Model 3 Highland Performance/Plaid Speculation [Car announced 04.23.2024]

[Gauss Guzzler]

Maybe, but the link you posted says exactly what I wrote above - that the front calipers are conventional hydraulic, not electric.

Who says Tesla doesn't allow 200kW regen? I randomly got 172kW on a cold battery so maybe it would allow even more if the battery was willing. But I don't think you'd want that much on pedal-lift anyway, it'd be much better to have modest regen for traffic negotiation paired with major regen for heavy braking. Assuming of course, that the major regen is well blended with good pedal feel and control.

 

[gearchruncher]

But the link you posted says exactly what I wrote above - that the front calipers are conventional hydraulic, not electric.

And the very next line says:

In production, a vehicle could have a split system like this, or it could have either the hydraulic or electric setup at all four corners.

Who says Tesla doesn't allow 200kW regen?

It sure won't on my car. I have hundreds of track datalogs and have never seen anything over 100kw.

 

[cappo3]

Maybe, but the link you posted says exactly what I wrote above - that the front calipers are conventional hydraulic, not electric.

Who says Tesla doesn't allow 200kW regen? I randomly got 172kW on a cold battery so maybe it would allow even more if the battery was willing. But I don't think you'd want that much on pedal-lift anyway, it'd be much better to have modest regen for traffic negotiation paired with major regen for heavy braking. Assuming of course, that the major regen is well blended with good pedal feel and control.

Last week I had a BMW i4 M50 and drove it a lot. It has this Adaptive Regen function, it was incredible. It knew when you wanted more (or less) regen and did it all by itself. The traffic light your coming onto is red? Stronger regen. There are cars in front of you? Stronger regen. Lift off on a free highway? Coast. When I got into the car and read "adaptive" on that screen, I thought "oh no, this is going to be horrible". But it wasn't!

 

[Watts_Up]

Last week I had a BMW i4 M50 and drove it a lot. It has this Adaptive Regen function, it was incredible. It knew when you wanted more (or less) regen and did it all by itself. The traffic light your coming onto is red? Stronger regen. There are cars in front of you? Stronger regen. Lift off on a free highway? Coast. When I got into the car and read "adaptive" on that screen, I thought "oh no, this is going to be horrible". But it wasn't!

What about driving on curvy roads?

Does the car drives at the speed limit posted, and then in middle of the curve applied energetically the breaks?
Or does the car starts to slow down when approaching a curve, and then when the end of the curve is visible
the car progressively starts to accelerate, like a more experienced driver will do?

 

[Gauss Guzzler]

...BMW i4 M50 and drove it a lot. It has this Adaptive Regen function, it was incredible. It knew when you wanted more (or less) regen and did it all by itself.

Yeah there's certainly some merit to that. Some people have difficulty modulating their foot well enough to drive smoothly within the limited travel range of the Gauss pedal and this is exacerbated by high power EVs and even further exacerbated by high regen power. So an auto-scaling regen logic helps to provide a more comfortable drive without giving up the benefits of one-pedal driving or regen efficiency. It's the exact opposite of what we'd want for performance driving however, since it makes the car less predictable.

 

[cappo3]

What about driving on curvy roads?

Does the car drives at the speed limit posted, and then in middle of the curve applied energetically the breaks?
Or does the car starts to slow down when approaching a curve, and then when the end of the curve is visible
the car progressively starts to accelerate, like a more experienced driver will do?

Are you now talking about driving on cruise control? It's not what I meant.
This adaptive functionality is still a one pedal driving kinda thing of which you are in control and, depending on the driving mode you're in, behaves differently.
The car did not do anything I was not expecting it to do, in contrast to what my M3s sometimes do on Autopilot.

Yeah there's certainly some merit to that. Some people have difficulty modulating their foot well enough to drive smoothly within the limited travel range of the Gauss pedal and this is exacerbated by high power EVs and even further exacerbated by high regen power. So an auto-scaling regen logic helps to provide a more comfortable drive without giving up the benefits of one-pedal driving or regen efficiency. It's the exact opposite of what we'd want for performance driving however, since it makes the car less predictable.

The car will react in a situation-aware manner in the Eco, Eco Pro and Comfort driving modes, while behaving differently in Sport and Sport Plus. I won't pretend to know the ins and outs of this car's brain, I can only testify to the great experience I had with it. Which is more than I can say about my own cars.

