Tesla Model 3 Gets CR Recommendation After Braking Update | Hacker News Reader

Tesla Model 3 Gets CR Recommendation After Braking Update

consumerreports.org

58 pointsfmihaila85 comments

I'm bothered by the fact that it took Consumer Reports testing to discover this issue and to have it fixed. Where was Tesla's QC on this before all this? The CR break tests don't appear to be anything unusual.

>> break tests

brake tests

The big automakers collectively ship defective vehicles by the millions as a matter of routine business and have been doing so for as long as there has been an automobile business. They're going to continue doing so. You'll regularly get pieces of mail from companies like Nissan, Ford or Hyundai, telling you to bring your vehicle in, years after they sold it, to fix mistakes.

Most of the other automakers will just leave you rolling around in a defective 3400 lbs pile of metal for years until they get to a fix or finally figure out there's a problem.

Tesla was able to quickly improve its product with a software update and without inconveniencing its customers. It's a superior solution.

Tesla has recalls as well, like this recent one for a faulty power steering component:

https://www.cnbc.com/2018/03/29/tesla-recalls-123000-model-s...

Rarely are recalls for serious defects (those are usually caught before a model is sold), but they are overall a good thing for safety and long term reliability.

It's not quite as bad as it seems. The first high speed stop CR did produced the expected and acceptable result. But when they repeated the test they started to get poor results. Under normal driving conditions you don't often repeatedly slam on the brakes from 60mph. Now this isn't a complete absolution of Tesla under the circumstances but it does go some way to explaining how this software bug got into the wild when it shouldn't have.

That’s still pretty bad, honestly. All it took was repeated testing to trigger the issue? That indicates that either Tesla’s testing was woefully insufficient here or that they took a late change to braking behavior without a proper re-testing regimen.

I can actually understand the latter possibility, but it’s still a giant screw up.

Yes I agree with you. I expect they made a code change in order to stress test a different component of the braking system and then forgot to revert that change.

If so, it's a big screw up for this kind of product development. But still, it's much better than "these brakes have a design flaw and can never work properly and we don't know how to make cars".

It was a very particular brake test, multiple stops going from 60 to 0. apparently they did some things to not being so aggressively stopping repeatedly. stopping once you'd stop much faster.

So, OTA software update improved the brakes (dear commenters, please spell this word correctly). Why not have the brakes with that strength from the beginning? If I were to speculate, could it be because this setting would cause high brake wear on the long run?

And, what's stopping them from selectively applying this setting? Heh, Volkswagen would probably have loved this capability a few years ago.

> If I were to speculate, could it be because this setting would cause high brake wear on the long run?

No idea why it wasn't this way in the first place. That said on an EV you'll almost never replace the brakes. On our Model S I only touch the brakes in panic situations or < 2mph to come to a stop. Regen braking cover all other cases sufficiently.

I doubt it was intentional. It seems that the anti-lock brake system was miscalibrated.

So, Model 3 has regenerative brake (uses magnetic resistance to slow car) and regular brakes that every other car has. I believe the tradeoff / timing about when to switch between the two braking systems was slightly off, favoring more energy efficiency rather than the best stopping distance. When this was pointed out they simply adjusted their algorithm for a better overall tradeoff.

It's extremely embarrassing/telling that it took Consumer Reports to point out their new car doesn't stop well.

This was not an issue on all cars.

Hit and miss braking on a single product line seems... kind of worse?

I had my Mercedes SLK350 try to launch me into an intersection once while stopped at the red light with my foot on the brake. The dealership reset the ECU and said “Let us know if it happens again.”

Why everyone thinks these issues are Tesla specific is beyond me. They can at least address them over the air.

When I think about it, my BMW 3 series, also had a strange loss of accelartion in certain situations (very minor). Was fixed during a service interval, by flashing the ECU.

Alternately, it's amazing and inspiring what just happened; other high-end car makers with a century of experience have shipped imperfect brakes for years (1) and only redesigned them on new models in response to complaints. Tesla, by comparison, is applying the Apple model (when Apple is at its best) of continuous over-the-air product improvement, which can change one’s sense of a product's value for money.

1. MERCEDES CANCELS BY-WIRE BRAKE SYSTEM; DECISION A BLOW TO TECHNOLOGY'S FUTURE. "When the system failed, the hydraulic system took over. But that resulted in a longer stopping distance and additional brake pedal effort by the driver." http://autoweek.com/article/car-news/mercedes-cancels-wire-b...

