What's wrong with the Tesla regen? Or my car? (Chart)

[McRat]

Yes, cars can generate power below 5 mph, that is at least 7 year old tech.

Linear load is not how a normal brake pedal feels. The energy needed to stop a car is based on the square of the velocity, but the amount of braking force with a constant pedal pressure drops in a linear fashion. 40 mph you are sweeping twice the surface area per second as 20 mph, but the energy you need to burn off is a square root function.

A great amount of work goes into the EV braking 'feel' to mimic the feel of ICE cars. Getting a test drive in a car when the brakes, accelerator, or steering feels unpredictable is not a selling point.

...
I recently took a chevy Bolt on a test drive and I really hated the Regen braking on that thing. Unlike other EV's in order to get regenerative braking on the Bolt I had to steps on the brake pedal. It was really odd. Of course the guy on the test drive had no clue if there was a mode to change that. But the Bolt forced me to drive with 2 pedals. When I pressed the brake the software in the car makes the decision if the motor or the brakes are being used. I guess they did it to allow the car can coast and work like a regular gas car and reduce the learning curve.

Yes, the Bolt will mimic a gas car with an automatic or do the EV stop trick.

For test drives, they should set the car to ICE style, unless the customer specifically asks for EV style. No sense in making the customer feel uncomfortable immediately.

 

[tedsk]

i3 is just doing some cleaver programming

Tesla is one big piece of clever programming. Why can't they do the same I wonder.

on the i3 you never have access to more than 80% of the battery available so the regen braking is always consistent no matter the charge.

That is not quite correct. Indeed i3 doesn't allow charge above 80%. Point is it doesn't allow to go over 80% even with regenerative braking. Otherwise owner living uphill could easily break the limit. In that matter i3 behavior is not consistent with regards to charge. If i3's battery is completely (80%) charged car is not doing regen but simulates it with friction brakes. Nice idea but it feels differently - I could easily tell.

I believe the bump in deceleration in phase 2 is caused by the front motor regen kicking on.

Not in my case - I have RWD.

The way I drive the Model S, I am able to do single pedal driving as well. Coming from a Leaf, I learned to pay attention much further out on what's happening (light changing, cars slowing down) to where I do not have to constantly go back and forth on pedals. Even in stop and go traffic, I time my acceleration to match the traffic conditions and do not have to actually brake. There are the occasional unavoidable cases where traffic stops longer than I anticipated but for the most part, I am mostly doing 1 pedal driving except at lights and stop signs.

In my opinion, I think this approach is better as it teaches me to avoid unnecessary acceleration where I need to depend on harder regen that requires braking.

I also practice 1-pedal driving in my Tesla and it goes well but it is still impossible to stop completely without moving foot and press conventional brake pedal. I disagree with you on what approach is better. Let the customer decide which is better - it should be an easy software fix, no hardware change.

 

[tedsk]

Linear load is not how a normal brake pedal feels. The energy needed to stop a car is based on the square of the velocity, but the amount of braking force with a constant pedal pressure drops in a linear fashion. 40 mph you are sweeping twice the surface area per second as 20 mph, but the energy you need to burn off is a square root function.

It is good remark that kinetic energy depends on speed squared. Thing is acceleration depends directly on force (remember F=m*a?) which is linear from regen power (which should be constant) assuming no significant friction and air resistance is going on (lets say speed below 50mph). So constant regen means constant deceleration without hypes. Catch with non linearity of kinetic energy vs speed is in distance. Work = energy recovered from regen = force by distance (not time!!!). Means that V squared is taken care of smaller and smaller distances being covered at each moment of time.

Linear load is not how a normal brake pedal feels

From my experience it is quite linear - constant pressure on brake pedal means linear deceleration which translates into predictability. You feel how much you should press on brakes to do a complete stop at desired point. Quite like that. Probably cars I had driven so far worked different than yours.

 

[derekmw]

I also practice 1-pedal driving in my Tesla and it goes well but it is still impossible to stop completely without moving foot and press conventional brake pedal. I disagree with you on what approach is better. Let the customer decide which is better - it should be an easy software fix, no hardware change.

Yeah, I wasn't trying to state that I am correct. I am just stating my opinion, the same as you are. As with many things, and as you can see with all the variety in how different manufacturers have chosen to deal with this 'feature' of EV's (I've owned/own Volt, Leaf, b250e and MS), there's no one solution to satisfy everyone. What one person wants, is what another person doesn't.

