, but since I don't know, maybe they never added it. I also have no idea if Jaguar has it. The brits seem kind of clueless sometimes, and clued other times. Tesla seems too distracted to care. It's something I'd expect to see on the Lucid Air, but they might have forgotten.

 

[Zetopan]

One point here, in hill hold mold there is no current going to the motor as it is using the friction brakes to hold the car from rolling. As soon as you apply pressure to the accelerator it releases the friction brakes.

Sorry, I should not have used "hill hold mode" in such an ambiguous manner. I was specifically referring to the original poster's method of holding a static position on a hill by using the accelerator rather than by using Tesla's built-in "hill hold mode".

 

[Ulmo]

Sorry, I should not have used "hill hold mode" in such an ambiguous manner. I was specifically referring to the original poster's method of holding a static position on a hill by using the accelerator rather than by using Tesla's built-in "hill hold mode".

Hill Hold seems to use both friction brakes and motors in my Tesla. There's a slight force from the motors, and a slight energy applied to the motors from the energy meter in the dash. I realized this when I found out my car slips sometimes during hill hold even though I was on a flat surface: the motors are pushing while the brakes are holding still, and sometimes, the brakes loose for a little bit and the car rolls forward at a stop light or stop sign, but not very far.

 

[Zetopan]

I don't think you have considered the current going through them and the rotor. I am no way a motor expert, but here is how I see it.

You should have stopped right there and considered looking up what you did not know (hint, I have been designing with electric motors for decades).

The phases current of the stator is set to some value and is driven in synchrony. Currents is induced in the rotor creating the counter magnetic field causing it to rotate with the phase of the stator.

NO! In order for any current to be induced into the rotor it CANNOT stay in phase with the rotation of the stator field. When the slip is zero (the rotor speed exactly matches the rotation of the stator magnetic field) the torque is zero since no currents can be induced in the rotor since it is not being subjected to changing magnetic fields that induce currents. Hence, to produce any torque in an induction motor, *by definition* the rotor has to be turning at a different speed than the stator field is rotating (i.e. the slip factor has to differ from zero).

In your hypothetical example the rotor catches up with the stator field rotation and at that point the motor is producing zero torque. Unless the vehicle was situated on a hill whose slope exactly matches the rolling resistance and drag so that it could coast down the hill at a constant speed that exactly matched the stator field rotation, non-zero torque is required to continue moving the vehicle.

What you may be thinking of is when accelerating from a stop. In that case the steady state torque required after a specific velocity is reached would generally be less than the torque required to accelerate from the stop. The initial acceleration requires more torque so the rotor slip factor will be higher (higher induced currents) than when a steady state speed is reached, where the slip factor reduces (lower induced currents). The only time that the rotor slip factor reaches zero is when coasting to a stop (regen off) while simultaneously letting up on the accelerator pedal at just the right rate to match the stator field rotation with the rotor speed, or else when coasting down a hill whose slope exactly overcomes all frictional and drag losses. No accelerator pedal changes are required for the latter case since the velocity remains constant.

 

[S4WRXTTCS]

Meanwhile I have to remind passengers to put their heads on the back of the seat just before it's go time.

 

[Zetopan]

Hill Hold seems to use both friction brakes and motors in my Tesla. There's a slight force from the motors, and a slight energy applied to the motors from the energy meter in the dash. I realized this when I found out my car slips sometimes during hill hold even though I was on a flat surface: the motors are pushing while the brakes are holding still, and sometimes, the brakes loose for a little bit and the car rolls forward at a stop light or stop sign, but not very far.

Are we even sure that the Tesla "Hill Hold Mode" uses the brakes to hold position? Since this could be done totally with the drive motors (which also matches the behavior that you have stated) I would suspect that the brakes are actually not being used at all. If the brakes were being used there is no need for any motor torque. So how could the vehicle slip during "Hill Hold Mode" when the brakes are sufficient to stop a 5,000 pound vehicle from 100 MPH to zero in about 4.2 seconds. On the other hand, an error in the "Hill Hold Mode" software or a faulty speed sensor could result in some amount of slip.

