Musk Says Model 3 Brakes Need Work, Firmware Update Coming in a ‘Few Days’

Responding to Consumer Reports’ criticism of the Model 3’s braking ability, Tesla CEO Elon Musk tweeted Monday that the company will “make sure all Model 3’s having amazing braking ability at no expense to customers.”

Consumer Reports said the Model 3’s stopping distance of 152 feet from 60 mph was far worse than any contemporary car they have tested. Musk called their results “strange” and inconsistent with other reviews. He suggested that performance may vary by vehicle and Consumer Reports may have had their hands on a dud.

https://twitter.com/x/status/998739763543736320

Musk later tweeted that braking tweaks can be made through a firmware update, which Tesla intends to ship in a few days. He said the company is dedicated to making the Model 3’s braking better than any comparable car.

Consumer Reports also said the Model 3’s stiff ride, unsupportive rear seat and excessive wind noise at highway speeds hurt its road-test score. Musk said the magazine was testing an early production car and those issues have since been addressed for cars coming off the assembly line today.

https://twitter.com/x/status/998778081492926465

Tesla has acknowledged that the Model 3’s braking needs improvement. The good news is that the company can seemingly avoid an expensive recall and make those performance improvements via a software update.

 

[Ben D]

Since the first emergency stop is good, then they experience fade, and then the braking performance doesn't come back even after cooling the car down overnight, it is not a software issue. It is a hardware issue, with the brake pad compound chosen. Sounds like the compound is too soft, and glazes too easily. Once they glaze, coefficient of friction drops and doesn't come back. They probably didn't pay as much attention to the friction brakes as they otherwise would, given that regen would be used most often. So I reckon they will try to cover the problem in the short term with electronics, but in the end they will probably need to provide brake pad upgrades, which is not a bad thing given the stock pads lasted less than 3 laps at laguna seca (i.e. they are crap). I bet the Performance versions will not have this problem as they have an upgraded friction brake package anyway, and have already been track tested.

 

[Reid]

It's almost certainly regen tuning, because that's the easiest to fix OTA. From the pictures, the 3 is certainly not lacking in brake hardware, and a single stop from 60 wouldn't fade the brakes that badly on anything this side of a Yaris/Corolla type car. And certainly not in a way that hurt performance once the brakes had cooled.
Obviously pad compound CAN make a difference, but this seems like way too little usage for that to come into play yet unless they used something laughably bad.
Thankfully, brake pads are cheap.

 

[mongo]

It's almost certainly regen tuning, because that's the easiest to fix OTA. From the pictures, the 3 is certainly not lacking in brake hardware, and a single stop from 60 wouldn't fade the brakes that badly on anything this side of a Yaris/Corolla type car. And certainly not in a way that hurt performance once the brakes had cooled.
Obviously pad compound CAN make a difference, but this seems like way too little usage for that to come into play yet unless they used something laughably bad.
Thankfully, brake pads are cheap.

ABS, not regen. Pedal braking should be able to lock wheels on its own, without any regen (cold or full pack for instance). No regen just means you push harder for the same deceleration.

 

[ILLCOMM]

ABS, not regen. Pedal braking should be able to lock wheels on its own, without any regen (cold or full pack for instance). No regen just means you push harder for the same deceleration.


View attachment 303195

The answer of "ABS calibration algorithm" is interesting. The *calibration* is the problem? I guess the logical conclusion is that there is a subset of cars whose calibration is flawed and needs to be reset/reinitiated. This could absolutely have to do with regen. In fact, it makes perfect sense.

 

[mongo]

The answer of "ABS calibration algorithm" is interesting. The *calibration* is the problem? I guess the logical conclusion is that there is a subset of cars whose calibration is flawed and needs to be reset/reinitiated. This could absolutely have to do with regen. In fact, it makes perfect sense.

In order for regen to be the dominant factor, the brakes would need to be too weak to lock up the tires on their own. In order for different tests to have different results, something need to be changing and staying changed. (regen should improve over multiple test runs due to lower SOC (until SOC is too low to accept full regen))

The car's initial braking distance was different (better) than later tests, including the next day. That says to me that the ABS system itself changed behavior due to the testing. If the system was left in factory/calibration mode, or the system always adapts itself to try to detect slip ahead of time, then it would produce a semi-permanent change in braking behavior.