I do agree that more settings around this would be nice. For example (I haven't really bothered enough to search if there's a setting), the b250e doesn't regen brake nearly as aggressive as the MS. It would be nice if all manufacturers simply had various settings around this. But for me, the MS just happens to be exactly how I like it so it's worked for me.

 

[Pezpunk]

I read somewhere in the TMC that below 5mph, there is no electricity generated so no regen. i3 seems to apply reverse power from motor or friction brakes to bring the car from 5mph to complete stop. At least that's what I heard.

that's interesting ... does that mean it might be possible to reduce stopping distance by augmenting the physical brakes with negative power (pardon my ignorance on the terms) to the motor in some way? granted it would be at the cost of range, but if you are in Ludicrous+ mode, i think it's safe to assume range is not a priority...

 

[JohnnyG]

...

Linear load is not how a normal brake pedal feels...

From my experience it is quite linear - constant pressure on brake pedal means linear deceleration...

I think what @McRat was trying to say is that, for example in a non-EV, at 50MPH apply a specific pressure on your brake pedal. Hold that specific pressure constant until you stop. The resulting lateral g-force at 40MPH will be significantly less than the resulting g-force at 4MPH. This directly relates to the rate of deceleration. This means that the rate of deceleration is not constant, which means that the rate of speed decrease is not linear. The correlation here is that the motor is producing a constant form of resistance until it disengages, giving the same effect. Of course, this is all unproven theory. We'd need more data to validate.

 

[tedsk]

Yeah, I wasn't trying to state that I am correct. I am just stating my opinion, the same as you are. As with many things, and as you can see with all the variety in how different manufacturers have chosen to deal with this 'feature' of EV's (I've owned/own Volt, Leaf, b250e and MS), there's no one solution to satisfy everyone. What one person wants, is what another person doesn't.

I do agree that more settings around this would be nice. For example (I haven't really bothered enough to search if there's a setting), the b250e doesn't regen brake nearly as aggressive as the MS. It would be nice if all manufacturers simply had various settings around this. But for me, the MS just happens to be exactly how I like it so it's worked for me.

Completely agree that there will be always someone unhappy. Can't be nice to everyone.
But! Non-linear non-predictable behavior of braking system is a safety concern not only inconvenience.
I don't see how anyone could be happy with unpredictability and non-linear braking.
Tesla owners are getting used and stop noticing - that is right, I got that. I got used to it myself.
Should it be that way? Don't think so.

 

[tedsk]

that's interesting ... does that mean it might be possible to reduce stopping distance by augmenting the physical brakes with negative power (pardon my ignorance on the terms) to the motor in some way? granted it would be at the cost of range, but if you are in Ludicrous+ mode, i think it's safe to assume range is not a priority...

Nope, not gonna happen. Friction brakes already make everything that is theoretically possible until ABS kicks in.
Adding more stopping power to the wheels will not help (it will block the wheels) - you can't increase friction between tarmac and tire with onboard motor.

 

[tedsk]

The resulting lateral g-force at 40MPH will be significantly less than the resulting g-force at 4MPH.

I see what you are saying but it still a subjective hypothesis.
Physics doesn't follow opinions and gut feeling.
Constant pressure on brake pedal makes constant force on braking pads which translates into constant deceleration rate (lateral g-force as you call it will still be same at 40 and 4 mph).
That assumes that braking system works properly and has linear dependence between brake pedal force and deceleration rate.

Of course we live in much more complicated world and there are lot of factors contributing but in general it works like that.

 

[Pipcecil]

My personal experiences with: owning a Nissan Leaf (2011), wife owning a Chevy Volt (2012), and extensive driving with the i3 (friends and co-works have, plus my work deals with the industry so I am very much exposed - probably drivin most EVs/plugins around actually).

1) Nissan Leaf - One petal driving was pretty impossible in the Nissan Leaf. They made changes in later models that made it easier, but the stopping distance was too much, you pretty much just ride the break. You also get MORE regen if you ride the "break" - but you must learn to ride it just at the max line and not to exceed to allow the friction breaks to kick in - this was easy IMO. You could not achieve max regen with one petal driving alone - you had to push the break peddle to give more regen. Regen coasting was linear

2) Volt - easier to regen coast in the volt - it regened hardered than the leaf. Coasting regen is linear. You could "ride the break peddle" but i found myself coasting more in the volt than in the leaf. Unfortunately the volt's simulated creep is REALLY hard - it creeps alot, something I disliked immensely.