 

[Saghost]

Are we even sure that the Tesla "Hill Hold Mode" uses the brakes to hold position? Since this could be done totally with the drive motors (which also matches the behavior that you have stated) I would suspect that the brakes are actually not being used at all. If the brakes were being used there is no need for any motor torque. So how could the vehicle slip during "Hill Hold Mode" when the brakes are sufficient to stop a 5,000 pound vehicle from 100 MPH to zero in about 4.2 seconds. On the other hand, an error in the "Hill Hold Mode" software or a faulty speed sensor could result in some amount of slip.

Need to be careful with the terminology here since we have a few different modes on a couple different system architectures.

Hill hold is a feature of pre AP cars, and from my limited experience with it I believe can only be applied if the car detects a noticeable incline, and is limited to a few seconds.

Vehicle Hold is a feature of AP cars, which can be engaged at any time the car is stopped and lasts for up to ten minutes, at which point the car seamlessly shifts to park.

There's also a Hold mode built into Autopilot in some circumstances which is something yet different.

There's no question Vehicle Hold is using the brakes - in addition to be explicitly described as such in the owner's manual, the brake pedal stays depressed while the mode is engaged.

I believe that the momentary "slipping" while it is engaged is deliberate - at least what I've felt is the car either releasing one end at a time or almost releasing and regripping, both likely ways to relieve the suspension tension resulting from a harder stop.

I don't know for certain how Hill Hold works in older cars, but the description suggests it uses the brakes and I know the competition does it that way as well.

 

[daniel]

Fascinating discussion! I'm learning stuff.

... American cars are squishy and completely undisciplined, feeling awful and going all over the place, very difficult to control the precise stop without that jerk and usually stopping in the wrong location...

My cars always stop in just the right location. I apply the brake as needed, and accept the fact that there will be a slight jerk at the end.

Meanwhile I have to remind passengers to put their heads on the back of the seat just before it's go time.

I never do that. Mwahahaha.

 

[No2DinosaurFuel]

NO! In order for any current to be induced into the rotor it CANNOT stay in phase with the rotation of the stator field.

Don't be so negative. Note I said "phase current," not "phase." Totally different things.

What you may be thinking of is when accelerating from a stop. In that case the steady state torque required after a specific velocity is reached would generally be less than the torque required to accelerate from the stop. The initial acceleration requires more torque so the rotor slip factor will be higher (higher induced currents) than when a steady state speed is reached, where the slip factor reduces (lower induced currents). The only time that the rotor slip factor reaches zero is when coasting to a stop (regen off) while simultaneously letting up on the accelerator pedal at just the right rate to match the stator field rotation with the rotor speed, or else when coasting down a hill whose slope exactly overcomes all frictional and drag losses. No accelerator pedal changes are required for the latter case since the velocity remains constant.

Ok so how are we in disagreement? If I am interpreting you right, you are saying the induced current is higher like accelerating from stop which is like the motor stalled, though maybe with lower power in this specific case.

You should have stopped right there and considered looking up what you did not know (hint, I have been designing with electric motors for decades).

OK... so I'll stop because we are obviously in the presences of someone who designs "with" motors... I suspect you did not think of the induce current both on the phase as well as the rotor as a possible issue. Initially I agreed with you after reading your first post, but then I thought about the energy argument and then tracked down where all that power went.

Prof. Knowitall challenges all the Real Nerds to calculate the power necessary to maintain the position on the hill @Zetopan describes. Remember that power and energy are different and pounds are really a measure of force not mass.