Mercedes had a hardware problem and Tesla had a software problem, I really don't see how this is a win for Tesla. Tesla has done physical recalls before as well, not everything is a software bug.

https://repairpal.com/recall/05V133000

Well, I have the car and stopping distance was never an issue at all. With regenerative braking you start braking as soon as you take your foot of the gas so you’re usually stopping short of where you actually need to stop, not having trouble stopping fast enough. I really don’t think it was that big of a deal, but it could be fixed and so they did it to get the recommendation. And im sure it’s safer too which is what’s most important.

That still doesn't answer why the braking distance differed between the first test, and all subsequent tests, results which were repeated on two different Model 3s.

Unless the Model 3 is very different from the other Tesla cars, there is no such tradeoff. In a Model S, regenerative braking is triggered by releasing the accelerator pedal. Mechanical braking is triggered by stepping on the brake pedal.

I assume that Tesla just screwed up the antilock breaks. If you push the brake pedal all the way down on a dry road, you’re supposed to get full braking power by a mechanical linkage. No software needed.

>> antilock breaks

antilock brakes

In a Camry Hybrid, the "tachometer" has a charging section, to show how much of the regeneration you're using when braking. Above the crawl speed, it also trims the car speed as regenerative braking. Pressing the brake pedal increases regenerative braking - up to maximum, and engages the mechanical brakes at some point.

I'd hope Teslas aren't wasting all of their stopping energy as brake heat, and instead regain some through regeneration as well. Ideally, both braking systems would be combined to maximize stopping ability. However not combining them might explain the Model 3's abysmal (longer than an F-150!) stopping distance.

> antilock breaks.

Break is what the car does when the brakes don't work. Petals will be on the flowers if you don't press the pedal. English is fun.

Toyota and Tesla do braking completely differently. Tesla has one pedal driving, same as the Bolt. Taking your foot off the accelerator does regenerative braking. Pressing the brake pedal also engages the mechanical brakes.

Seems like an update on the ABS, so less to do with strength, more to do with how ABS are performed. My guess is this probably won't have much effect our everyday driving, but when in emergency, the updated ABS will perform better.

The original CR article noted inconsistent braking performance, performing very well the first time they tested it and very poorly every time afterward. Frankly it just seems like a bug that Tesla never noticed in their own testing. The fact that they screwed up the brake control software and did not realize this until some journalists complained about it makes me very concerned about Tesla's overall design/QA/QC process.

Move fast and brake things?

This sounds awfully similar to a well-known Prius problem, where the transition from regenerative to friction braking takes on the order of a second and can lead to significantly increased stopping distance.

What prevents a new update to revert this or cause other safety problem?

Pushing updates in a rush is not a good sigh for critical software.

I'd like to think they have some kind of automated CI with a huge warehouse full of cars that automatically drives around and tests everything with each Git push.

The cars are full of actual chaos monkeys.

>The cars are full of actual chaos monkeys.

What a coincidence! Car Full of Chaos Monkeys is the name of my Fountains of Wayne tribute band.

They sort of do, it's the general public.

And they've only killed a few people so far.

And it's a demolition derby where only the cars that survive get to pass on their genes/settings to the next generation on the production line.

If they did, they wouldn't need the media to find the problems for them. Without CR reviewing every update, how confident can you be that they won't regress the braking?

This image delights me. "Jenkins, hit the track!" "Immediately, Sir."

Please cite your evidence that this was done in a "rush."

It's not like Tesla makes changes in production. They do it first in staging/test/SDDV/etc... like the rest of us.

A feature landing a few days after Elon Musk asked for it, how much testing/miles was this actually tested?

If you have development experience you know that even trivial changes can cause unexpected issues.

I've had development experiences involving "all hands on deck" overnight testing of an upgrade if the boss thought it was important. Does that count?

Yes, me too, and adding a new thing in a rush because the boss wants it today it usually involves hacks, ugly workarounds, incomplete functionality. Also putting a fix late in the day increases the chance of making mistakes.

Btw, why is braking and software connected, I am just curious how can you affect braking with an update, does this mean that a bug/crash will make the brakes not work at all or work not as good as a regular car.

Also in response to an unfavorable review from a magazine. Is Tesla going to push out fixes after a wave of critical tweets?

