P85D - Electric Mechanical Braking System

[Lloyd]

Consider the instance when you are in a skid. Your reaction in an ICE car is to take your foot off of the brake or accelerator to give the best 'neutral ' acceleration. In a Tesla as designed now taking your foot off the accelerator and being in full regen has the potential to put you in a reverse skid. This would be accentuated in icy or wet conditions.

Perhaps putting some of the regen on the brake would provide a more natural reaction in recovering from a skid and provide a little more safety in those conditions. It sounds like this master cylinder has the potential to do this for us.

 

[Saghost]

Just wondering why you guys think 120kW of regen? In the case of the P85D the front motor is not as big as the back motor. In the case of an 60D/85D the front and back motors are smaller than the RWD equivalents. Maybe I'm missing something but shouldn't these smaller motors be capable of less regen?

Yes: a smaller motor is capable of less regeneration.

However, the current car's regeneration isn't set by what the rear motor is capable of, but rather by some combination of what was acceptable for the accelerator pedal and what could safely be applied to the rear wheels without disrupting the car's handling.

Both of those limits are out the window in the proposed system. A synchronous AC device like all modern EV drive motors can generally produce about equal amounts of power as a motor or as a generator.

So there's little question the motors and inverters could manage a while lot more than the current 60 kW - what the battery can safely accept would likely become the next limit, and folks are fixated on 120 kW because of the Superchargers.


The fact is that for a short term surge event like regen the battery can safely handle a lot more than it can for a sustained event like DCFC - most likely it wouldn't have any issues with the whole front motor output plus the current 60 kW from the rear - but we know for certain that 120 kW won't be an issue in the bottom half of the battery.
Walter

 

[jerry33]

During that rare occasion when you do have to press the brake pedal (because you need to stop faster than the current braking effect from accel lift off) imagine you get the exact same additional braking effect. But the only difference is: instead of making brake dust, you are making electricity. Can we all agree that making electricity is better than making brake dust?

But not at the expense of the smooth brake feel. And I doubt that one mile of extra range per month is worth the additional complexity and loss of smooth braking.

 

[Saghost]

But not at the expense of the smooth brake feel. And I doubt that one mile of extra range per month is worth the additional complexity and loss of smooth braking.

The complexity already happened, when Tesla decided they needed electromechanical braking for Autopilot. Every Tesla built in the last month or so has the hardware for this, and all future ones will.

I'm not convinced the braking feel aspect is inevitable, especially since the new Tesla system works on different principles and hardware than anything else before it.

I think I'd agree with you if we knew it had to have a large negative effect on pedal feel - but I also think it could be done seamlessly with the new hardware, in which case I have trouble understanding how it could be a bad thing.
Walter

 

[JRP3]

The fact is that for a short term surge event like regen the battery can safely handle a lot more than it can for a sustained event like DCFC - most likely it wouldn't have any issues with the whole front motor output plus the current 60 kW from the rear - but we know for certain that 120 kW won't be an issue in the bottom half of the battery.
Walter

But it will be a big problem in the "top half", or more accurately when the pack is at a higher SOC. The concern is not just heat but voltage, and pushing a large amount of current into a pack at a higher SOC will drive cell voltages higher instantaneously. So that would mean the regen would have to be actively reduced depending on pack SOC, the way it is now in cold weather with fully charged packs, but more so. I think for consistent A pedal feel it might make more sense to put the extra regen on the brake pedal in that case, if the unit can be programed to keep a consistent brake pedal feel by the booster compensating for when regen needs to be reduced. Ultimately the energy recovered will be minimal, unless you spend a lot of time slamming on your brakes, so may not be worth the extra engineering.

The complexity already happened, when Tesla decided they needed electromechanical braking for Autopilot. Every Tesla built in the last month or so has the hardware for this, and all future ones will.

That is a reasonable point. So if the programming is not too complex to make it seamless then it might make sense to integrate it.