 

[OleJohnny]

I'm not aware of any car on the market with nicely blended brakes. Both Hitachi and Brembo even list well blended brakes as one of the core functionalities made possible by electric calipers. And what can a Prius even do anyway? Maybe 12kW? A driver can't tell the difference between squishy or smushy when lightly touching the pedal in a parking lot. But crank the regen up to 200kW and drivers are sure as heck gonna be sensitive to transition feel under medium/hard braking.

Also, there's little indication that fast charging is a significant factor in battery degradation, so a couple seconds of 172kW regen once a week is surely negligible.

I haven't seen a good description of Hitachi's system, but Brembo's looks like it uses a souped up parking brake motor on the rear calipers and a pair of ABS pumps to drive conventional hydraulic front calipers. The primary function and feedback are electric, but a conventional master cylinder will actuate the front calipers in the event of electrical failure.

The Taycan has indistinguishable blended braking. Up to 275kw regen via braking as well.

 

[Tiger]

Last week I had a BMW i4 M50 and drove it a lot. It has this Adaptive Regen function, it was incredible. It knew when you wanted more (or less) regen and did it all by itself. The traffic light your coming onto is red? Stronger regen. There are cars in front of you? Stronger regen. Lift off on a free highway? Coast. When I got into the car and read "adaptive" on that screen, I thought "oh no, this is going to be horrible". But it wasn't!

Tesla also does this but to a lesser extent. Regen is lower when you have a set speed (even if it is much lower than current speed). When arriving at red light or stopped cars, regen tries to stop the car correctly, providing stronger regen if needed.

 

[gearchruncher]

Tesla also does this but to a lesser extent. Regen is lower when you have a set speed (even if it is much lower than current speed). When arriving at red light or stopped cars, regen tries to stop the car correctly, providing stronger regen if needed.

Are you saying it does this on Autopilot, or even when autopilot or TACC is not being used at all?
Got any evidence?

 

[Tiger]

Are you saying it does this on Autopilot, or even when autopilot or TACC is not being used at all?
Got any evidence?

If TACC is used it reduces regen at high speed regardless of what the TACC speed setting is. This is very unintuitive, though, and I typically end up cancelling the TACC to get normal regen back.

When arriving at red lights or a queue it adapts regen so that the car stops in time when TACC is not used.

I use Autopilot only for coasting so I have not paid attention to it.

 

[gearchruncher]

When arriving at red lights or a queue it adapts regen so that the car stops in time when TACC is not used.

I use Autopilot only for coasting so I have not paid attention to it.

OK, so you are discussing what the car does when on AP. That's not what other people are discussing.

All you are saying is that the car doesn't either use full regen or no regen. Which is exactly what I do with my foot too as I come to a stop light. But that doesn't mean AP uses MORE regen than you can get with your foot.

 

[Gauss Guzzler]

AP does use more regen than you can get with your foot, as you might expect. I don't know how much it uses, but it will exceed the 85kW limit in street mode.

 

[Tiger]

OK, so you are discussing what the car does when on AP. That's not what other people are discussing.

Not what I said. Specifically, I meant AP is not part of the conversation as I use it only for constant speed coasting (regen doesn't come to play at constant speed on flat ground). Tesla TACC is not autopilot.

All you are saying is that the car doesn't either use full regen or no regen. Which is exactly what I do with my foot too as I come to a stop light. But that doesn't mean AP uses MORE regen than you can get with your foot.

We were discussing whether the foot regen is stronger or weaker depending on various driving situations. In Tesla it is too, but not very obviously and not always very intuitively, in contrary to the described BMW experience.

 

[gearchruncher]

Not what I said. Specifically, I meant AP is not part of the conversation as I use it only for constant speed coasting (regen doesn't come to play at constant speed on flat ground). Tesla TACC is not autopilot.

That's where I am lost. You also said this:

When arriving at red lights or a queue it adapts regen so that the car stops in time when TACC is not used.

Are you claiming that the car sees red lights and traffic and behaves differently even when not using AP or TACC? You're saying that if I take my feet off both pedals as I slow down to an intersection with a light, the max regen I get is dependent on the color of the light?

Can you describe a test I can go run to prove this? I can datalog the CAN bus for regen power, so just tell me what situations to put the car in where I'll see different regen amounts. I'll go try this out right away and bring back datalogs.