> Is Tesla going to push out fixes after a wave of critical tweets?

If they uncover problems, why shouldn't they?

They should test them more, also why did they not handle the issue before it caused bad PR, it is like they don't care as much for safety as others.

> They should test them more,

They should test them more, yes, as should probably all car manufacturers. And computer programmers. Complex bug-free software doesn't exist.

> also why did they not handle the issue before it caused bad PR, it is like they don't care as much for safety as others.

Most likely because they were not aware of the problem. When CR pointed it out, they investigated and fixed it.

Do you think they are other safety related issues that they are not aware and they are waiting some big negative PR until they notice(people reported issues in the exact place the Tesla killed the driver but was not enough)

> What prevents a new update to revert this or cause other safety problem?

One would hope it's a culture of safety first at Tesla where all software changes are extensively reviewed, heavily tested in the real world and using whatever kind of simulations are possible, deployed in a reasonable manner to a small set of cars and only deployed to the main population of cars after a satisfactory amount of time.

> Pushing updates in a rush is not a good sigh for critical software.

Yeah, it's pretty scary to know the braking software in your car was tested for less than a week.

  - const brakingDistance = 152;
  + const brakingDistance = 133;

braking

oh, right... thank you! (edited)

  const breakingDistance = warrantyMiles + 1;

>> const breakingDistance = warrantyMiles + 1;

const brakingDistance = warrantyMiles + 1;

Read the chain of comments again

Woosh

if (autoPilot.active) brakingDistance = null;

It was an ABS issue for anyone wondering. https://twitter.com/elonmusk/status/998778751767871488?lang=...

Controlling the ABS hydraulic modulator and ABS diagnostic bleed are one of the most dangerous cyber-attacks that can be performed on a car.

The fact that these systems are (indirectly) connected to the Internet is irresponsible.

>In retesting after the software update was downloaded, the sedan stopped in 133 feet from 60 mph, an improvement of 19 feet.

Combining from a previous thread[0]:

  Car           distance (feet)
  Model 3(orig) 152
  Model 3 (new) 133
  F-150         129
  Model X       127
  Camry Hybrid  125
  F-150 Lariat  119
  Chrysler 300S 109

It's noteworthy that the stopping distance is still longer than the heavier Model X.

[0]https://news.ycombinator.com/item?id=17124967

Did they do it by reducing use of regenerative braking? Does it mean now that M3's brake pads will wear even more quickly? It looks like they already were wearing off faster than expected [1].

[1] https://www.autoevolution.com/news/tesla-model-3-brake-pads-...

That link is about brake performance on a race track at speeds greater than 100mph. I'm not sure what "expected" is in that context, and I doubt the results generalize to more normal usage.

CR article mentions 60mph, not 100.

I was referring to the Autoevolution post, not the Consumer Reports article. It specifically mentions a maneuver at 103.9mph.

Presumably they're comparing their observations to other vehicles they've driven on Laguna Seca.

My suspicion is that the model 3 has undersized brakes for a vehicle of its mass. It's an obvious place to cut the costs.

edit:

To clarify, the autoevolution report noted not only the pads "were cooked" but that the front rotors appeared to need replacing.

Typically only the pads will wear out when it's a simple matter of requiring a more track-oriented pad (higher temp range). The rotors do not get worn in the process, though they may need to be machined (AKA turned) to remove non-uniformly deposited pad material.

When you've got both front rotors and pads ruined after a few hot laps at the track, your options are basically larger diameter rotors and/or a different rotor material. Putting back the same style rotors with a higher temp pad is only going to eat the rotors even faster than before.

These long-range EVs are not light vehicles, they have a very high potential to stress the braking system.

Undoubtedly the brakes are undersized for track racing. But you don’t need brakes that can hold up to timed laps on Laguna Seca to commute 30 minutes to work. Especially in a car where most braking is done regeneratively.

It doesn't require a track to repeatedly accelerate and brake agressively a dozen times in a short period of time, an activity a relatively quick EV practically encourages with its silent operation:

"It makes all of its power at zero rpm, it has perfect traction control, it never squeaks the tires at all, and it's totallly quiet. So, every stop sign you go to, literally like 90% of the times I come to a stop, if it's a straight line ahead of me, I put my foot to the floor and you just get rubber-banded ahead, it's not anti... If I did that with the Testarossa, it was anti-social, I mean people could hear me three miles away when I would get on it on that car so it's just not something that you do every little corner on there." - John Carmack speaking about his Tesla Roadster [1]

I think most people won't expect to have to replace their brakes after a single charge worth of likely shenanigans.