3) i3 - Again linear regen breaking. I have to say while its nicer compared to the leaf and volt for no creep and easier one petal driving and harder regen overall i HATE its simulation of stopping the car. Period. To me, its the same as the leaf and volt doing the opposite and adding creep. STOP FABRICATING THINGS. Uggh... Yea I disliked it that much. Instead of creeping it stops for you - lets do the exact opposite. I really don't like that. I had too many close calls with its "smart" self stopping the car when I totally didn't want to. I found myself accelerating more often because it decided to stop for itself.

With the Tesla I know its gonna just leave it alone and allow the natural momentum of the car to do as is. I like that. Its much more predictable and much easier to learn and drive. No fabrication, no pretending its slowing down and stopping for me or pretending to add an ICE creep simulation to it. Pure and predictable. Also the non-linear breaking is a awesome. It applies the right amount of break and regen at the right amount of speed. The linear regen of the other three made for some hard slowing down when I really didn't want it and felt it was more unsafe.

 

[tedsk]

I disagree with you but thank you for sharing your experience. Sure it is a matter of personal preference.
I will not exchange linear predictable deceleration to what Tesla is doing - it just doesn't work. You will not stop where you want without hitting brakes in Tesla. In i3 it is easy-peasy. Highway - having coasting is good, city - not so much.

Did you try not to let your foot off from accelerator completely but regulate deceleration rate with it in i3? Exactly as you would do with the brake pedal but in opposite direction. It works almost flawlessly. You can even coast in i3 if you want to.

Tesla's regen is a mess, sorry. I hate Tesla one-pedal driving.

 

[tedsk]

My personal experiences with: owning a Nissan Leaf (2011), wife owning a Chevy Volt (2012), and extensive driving with the i3 (friends and co-works have, plus my work deals with the industry so I am very much exposed - probably drivin most EVs/plugins around actually).

1) Nissan Leaf - One petal driving was pretty impossible in the Nissan Leaf. They made changes in later models that made it easier, but the stopping distance was too much, you pretty much just ride the break. You also get MORE regen if you ride the "break" - but you must learn to ride it just at the max line and not to exceed to allow the friction breaks to kick in - this was easy IMO. You could not achieve max regen with one petal driving alone - you had to push the break peddle to give more regen. Regen coasting was linear

2) Volt - easier to regen coast in the volt - it regened hardered than the leaf. Coasting regen is linear. You could "ride the break peddle" but i found myself coasting more in the volt than in the leaf. Unfortunately the volt's simulated creep is REALLY hard - it creeps alot, something I disliked immensely.

3) i3 - Again linear regen breaking. I have to say while its nicer compared to the leaf and volt for no creep and easier one petal driving and harder regen overall i HATE its simulation of stopping the car. Period. To me, its the same as the leaf and volt doing the opposite and adding creep. STOP FABRICATING THINGS. Uggh... Yea I disliked it that much. Instead of creeping it stops for you - lets do the exact opposite. I really don't like that. I had too many close calls with its "smart" self stopping the car when I totally didn't want to. I found myself accelerating more often because it decided to stop for itself.

With the Tesla I know its gonna just leave it alone and allow the natural momentum of the car to do as is. I like that. Its much more predictable and much easier to learn and drive. No fabrication, no pretending its slowing down and stopping for me or pretending to add an ICE creep simulation to it. Pure and predictable. Also the non-linear breaking is a awesome. It applies the right amount of break and regen at the right amount of speed. The linear regen of the other three made for some hard slowing down when I really didn't want it and felt it was more unsafe.

Look at this deceleration rate that Tesla has. It is just stupid.
It has to be damn straight line, not these crappy spikes.
Gosh.

P.S.: This line is exact measure of how you feel deceleration. It is force that stops your body when you sit in your car and let accelerator go.

 

[tedsk]

As I noted previously during the test my regen was limited to around 16kW which is shy 1/4 of full regen power on Tesla. I will redo the test when it is warmer outside and regen is not limited. Probably effect that bothers me disappears when battery is able to accept charge at high rate.

To be honest I notice many technical design flaws in Tesla specific to colder climates. As many folks noted before - car was designed with California climate in mind.

 

[tedsk]

Confirmed - with full regen available when the battery is warm deceleration is almost linear without spikes and gives up between 10-5mph.
There is no hype in deceleration when battery is warm.
This proves one more time that Tesla didn't design MS with cold climates in mind.
Here is my acceleration with following slowdown without friction brakes to a complete stop.

 

[green1]

Others have said in the past that the lack of regen at low speeds in the Tesla is due to the AC motor vs the permanent magnet motors used in other EVs, something about the AC motor not being able to do the regen at low speeds like the permanent magnet motors can.