Not what I said. Specifically, I meant AP is not part of the conversation as I use it only for constant speed coasting (regen doesn't come to play at constant speed on flat ground). Tesla TACC is not autopilot.

This has to be one of the weirdest uses of AP I've ever heard of. As you say, AP is not TACC. But also, TACC is longitudinal control and AP is lateral. So you're saying that any time TACC needs to speed up, slow down, add or remove any power in order to go up/down a hill, you disable AP but keep TACC? But when the car needs to turn on flat ground, you keep AP on? Bizarre. Also only possible in super flat areas with zero traffic and negates one of the most useful environments for AP which is stop and go traffic. What don't you trust about having TACC changing speed while AP is also on?

Finally, the definition of coasting is "moving easily without using power."- So you're mis-using that which is part of the confusion. "Constant speed coasting" by definition means you are going down a hill that is steep enough to maintain your speed, and possibly steep enough to use.

 

[Tiger]

That's where I am lost. You also said this:


Are you claiming that the car sees red lights and traffic and behaves differently even when not using AP or TACC? You're saying that if I take my feet off both pedals as I slow down to an intersection with a light, the max regen I get is dependent on the color of the light?

On TACC without activating AP. Yes, I suspect it is some sort of AEB related functionality, but very smooth. It is more noticeable when there are cars in front, but it feels like there is a difference also if the lights are red.

Can you describe a test I can go run to prove this? I can datalog the CAN bus for regen power, so just tell me what situations to put the car in where I'll see different regen amounts. I'll go try this out right away and bring back datalogs.

Cool.

• Compare flat ground, no obstructions (think empty parking lot or some field, race track, airport, etc.) and test the regen. No AP no TACC. Then test the regen when arriving at traffic lights where there are cars stopped in front of you and lights are red. No AP no TACC.
• Then set TACC to some 40mph and drive on highway, drive 75mph (exceed the set TACC speed by pressing speed pedal). Then take your foot off the speed pedal and compare regen to situation where TACC is switched off.

This has to be one of the weirdest uses of AP I've ever heard of. As you say, AP is not TACC. But also, TACC is longitudinal control and AP is lateral. So you're saying that any time TACC needs to speed up, slow down, add or remove any power in order to go up/down a hill, you disable AP but keep TACC? But when the car needs to turn on flat ground, you keep AP on?

Nope. I just don't use AP very much because I enjoy driving (and I dislike how AP drives, or cannot handle curves etc.). I do almost 2k miles per month.

Bizarre. Also only possible in super flat areas with zero traffic and negates one of the most useful environments for AP which is stop and go traffic. What don't you trust about having TACC changing speed while AP is also on?

I don't do stop and go traffic, I either drive the bus lane with EV privileges or take a shortcut. If traffic is smooth TACC feels sufficient.

I use AP mostly at high speed on multi lane highways while coasting on the overtaking lane (yes I watch out rear mirror to make way if there is someone who wants to pass).

Finally, the definition of coasting is "moving easily without using power."- So you're mis-using that which is part of the confusion. "Constant speed coasting" by definition means you are going down a hill that is steep enough to maintain your speed, and possibly steep enough to use.

Well... I'd say you digress And remind me again how this is related to Highland Plaid

 

[gearchruncher]

Cool.

• Compare flat ground, no obstructions (think empty parking lot or some field, race track, airport, etc.) and test the regen. No AP no TACC. Then test the regen when arriving at traffic lights where there are cars stopped in front of you and lights are red. No AP no TACC.
• Then set TACC to some 40mph and drive on highway, drive 75mph (exceed the set TACC speed by pressing speed pedal). Then take your foot off the speed pedal and compare regen to situation where TACC is switched off.

I think I can accomplish this even easier with just a sheet of paper covering the camera. Exact same road, exact same situations. Just in one case the car can see forward with the camera and see the cars or red light, in the other case it cannot. Would you agree that would work?

I think you can do this even without logging power.

1) Go 50 MPH towards an intersection with the camera blocked
2) Identify a mark where if you take your foot off both pedals, you roll to a stop using regen just past the intersection stop line
3) Repeat #2 a few times to be repeatable.
4) Remove camera block
5) Repeat with red light and see if car stops under regen before stop line
6) Repeat with cars at light and see if car stops under regen before hitting the car

 

[Tiger]

I think I can accomplish this even easier with just a sheet of paper covering the camera. Exact same road, exact same situations. Just in one case the car can see forward with the camera and see the cars or red light, in the other case it cannot. Would you agree that would work?