[1] https://www.youtube.com/watch?v=L51eoUFp_YA

It doesn't take a track, just an aggressively inappropriate driving style. In a family sedan. :\

Flooring it between lights is stupid and unsafe and to the extent that it's causing massive brake fade you should detect it (and stop) unless your recklessness is accompanied by incompetence.

If you're driving so aggressively that you manage to destroy your brake pads in a few miles, you're either on a track or an utter moron.

There's no shortage of jalopies throughout automotive history which would want their brakes serviced after a single agressive stop from highway speeds.

That was before they carried a floor full of batteries and accelerated with their peak torque from 0 RPM. This situation is exacerbated with the paradoxically affordable, massive, and quick (in a straight line) EVs.

It won't surprise me one bit if Tesla's first "budget" electric car for the masses cut enough corners to join that list. Skimping on things like brakes in favor of retaining impressive, car-selling acceleration and range figures.

It would surprise me very much if the Model 3 couldn’t handle a single stop from highway speed since it took 4 laps to fry the brakes at Laguna Seca.

There's a big difference between brakes being worn to their service limit and wanting service due to runout caused by overheated pads depositing pad material on the rotor surface.

Most people (I suspect) would take a vehicle in for service when the brake pedal pulsates or the steering judders under braking. This is fairly easy to cause on heavy vehicles with undersized brakes.

We'll see what happens with the Model 3 as the numbers climb and we begin to get a real sense of just what Tesla has been selling to people.

Given that most Tesla owners rarely use the brakes, I suspect you won't find much of a problem. I have 40k miles on my car, and the original brake pads are in great shape.

Possibly, but I presume you're driving a Model S or X?

It's already been noted by multiple reviewers that the regenerative braking of the Model 3 is less significant than previous models.

They did switch to a completely different motor architecture in part to cut costs, the two may be related.

Edit: More significantly there have been numerous reports of Model 3s disabling regenerative braking altogether due to thermal constraints in fairly normal climatic conditions like 68F ambient temps.

What I think is most likely is the Model 3 just won't have particularly good service intervals for the front brakes. They at least appear to have a performance-oriented brake configuration with four piston front calipers, they will probably just wear quicker than expected.

I don't observe very many Tesla owners braking aggressively -- you can easily accelerate aggressively and do almost all your braking with regeneration. Personally I have 40k miles, original brake pads in pristine condition.

From the article it seems like it was an update tweaking the ABS, which is really only triggered during an emergency braking.

A question for those in the know: To what extent do other auto manufacturers make this level of in-line changes?

As someone who has only ever bought used cars, the prospect of substantial in-line changes for a specific model year is quite daunting. With Tesla, it would appear that they are making substantial changes in build technique/quality/parameters weekly...

Performing code changes/updates whenever a car comes into the dealership is pretty standard practice. OTA updates are a technical capability many manufactures have but I'm not sure the extent to which they're used.

My understanding is that almost no one else does OTA updates because their dealer networks demand that any repairs or updates be performed by dealers. The legacy car manufacturers are (sadly) at the mercy of their dealers.

Newer FCA products can update OTA - Our 2015 Dodge Durango updated a month ago.

I know there is a 99.9% chance that my paranoia is just paranoia, but when I first read this I thought tesla just changed the car's behavior to work better on CR's specific tests. The diesel fiasco has scarred me.

The practical question is how good the Model 3 will be when they get production up to the point where the large number of people who want to buy one are actually able to do so.

So to clarify a bit...maximum/emergency braking like this is not limited by regen or by the hydraulic brakes. It is limited by tires, grip, and weight transfer.

Clearly in this case it was limited by a software bug/limitation. Tesla did not push an over the air update that changed the tires, grip, or weight transfer.

(Actually, I guess they did change the weight transfer, by increasing deceleration.)

oh totally agreed. What I find confounding is that up until now the limit on braking performance appears to not have been tire grip but software.

Cars are not a Flappy Bird app: Crashing matters.

This clearly points to a culture inside Tesla of a lackadaisical concern for product safety testing. I say that as a big fan of Tesla and the overall vision Musk and his companies are pursuing.

"We fixed it so it brakes."

Reply on news.ycombinator.com