I would take on your challenge, but I have better use of my time. What I will do is point you in the right direction.
1) Gravity * mass of car is the force you need to overcome
2) This force induces a torque on the 4 wheels, but we can simplify by imagining only 1 wheel with 1 torque is the sum of the 4 wheels. We can obtain the torque because we have the radius of our 4 wheels.
3) This is the torque you need to counter
4) Ignore gearing ratio for now because we can always scale with some friction factored in
5) Torque is a function of the voltage and duty cycle of the phase current in the stator assuming fixed phase frequencies. Higher voltage just means higher current flow when the PowerFET is open or higher instantaneous torque which will average out as higher average torque. Similarly longer duty cycle means longer constant torque on which can be average out to be higher average torque.
6) Assuming the drive unit changes the torque using duty cycle on the phase current instead of changing the battery voltage, then we just need to calculate the duty cycle @ a fix phase frequency to produce the counter torque required to counter the torque we got in 3.
7) Typically a drive unit has a max current draw in its specs, we can simplify by assuming max continuous current draw is 100% duty cycle. Therefore, our duty cycle is the percentage of the max current the drive unit can draw from the battery.
8) Multiply percentage current draw by the battery voltage and you get your power required to keep the car stationary on an incline which you can probably easily get by experiment by actually doing it in your car.

This is definitely an interesting topic to think about. Obviously this is not the optimal way to keep an object stationary on an incline. There are countless other sub-optimal ways to keep the object stationary on an incline like using thruster attached to the car. Or using a petrol engine and riding the clutch. But the optimal is to use 0 power by applying your brakes.

Again like I said before, you should think of the system as a whole. There is definitely some power required to keep the car stationary on an incline if you are solely using your motor. You can verify this by looking at the power meter on your dash. Since you are not rotating/moving/doing work. This power has to go somewhere. Power is not dissipated through a loud buzz or what not so it has to turn into heat. Heat that can damage the drive unit and motor if the duration and power is long enough. But I am sure tesla engineers have all sorts of protections built into the drive unit and motors to prevent this.

 

[mal_tsla]

My understanding is that the motor holds the car still by slowly rotating three slightly out of phase magnetic fields around the rotor so as to generate the torque necessary. Thus, there's not stalling in the motor like you would have with an old school DC motor, even when the car is at 0 mph on an incline via the accelerator.

In my observation on slopes near my home (I tried this yesterday), I can barely register any current on the energy meter on the dash while holding the car on a hill with the motor. I can register more current by turning on BioWeapon Defense mode.

I think this is a non-issue in any case, due to the low power needed, but becomes an interesting thought experiment if you imagine extreme slopes that would require more power. Even in those cases, I think because of the way the three phase induction motor works, it doesn't create wear different than driving the car beyond the fact that holding high amounts of current normally results in running out of road, going too fast and going to jail, or hitting the top speed of the car.

I can imagine hypothetical scenarios where on a ridiculously steep hill you could hold high current usage longer than is practical in driving, and I would expect the car to behave just like it does at the race track in that situation -- it'll enter a thermal protect mode where power is limited. If the slope were steep enough, you'd find yourself unable to hold on the hill anymore with the accelerator and you'd activate the brakes until the system cooled.

Again, none of this should ever happen in real life... but it would be interesting to know the "hold times" to cause this situation to occur for various slopes. I imagine it approaches "infinite time" until you get to crazy slopes (45 degrees+?)

 

[Saghost]

My understanding is that the motor holds the car still by slowly rotating three slightly out of phase magnetic fields around the rotor so as to generate the torque necessary. Thus, there's not stalling in the motor like you would have with an old school DC motor, even when the car is at 0 mph on an incline via the accelerator.

In my observation on slopes near my home (I tried this yesterday), I can barely register any current on the energy meter on the dash while holding the car on a hill with the motor. I can register more current by turning on BioWeapon Defense mode.

I think this is a non-issue in any case, due to the low power needed, but becomes an interesting though experiment if you imagine extreme slopes that would require more power. Even in those cases, I think because of the way the three phase induction motor works, it doesn't create wear different than driving the car beyond the fact that holding high amounts of current normally results in running out of road, going too fast and going to jail, or hitting the top speed of the car.

I can imagine hypothetical scenarios where on a ridiculously steep hill you could hold high current usage longer than is practical in driving, and I would expect the car to behave just like it does at the race track in that situation -- it'll enter a thermal protect mode where power is limited. If the slope were steep enough, you'd find yourself unable to hold on the hill anymore with the accelerator and you'd activate the brakes until the system cooled.