 

[Cottonwood]

Consider the instance when you are in a skid. Your reaction in an ICE car is to take your foot off of the brake or accelerator to give the best 'neutral ' acceleration. In a Tesla as designed now taking your foot off the accelerator and being in full regen has the potential to put you in a reverse skid. This would be accentuated in icy or wet conditions.

Perhaps putting some of the regen on the brake would provide a more natural reaction in recovering from a skid and provide a little more safety in those conditions. It sounds like this master cylinder has the potential to do this for us.

Exactly!

I had exactly this situation happen descending a long steep hill in an ICE low gear using engine braking. When I hit some ice, the car went into a skid. Luckily, I was able to think fast, add some accelerator, and upshift quickly, arresting the skid. I therefore strongly recommend not using engine braking in marginal traction situations.

OTOH, the MS has an amazing antl-lock brake system. If that could be programmed to reduce regen in a skid situation, the MS could be safe. In fact, it may be my imagination, but I think it does that already...

 

[bhzmark]

The complexity already happened, when Tesla decided they needed electromechanical braking for Autopilot. Every Tesla built in the last month or so has the hardware for this, and all future ones will.

I'm not convinced the braking feel aspect is inevitable, especially since the new Tesla system works on different principles and hardware than anything else before it.

I think I'd agree with you if we knew it had to have a large negative effect on pedal feel - but I also think it could be done seamlessly with the new hardware, in which case I have trouble understanding how it could be a bad thing.
Walter

Exactly. I think it is very possible that we will find out that the new electomechanical braking already is designed to trigger additional regen from the brake pedal before triggering the old fashion brake dust makers. Then it will be amusing to see all these people who are defending the status quo have their heads explode as they contemplate the tragedy of additional regen coming from the brake pedal. horrors!

 

[jerry33]

Exactly. I think it is very possible that we will find out that the new electomechanical braking already is designed to trigger additional regen from the brake pedal before triggering the old fashion brake dust makers. Then it will be amusing to see all these people who are defending the status quo have their heads explode as they contemplate the tragedy of additional regen coming from the brake pedal. horrors!

If it can be done smoothly with no additional friction in all circumstances, I have no problem with it. However, no one else has been able to put regeneration on the brake pedal without poor brake feel, extra drag, or both. Some companies have been at it for twenty years and still haven't achieved it. Doesn't mean Tesla can't but...
And there's still the additional complexity. Ideally, the safety systems should be as simple as possible.

 

[bhzmark]

However, no one else has been able to [fill in the blank] . Some companies have been at it for twenty years and still haven't achieved it. Doesn't mean Tesla can't but....

Indeed. It is silliest argument to say "but others haven't done it -- therefore Tesla can't do it."

And in fact it looks Tesla plus Bosch have done it. This video on the ibooster components strongly suggests that the brake pedal will trigger additional regen: EN | IAA 2013: Bosch iBooster - YouTube

Which makes sense since the autopilot braking wouldn't be limited by the driver experience cap on the accel lift off regen braking effect. so if autopilot could trigger additional braking regen, the electric brake pedal could as well.

And in the Bosch video they specifically say that it will be smooth. But that was a silly concern to begin with since Toyota/Lexus have already been doing it for over ten years -- my 2004 Prius brake pedal was perfectly smooth and after 150k+ plus miles still had 90% of brake pads left. Plus, even if Toyota hadn't figured this out back in 2002, I have little doubt that Tesla could do it.

Welcome to the world of stronger regen braking from the brake pedal and from autopilot.

 

[jerry33]

But that was a silly concern to begin with since Toyota/Lexus have already been doing it for over ten years -- my 2004 Prius brake pedal was perfectly smooth and after 150k+ plus miles still had 90% of brake pads left. Plus, even if Toyota hadn't figured this out back in 2002, I have little doubt that Tesla could do it.

Except that Toyota hasn't done it. You get used to the feel, but it's not even remotely smooth if you're not braking on perfectly smooth roads. There is a transition between regen and friction. In the early 2010 models it was so bad Toyota had to issue a recall.