The test driver is applying maximum pressure to the brake pedal. This level of force should lock the brakes in a properly designed system with no ABS. The Motor Trend data showed the brakes are sufficient for a ~120ft stopping distance. To prevent persistent wheel lock up, the ABS system should kick in. If the activation threshold changes or the dwell time for the isolation, pressure release, or re-pressurization events change, the average braking force can be reduced, thus increasing the braking distance.

 

[ILLCOMM]

If the system was left in factory/calibration mode, or the system always adapts itself to try to detect slip ahead of time, then it would produce a semi-permanent change in braking behavior.

I have no idea whether it is typical for ABS systems to "learn" and adapt over time. I always assumed that ABS is static. If not, your explanation makes sense.

 

[mongo]

I have no idea whether it is typical for ABS systems to "learn" and adapt over time. I always assumed that ABS is static. If not, your explanation makes sense.

Yah, I'd never really thought about it before this, but things like caliper release time seem like they would shift over vehicle life. Force to engage, vs force to disengage. Interesting .

Maybe there is an 'max pressure after locking event' variable which gets tweaked to compensate for tire/ road variations, but the time constant was set too long...

 

[MVuu]

I don’t understand Tesla trolls. Like Tesla, STAY! Don’t like Tesla, LEAVE. Such pessimist. My TM3 has been a blast! Amazing handling and performance.

 

[aronth5]

Are people happy with the new suspension? Are performance people satisfied with it? And does it compare to Toyota or Lexus in ride?
If so, problem solved. If not, Tesla may want to explore splitting into two options or investigating active suspensions.

I have the new suspension and its still harsher then I expected but I understand others like it. Today I rode my Cannondale Synapse Ultegra road bike to work and tried to decide which is smoother my bike or my Model 3.
Probably a tie but if I had to make a decision I'd say the bike is smoother which isn't good

 

[Peter Thomas]

Elon doesn't seem to hear me here on Tesla Motors blog or on his twitter or other twitter accounts like SpaceX or Tesla.
Seems like I am being blocked from communicating with him.
Just wanted to let him know that we are now connected to God through the matrix until China works out the issue with Ethan.
Hopefully people will start feeling better. Its been getting pretty crazy around here. I almost felt like I was on another planet.

Special thanks to Evan for connecting me to God.

I think I will try to get back to work now . I have a lot to catch up on. Its a long story that some of you may be interested in hearing.
I will be setting up a website Interstellar Friends in a few weeks in case anyone gets this message.

 

[scottf200]

Reasonable CR youtube on the situation.

Published on May 23, 2018
Consumer Reports has revealed that it would not be recommending the Tesla Model 3, in part due to poor braking performance. In response, Tesla CEO, Elon Musk agreed to meet with Jake Fisher, CR Auto Test Director, to talk [for about an hour regarding] our findings and discuss immediate updates Tesla might make. Hear all about that exclusive conversation this week on Talking Cars.

 

[scaesare]

Elon doesn't seem to hear me here on Tesla Motors blog or on his twitter or other twitter accounts like SpaceX or Tesla.
Seems like I am being blocked from communicating with him.
Just wanted to let him know that we are now connected to God through the matrix until China works out the issue with Ethan.
Hopefully people will start feeling better. Its been getting pretty crazy around here. I almost felt like I was on another planet.

Special thanks to Evan for connecting me to God.

I think I will try to get back to work now . I have a lot to catch up on. Its a long story that some of you may be interested in hearing.
I will be setting up a website Interstellar Friends in a few weeks in case anyone gets this message.

I can't imagine why you would be blocked...

 

[whitex]

@bro1999 Elon never offered to provide them a car for testing. All he said was "Also, Consumer Reports has an early production car. Model 3 now has improved ride comfort, lower wind noise & many other small improvements. Will request that they test current production."

In other words they know early production had issues, so how about you try to get a car more recently built to see how we are doing now instead of how we were doing almost a year ago.

Are they recalling all those early production cars and offering to fix those known early production issues, or is this typical Elon style, as long as it's fixed in future cars it's considered ok? Is bad braking an early production issue?

 

[whitex]

It's almost certainly regen tuning, because that's the easiest to fix OTA.

The first thing I bet that will find a way into the OTA is tracking of stopping distances. This way Tesla can call out anyone making up numbers, and Elon can use the numbers to play his usual tactic along the lines of "statistically Model 3 stopping distance is ok".

 

[eigenv1]

Elon,
Besides the software you may want to check the brake pads. I have had experience with hard long lasting brake pads which degrade the braking performance substantially.compared to the brake pads optimized for braking performance.