Again, none of this should ever happen in real life... but it would be interesting to know the "hold times" to cause this situation to occur for various slopes. I imagine it approaches "infinite time" until you get to crazy slopes (45 degrees+?)

The details of the cooling system may matter for your gedankenexperiment. If rotor movement is used to pump coolant or aid cooling, the stalled case won't get that extra cooling.

What might be worse is if the coolant is delivered from a port on the side of the rotor, so you get local hot/cold zones.

As you said, the power involved in real life circumstances is trivial enough that I'm sure it's a non issue, but in the theoretical limits...

 

[_jal_]

6) Assuming the drive unit changes the torque using duty cycle on the phase current instead of changing the battery voltage, then we just need to calculate the duty cycle @ a fix phase frequency to produce the counter torque required to counter the torque we got in 3.
7) Typically a drive unit has a max current draw in its specs, we can simplify by assuming max continuous current draw is 100% duty cycle. Therefore, our duty cycle is the percentage of the max current the drive unit can draw from the battery.
8) Multiply percentage current draw by the battery voltage and you get your power required to keep the car stationary on an incline which you can probably easily get by experiment by actually doing it in your car.

It is my understanding that the inverter for Teslas, like most inverters for driving induction motors, modulate frequency and change wave amplitude to control speed. They do not rely on duty cycle or pulse width modulation or whatever to adjust power. This is 10 years old, but at least at that time this was the state of the art for inverters.

Induction Versus DC Brushless Motors

I'm not sure how to solve my challenge other than holding it on the hill and measuring the waveform from at least one leg into the motor. Because we don't know a LOT of variables. The more conductive the rotor is, for instance, the less current needed because the induced current will be higher in the rotor and thus it will produce more magnetic field to couple with the stator's field. But there is a current, and it's not a superconductor, so there will be heat. It'd be interesting to calculate.

You could also weigh the car before and after holding for so long on the hill, and the difference will be the change of the quantum state of the lithium ions. Then a simple E=mc^2 will give you energy consumed and then you can subtract use of electricity from the radio you were listening to and seat adjustments and then assume the rest is lost to heat. You'll need a really really accurate scale though. Really accurate

 

[No2DinosaurFuel]

It is my understanding that the inverter for Teslas, like most inverters for driving induction motors, modulate frequency and change wave amplitude to control speed. They do not rely on duty cycle or pulse width modulation or whatever to adjust power. This is 10 years old, but at least at that time this was the state of the art for inverters.

Induction Versus DC Brushless Motors

I'm not sure how to solve my challenge other than holding it on the hill and measuring the waveform from at least one leg into the motor. Because we don't know a LOT of variables. The more conductive the rotor is, for instance, the less current needed because the induced current will be higher in the rotor and thus it will produce more magnetic field to couple with the stator's field. But there is a current, and it's not a superconductor, so there will be heat. It'd be interesting to calculate.

You could also weigh the car before and after holding for so long on the hill, and the difference will be the change of the quantum state of the lithium ions. Then a simple E=mc^2 will give you energy consumed and then you can subtract use of electricity from the radio you were listening to and seat adjustments and then assume the rest is lost to heat. You'll need a really really accurate scale though. Really accurate

In actuality the waveform are digitally generated by pulse width and averaged out. The inverter generates the 3 sine wave of the induction motor using pulse width because they use PowerFETs and FETs are best used when they are on or off. Through a series of smoothing circuit as well as the load the output will be averages out to a sine wave.

What I meant by duty cycle is the average duty cycle of all the 3 phases. I am trying to keep it simple and not get bog down on the details of measuring the impulses and accumulating the work and dividing by time, etc. to get power.

 

[HankLloydRight]

 

[_jal_]

In actuality the waveform are digitally generated by pulse width and averaged out. The inverter generates the 3 sine wave of the induction motor using pulse width because they use PowerFETs and FETs are best used when they are on or off. Through a series of smoothing circuit as well as the load the output will be averages out to a sine wave.