 

[EarlyAdopter]

Except that Toyota hasn't done it. You get used to the feel, but it's not even remotely smooth if you're not braking on perfectly smooth roads. There is a transition between regen and friction. In the early 2010 models it was so bad Toyota had to issue a recall.

If Tesla ends up doing this (and that is still a big IF), here's hoping they can do better than Toyota.

Safe bet, that one I think.

 

[Lloyd]

Except that Toyota hasn't done it. You get used to the feel, but it's not even remotely smooth if you're not braking on perfectly smooth roads. There is a transition between regen and friction. In the early 2010 models it was so bad Toyota had to issue a recall.

I find the regen on the brake pedel on the Rav4EV (new version) totally smooth.

 

[tezco]

I guess we're a bit off topic, but what the hay...

Exactly! I had exactly this situation happen descending a long steep hill in an ICE low gear using engine braking. When I hit some ice, the car went into a skid. Luckily, I was able to think fast, add some accelerator, and upshift quickly, arresting the skid. I therefore strongly recommend not using engine braking in marginal traction situations.

(If you really want to test your skills, get in a car with a hand lever emergency brake, put the car in neutral and pull the brake while you are going downhill on ice and try to stay in your lane...)

OTOH, the MS has an amazing antl-lock brake system. If that could be programmed to reduce regen in a skid situation, the MS could be safe. In fact, it may be my imagination, but I think it does that already...

With regen spread across 4 AWD wheels instead of 2, the rear wheels in the D should be less likely to break free. I think the ABS and traction control in the RWD S already reduces regen if the rears slow down more than the fronts in the current RWD model. I suppose it might be a little more difficult to detect in AWD, but I suspect one wheel will break loose first in most situations; otherwise existing AWD cars would have trouble when decelerating.

On the opposite side of the curve, my old Audi will control wheel spin with selective brake application, but it doesn't cut power even when all four break loose upon acceleration, which I really enjoy since you can get into a 4 wheel power skid while still retaining full power. My SLK and the S cut power so drastically that I can't even get any wheel spin with traction on. Where's the fun in that? I like to see some sort of override at high throttle settings; then you have traction control on for everyday use, but instant override when wheel spin is desired.

As far as the regen argument goes, I'd vote to try for a bit more regen on the AWD S accelerator pedal, but even with the current 60, I occasionally find myself see-sawing a bit, particularly when I transition in from my Leaf (with anemic regen), or my nearly free wheeling Audi), so it just may not prove practical with only so much pedal travel available. Maybe some user settable variation would help satisfy the spectrum of experienced S drivers, with a default to a lower regen limit in valet mode?

I'd like to see any additional available regen put on the brake pedal, not for panic stops, but for increased deceleration on steep hills and downhill freeway off ramps, where there is some significant energy harvesting to be had.

 

[andrewket]

Tesla could implement brake pedal regen, as well as increased regen capacity w/AWD, but then give the driver selectable options on what they prefer. It adds some complexity, but with sane defaults I think it can be done. Tesla could do this in the future with a software release. The hardware appears to be there today.

 

[GSP]

Bosch iBooster

Thanks TEG for the link with nice photo of the booster-master cylinder assembly.

From the Bosch info at TEG's link:

by ensuring deceleration rates of up to 0.3 g are achieved using the electric motor alone. It thus covers all common braking maneuvers in everyday traffic. If the brakes to be applied harder, the iBooster generates the additional braking pressure needed in the traditional way, using the brake master cylinder. The driver does not notice this harmonious interplay of motor and brakes, as pedal feel remains absolutely normal.

GSP

 

[Johan]

Tesla could implement brake pedal regen, as well as increased regen capacity w/AWD, but then give the driver selectable options on what they prefer. It adds some complexity, but with sane defaults I think it can be done. Tesla could do this in the future with a software release. The hardware appears to be there today.