I do not know how Tesla AI works so I do not know if blocking cameras will give the same result. And I don't understand why you would want to change the test setup, it only introduces more variables to the equation. If you insist on trying it, then I would recommend you add it as a third permutation for comparison instead of replacing the no traffic version.

I think you can do this even without logging power.

1) Go 50 MPH towards an intersection with the camera blocked
2) Identify a mark where if you take your foot off both pedals, you roll to a stop using regen just past the intersection stop line
3) Repeat #2 a few times to be repeatable.
4) Remove camera block
5) Repeat with red light and see if car stops under regen before stop line
6) Repeat with cars at light and see if car stops under regen before hitting the car

Will be interesting to see the logs.

 

[gearchruncher]

I do not know how Tesla AI works so I do not know if blocking cameras will give the same result.

You don't know how Tesla "AI" works, yet you are SURE that the car regen is different based on red lights and cars ahead? Tesla famously only uses cameras for their object detection. It's impossible for the car to detect these things without seeing them.

I'm suddenly very, very skeptical that you have any data that Tesla's regen is based on vehicles ahead if you don't even know the basics of Tesla's ADAS.

And I don't understand why you would want to change the test setup, it only introduces more variables to the equation.

Your test setup involves doing initial regen testing in a different place (race track, empty parking lot) than the test with traffic. This is a huge variable.

Mine does everything in the exact same location and the only variable that changes is the car's ability to perceive, which is the fundamental thing you are claiming Tesla's regen logic reacts to. It also fundamentally tests the place the stop occurs, which is what you claim changes instead of intuiting it from regen power.

What is wild about this is that if all of this is happening, Tesla doesn't document any of it. Why go through all this work and create all this inconsistency in user experience if you're not going to sell it as a feature?

 

[Tiger]

You don't know how Tesla "AI" works, yet you are SURE that the car regen is different based on red lights and cars ahead?

Yes, I know the AI in my butt. Also I have learned to adapt (my driving) to it due to its predictability by the patterns I described.

Your test setup involves doing initial regen testing in a different place (race track, empty parking lot) than the test with traffic. This is a huge variable.

If you have a road with empty traffic you can do it there too. Yes, a strange environment like parking lot might be a significant variable.

Mine does everything in the exact same location and the only variable that changes is the car's ability to perceive, which is the fundamental thing you are claiming Tesla's regen logic reacts to. It also fundamentally tests the place the stop occurs, which is what you claim changes instead of intuiting it from regen power.

I am not sure if the car uses regen level to accomplish the varying one foot braking but that's just an assumption. My main observation is that the one foot braking strength (g forces) varies in a predictable pattern.

What is wild about this is that if all of this is happening, Tesla doesn't document any of it. Why go through all this work and create all this inconsistency in user experience if you're not going to sell it as a feature?

They have so many "fuzzy smart" features in the car they cannot document all, and some of them might be beta and might be changed or removed in future updates.

For example you can put the car to reverse without brake pedal in some situations, but not all. If the parking lot looks too empty, the car will assume you are driving forward, and you will need to press brake. But if there are cars parked, it will assume you are parking and allow to switch to reverse on the fly without stopping. And it is not 100% consistent as noting is with AI vision.

 

[gearchruncher]

For example you can put the car to reverse without brake pedal in some situations, but not all.

Yeah, you linked to your own opinion that this exists with no evidence, and then were immediately refuted by other people. You can't cite yourself as fact that this exists.

It's been an interesting conversation, but for me it's clear that you just believe these things out of your own experiences without any actual evidence, and a generalized sense that the car is somehow using all sorts of sensors all the time. The preponderance of evidence from 4.9 million other Tesla drivers is that these things don't happen, and logic says there is no reason for Tesla to do them. I think you're also tainted by the specific advertised, documented, and indicated existence of Adaptive Regen on BMW's and are now experiencing a placebo effect in the Tesla hoping that it has some of the same tech.

I do hope you can come up with some evidence besides your "butt," but until then I'm done, although I will be paying more attention as I drive.