Apparently, 3rd party options are cropping up...some not so expensive...

Unplugged Performance Sport Brake Pad Set for Tesla Model 3

and some much more expensive...

Unplugged Performance Superlight Carbon Ceramic Brake Kit for Model 3

 

[dhanson865]

Are they recalling all those early production cars and offering to fix those known early production issues, or is this typical Elon style, as long as it's fixed in future cars it's considered ok? Is bad braking an early production issue?

There was a coil upgrade that Tesla offered for free to 2017 Model 3 owners. There is now a "damper" aka shocks upgrade that presumably will be a free retrofit as well.

But it isn't a recall as it is a comfort/performance preference not a safety issue. You have to request the change because you don't like the ride. If you do like the ride three is no reason to replace the springs or dampers.

2017 Model 3 could get springs and dampers
early 2018 Model 3 could get dampers
current Model 3 already has both

No reason for a recall, it's a minor change for comfort at the cost of performance.


For the wind noise issue, they remove/replace the glass. New installation method. Again if you find the wind noise objectionable let your service center know. It's not a safety issue so it isn't a recall.


The braking issue is not a hardware issue, it's a software issue. No recall, because OTA updates will fix. If for any reason the OTA update won't fix it I'm sure there would be a recall for the brake issue. That is safety related and there is no way Tesla won't do what it takes to fix on every Model 3.

 

[RDoc]

ABS at the basic level only kicks on when a wheel locks up. Up until that point the braking force is 100% based on the driver pedal pressure * booster assist. If you want to brake harder you push harder, so any fixed offset due to assumed regen can be bypassed by pushing harder.

Where it gets funky is if the ABS system anticipates the force need to lock up a wheel. If (made up number) 15kPa locked the wheels the last two braking events, maybe it starts backing off at 14kPa. If that were the setup, then the extra torque of regen would be a critical factor in the calculation (how much regen during the two lock up events vs how much now).
This type of system would be great in adverse conditions due to ability to reduce the number of lock up events and thus reduce stopping distance. The downside being if it 'learns' wrongly and overly reduces maximum braking, then you end up taking longer to stop.

This seems to me like a pretty poor way to program the system. ABS reads the wheel speed and adjusts the braking pressure in real time. If it did what you are suggesting, then if the car is on a typical snowy road with patches of snow/ice and dry pavement, the system could easily wind up extending the braking distance when on dry pavement because the last couple of braking events were on the snow.

A similar argument can be made against the ABS trying to adjust for regen. Keep it simple. All the ABS has to do is look at the wheel rotation speed and adjust the brake pressure to keep it rotating irrespective of what the regen is doing. At the limit, where the brake pressure is zero, it needs to tell the regen to stop, but beyond that, why make it more complex? The braking valves can demonstrably respond fast enough to adjust in a single braking event that extends from a slippery surface to dry pavement as it is. There's no need to add some kind of anticipation algorithm.

 

[jelloslug]

Apparently, 3rd party options are cropping up...some not so expensive...

Unplugged Performance Sport Brake Pad Set for Tesla Model 3

and some much more expensive...

Unplugged Performance Superlight Carbon Ceramic Brake Kit for Model 3

Those are massively overpriced.

 

[mongo]

I'm not saying its a good idea, only theorizing about how the SW could change parameter and why it might do so . Totally agree it needs to not carry parameters between disparate braking events.

ABS reads the wheel speed and adjusts the braking pressure in real time.

And to do so, it needs a pressure target to control to.

Doing more research, the cool thing is, the system can see the slip change before lock up. (max braking involves 8/10% slip). So it could be the code that calculates either the current slip and/or the slip target.

The current slip value calculation could have issues if the vehicle speed estimator has a carry over vehicle mass or braking force to deceleration value... The speed can't be directly measured due to all wheels having slip.

Reducing the slip target reduces the braking rate and increases the distance. Tthis parameter would be surface dependent. That might be the parameter that is getting adjusted wrongly. (system becoming more conservation over multiple lock up events)

Simulation article:
Modeling an Anti-Lock Braking System - MATLAB & Simulink
Research paper:
http://www.mdpi.com/1996-1073/7/10/6459/pdf

 

[Pollux]

Agree with @MrAustraliaTax. Refreshing to see the Musk of somewhat older times... Forthright, willing to take responsibility, willing to set the right goal ("we're gonna be the best at braking") and even be willing to eat some money if necessary.

That's the guy I respect so highly.

Alan