I agree that this whole ting could be much more customizable. However, I'm in the camp of keeping regen on the accelerator and leaving the brake pedal out of it, be it and electromechanical brake system or a traditional hydraulic one. The reasons being I think the added benefit (range wise) with normal use would be very small (track use is something else), it adds complexity and could give rise to unexpected handling on emergency braking.

I played with the idea of customizable drive modes a good while ago:
Feature idea: Custom Driving Modes

I've been thinking some about a feature I would love to have in my car. This feature would be sort of a feature for an "advanced user" and would perhaps best be in a submenu of the Driving tab. The idea would to be to allow for user definable accelerator pedal mapping.

First of all, the way in which the accelerator pedal commands the drive train is obviously via an analog to digital interface of sorts - the travel of the pedal as you either push down or let go is an analog function that is converted to a digital control signal which in turn controls the current flow to the motor. So we know that there is already mapping. We also know that a lot of user have reported feeling a difference in mapping (how aggressive the acceleration is) with firmware updates. We also know that regen is capped when battery temp is low (yes there is a physical reason for this but the capping must happen in the control software).

As of today we can only change one setting: Regen normal or Regen low. We can't turn regen off. We also cannot cap or modulate acceleration. Valet mode is promised in future upgrades, which would be an example of this.

My actual suggestion is to create a menu where the user is able to create new driving modes, choose between preset modes and also edit these. These presets would be connected to driver profiles. There would also be an option for the presets to password/PIN protect them working in the way that if that mode is set (for example Valet or Teenager mode) you would need to PIN/password to get out of that mode and in to another ("Daddys racing mode" for example).

The way I envision the IU for this is simple. You present the user with a diagram ranging from -60 to +320 kW. There are 6 points that the user can drag vertically - similar to a sound system EQ. The curve between the points interpolates and smooths. This curve represents the accelerator pedal mapping. Here is a crude rendering of let's say what is the current/standard mapping (note the Y-axis is not at all linear):

View attachment 42403

Let's say you want a profile for hypermiling where to want only very cautious regen (kicks in very softly and capped at 20kW) and a dampened acceleration curve, no problem, just drag:

View attachment 42406

Teenage son borrowing the car, OK cap acceleration and accelerator responsiveness:

View attachment 42407

For the race track:

View attachment 42408
OK, so you get the idea (sorry for the nasty graphics).


Version 2 of this feature would also include, if possible, the option to map regen to the brake pedal. Now, Tesla haven't done it this way and I respect that, but some might like to have regen on the brake pedal only, some want some regen when letting go the accelerator and the rest on the brake. Also, the mechanical brakes would have to come on gradually and the speed with which the brake is depressed would have to be taken in to account (a sudden hard push on the pedal should engage the mechanical brakes instantly for example).

 

[PeterK]

Why are people pushing for regen to be moved to the brakes? Keeping regen on the accelerator pedal is, IMHO, one of the best "driving feeling" decisions Tesla's made. Not just because it avoids the complexity of trying to keep natural braking feel, but because it lets you control the car in most driving situations without moving your foot very much.

+1. 20 months in, and having owned a Toyota hybrid, I much prefer the one-pedal driving experience and can't imagine going back. I love the ability to finely control the speed and position of the car vs the cars around me.

 

[PeterK]

If and only if the Bosch iBooster can capture additional regen seamlessly, then it makes sense to do so. But like others above I don't think the incremental range would be worth compromising the brake feel.

 

[jerry33]

If and only if the Bosch iBooster can capture additional regen seamlessly, then it makes sense to do so. But like others above I don't think the incremental range would be worth compromising the brake feel.

And seamlessly means going over an expansion joint or washboard surface while braking and not having the regen pulled out from under you so that the car appears to lose braking ability.

 

[Saghost]

And seamlessly means going over an expansion joint or washboard surface while braking and not having the regen pulled out from under you so that the car appears to lose braking ability.

I'm not sure that's possible without risking a skid.

However, the iBooster car should be able to instantly ramp up the friction brakes to hold the rate of deceleration, then slowly phase regen back in once the wheel speeds settle down, which is close enough in my book.

Walter