As for the rest, yeah, until the battery pack warms up (which, in the winter, in city driving, it basically never does) many things on the car are not "right" and regen is the biggest of them.

I also 100% agree about predictability, if your battery can't accept the power due to full battery or cold weather, or anything else, you should either find somewhere else to dump the energy, or simulate that same regen with the friction brakes. Predictability is a safety issue.

 

[Saghost]

I've been driving BMW i3 for couple years and recently traded it for Tesla Model S. Right away I noticed difference in regenerative braking behavior. First Tesla doesn't stop completely like BMW but starts coasting at around 5mph - forces to use friction brakes which wasn't necessary in BMW. Second and more annoying one is non-linear deceleration (it makes it very hard to predict where car will stop and forces me to depress accelerator closer to the end of braking). My butt-dyno was screaming - something fishy is going on at the end off deceleration right before regen gives up at around 10mph. My assumption was that it is software glitch in regen rate limiter and I decided to check exact metrics. It showed that my assumption was incorrect (regen current to battery pack was linear) but butt-dyno didn't lie. So what is the heck going on there?
Need help from gurus and have complete CAN3 log for the measuring period. I highlighted 4 specific periods: 1 - linear deceleration, current (what one may expect); 2 - deceleration kicks-in while regen current stays constant; 3 - regen gives up and car starts coasting at around 5mph; 4 - I hit friction brakes.
What is happening on in period #2? If regen doesn't increase what makes car brake 2 times more heavily?
Legend: Green line is calculated and scaled deceleration rate to fit on the screen, it lags a little since it is 1 second moving average; Yellow - pack current in Amperes; Blue - speed in mph;
View attachment 218492

Remember, a constant rate of acceleration (or deceleration) requires an increasing amount of power at higher speeds. (Acceleration is proportional to torque, horsepower to torque times rpm.)

With regeneration, what you're ding is converting kinetic energy into electricity with the drive motors. The car's kinetic energy is equal to one half of the car's mass times the car's velocity squared.

So the reason that the deceleration increases so much while the regeneration remains constant is physics. Tesla limits the car to a maximum amount of power for regeneration, presumably to protect the inverters or battery pack. At lower speeds, that same amount of power coming out of the car's pool of kinetic energy translates into a steadily increasing rate of deceleration.

My Volt did something similar, but from about 40 mph on down, it locked on to a maximum deceleration rate instead (with steadily decreasing regeneration power as a result,) while Tesla lets the rate increase a ways further to get more energy out of the car.

What you're seeing as region 1 and region 2 is actually a single long exponential curve in deceleration (constant regeneration power,) which is flat enough in the first portion that the aberrations hide the curve.

Note that Tesla also smooths the regeneration input - it takes a second or two for full regen to build if you drop the accelerator but don't hit the brakes, and reduces regen for full or cold batteries.

 

[Saghost]

Confirmed - with full regen available when the battery is warm deceleration is almost linear without spikes and gives up between 10-5mph.
There is no hype in deceleration when battery is warm.
This proves one more time that Tesla didn't design MS with cold climates in mind.
Here is my acceleration with following slowdown without friction brakes to a complete stop.
View attachment 218872

It may look linear to the naked eye, but it isn't. Even a simple straight edge on the screen will confirm your speed curve has a steadily increasing slope in the regen portion, let alone real numerical analysis.

 

[tedsk]

It may look linear to the naked eye, but it isn't

It is very easy to check. Here is amended graph with acceleration (green) and speed (blue).
Indeed deceleration is not linear but it is very close. No spikes and relatively small change.
Interestingly acceleration slows down very abruptly after 42mph - doesn't look like air drag (this component should phase in non-linearly).
Seems like power limit you mentioned.

 

[tedsk]

My measurement is very consistent with typical power-torque curve of electric motor:
Until 42mph motor is torque-limited, after that there is max power limiting torque so acceleration decreases.

 

[Saghost]

It is very easy to check. Here is amended graph with acceleration (green) and speed (blue).
Indeed deceleration is not linear but it is very close. No spikes and relatively small change.
Interestingly acceleration slows down very abruptly after 42mph - doesn't look like air drag (this component should phase in non-linearly).
Seems like power limit you mentioned.View attachment 218891

I'm still wrapping my head around this chart.

It looks like the rate of change in the deceleration is constant or close to it for the whole deceleration period - the deceleration rate increases steadily over time, with about a 50% increase in rate over the course of the cycle.

In another thread we were talking about the character of the acceleration, and yes, it looks like your car hits the battery/inverter power limit at 42 mph, with acceleration suddenly changing from constant to declining at that point.