What I meant by duty cycle is the average duty cycle of all the 3 phases. I am trying to keep it simple and not get bog down on the details of measuring the impulses and accumulating the work and dividing by time, etc. to get power.

Got it. I was a step behind you in that at the end of the day the coils gets a sine wave, not a pulse width modulated sine wave. In other words, just looking at the wave at the motor you wouldn't know that there was any modulation happening. The power is modulated in the inverter - not to the actual motor leads. When I think of duty cycle I think of the product of the wave and a unit impulse function of whatever width. Short of getting at the internals of the inverter, I don't know how you'd measure that to determine the power used to keep the car still. We might never know! Then what?

 

[Zetopan]

In actuality the waveform are digitally generated by pulse width and averaged out. The inverter generates the 3 sine wave of the induction motor using pulse width because they use PowerFETs and FETs are best used when they are on or off.

Close, but not quite correct. Most high power inverters (including Tesla) use IGBT (Insulated Gate Bipolar Transistors) which are not quite as fast as power FETs but are more efficient at high current levels, can handle far higher voltages, while also having a huge SOAs (Safe Operating Areas). You can view this semiconductor as a hybrid between a FET and a BJT (Bipolar Junction Transistor) with significant advantages over both. These inverters generally switch at supersonic frequencies and produce essentially full amplitude square pulses on the motor stator windings. The self inductance of the motor stator windings does the filtering to produce the lower frequency sine wave currents that are needed to drive the 3 phase stator windings. These inverters are also technically know as class D amplifiers since the outputs are either ON or OFF, with a variable duty cycle to control the actual current amplitude in most commonly inductive loads. Resistive loads can also be used, as for example in heating applications like kilns and furnaces.

 

[evp]

I was going to add, I think the concern is that when you stall the EV motor by holding it in one position, the stator is energized with the rotor in one fixed position, pushing current through the same set of stator coils, which are usually only energized for short moments when the motor is normally rotating, generating more heat in those coils than normal.

I think this is incorrect. An induction motor needs to move the external field a little faster than the rotor's rotation in order for the rotor's coils to see a moving magnetic field which creates the current in the rotor coils that in turn reacts against the external field to produce force. (This is called "slip".) If the rotor were superconducting, this would not be true, but replenishment of that current is necessary due to resisitive losses. Therefore, even if the rotor is motionless, the controller would be (slowly) rotating the external field to apply torque.

 

[HankLloydRight]

Yes, we established I was indeed misinformed yesterday. Do we need to rub it in???

 

[Ulmo]

Need to be careful with the terminology here since we have a few different modes on a couple different system architectures.

Hill hold is a feature of pre AP cars, and from my limited experience with it I believe can only be applied if the car detects a noticeable incline, and is limited to a few seconds.

Vehicle Hold is a feature of AP cars, which can be engaged at any time the car is stopped and lasts for up to ten minutes, at which point the car seamlessly shifts to park.

There's also a Hold mode built into Autopilot in some circumstances which is something yet different.

There's no question Vehicle Hold is using the brakes - in addition to be explicitly described as such in the owner's manual, the brake pedal stays depressed while the mode is engaged.

I believe that the momentary "slipping" while it is engaged is deliberate - at least what I've felt is the car either releasing one end at a time or almost releasing and regripping, both likely ways to relieve the suspension tension resulting from a harder stop.

I don't know for certain how Hill Hold works in older cars, but the description suggests it uses the brakes and I know the competition does it that way as well.

Can you believe I never read the Tesla car manual for my car, and don't know what the H with a circle around it is called for my car officially? I have a November 2016 AP2 car. By your description, it might be called "Vehicle Hold".

 

[Shock-On-T]

To clarify - I do have vehicle hold, but there's still a slight (I'm being very picky here) jerky motion when using the brake pedal to stop on an upward hill. I have to time it so that the instant (H) activates, I'm neither going forward nor rolling backwards - the latter gives a worse experience. It's more of an obsession than necessity

I totally understand, I do the exact same thing.