Report: Toyota To Recall 2010 Prius In Japan, US Over Brake Problem
04 February 2010
The Nikkei reports that Toyota Motor Corp. will recall an estimated 270,000 units of its new third-generation MY 2010 Prius in Japan and the US to fix a brake problem.
The US National Highway Traffic Safety Administration (NHSTA) has opened a formal investigation of the Toyota Prius Hybrid model year 2010 to look into allegations of momentary loss of braking capability while traveling over an uneven road surface, pothole or bump. The agency has received 124 complaints from consumers, including four alleging that crashes occurred.
Some Prius customers have complained of inconsistent brake feel during slow and steady application of brakes on rough or slick road surfaces when the anti-lock brake system (ABS) is activated in an effort to maintain tire traction. The system, in normal operation, engages and disengages rapidly (many times per second) as the control system senses and reacts to tire slippage. Toyota introduced a running production change last month, improving the ABS system’s response time, as well as the system’s overall sensitivity to tire slippage.
The leading Japanese automaker is to shortly file with the Transport Ministry and the US Department of Transportation for the recall to change the control program for the vehicle’s anti-lock brake system. Toyota had denied the existence of any structural defects in the Prius braking system. But its response to quality problems has come under increasing fire at home and in the US, which announced Thursday that it would investigate into the braking issue.
The scope of recalls and voluntary repair offers may expand further because the Sai and Lexus HS250h hybrids, which were released after the third-generation Prius, use the same braking system. The company has begun examining these vehicles for braking problems.
—Nikkei
This condition is not related to either the floor mat entrapment recall or the sticky pedal recall currently in action.
Ford updates brake system software. Separately, Ford announced a customer satisfaction program to update the software of the regenerative brake system of some 2010-model Ford Fusion Hybrids and Mercury Milan Hybrids.
Ford has received reports that some drivers have experienced a different brake feel when the hybrid’s regenerative brakes switch to conventional hydraulic braking. While the vehicles maintain full braking capability, customers may initially perceive the condition as loss of brakes.
The Fusion and Milan Hybrids’ brake system maintains full conventional brakes and full ABS function even as the customer sees visual indicators and hears a chime. The software threshold to transition from regenerative brakes to conventional brakes can cause the system to transition to conventional brakes unnecessarily. The software upgrade will reduce unnecessary occurrences of the vehicle switching from regenerative braking to conventional hydraulic brakes.
Customers with affected vehicles will receive a notice in the mail. Ford is asking owners of affected vehicles to have vehicle software reprogrammed at dealers at no charge.
I'm a Prius fan but:
A. There were enough reports of these problems in the 2nd generation Prius
B. One thought they were fixed in the third generation Prius
Posted by: kelly | 04 February 2010 at 11:39 PM
Every ABS-equipped car I've ever driven regularly has exhibited this problem to some extent. Living in a city with very rough roads, I've had plenty of experience with the issue. It's not a hybrid-specific issue. When ABS sees the first indications of wheel slip at a single wheel, it cuts braking force at all wheels for fear of braking the two sides of the car unevenly and putting it into a spin, and waits for about a second to see how the vehicle responds before allowing firmer braking again. It's a logical and helpful reaction on icy roads, but when you hit a pothole while braking on a dry road and the brakes don't give you all the braking you wanted for a second afterwards, regardless of how much pedal pressure you apply, it's problematic. The worst car I drove/owned with this problem was a '94 Taurus, which did it frequently and dramatically, and felt like it was taking forever for braking to return. It provided plenty of scary moments. People do complain about the same feeling in the '04-09 Prius (I have an '07), but what I have experienced in that Prius isn't quite the same. The Prius has a hairtrigger reflex about cutting regenerative braking when you hit even the tiniest bump, but if you push the pedal a little more, full friction braking is available without hesitation. It's only scary to the extent that you have to add a little pedal pressure unexpectedly, but at least you get a response in actual braking to the extra pressure. I have not had any trouble with it triggering a full-blown ABS response to ripples and potholes in the way that my older ABS-equipped cars did.
I would expect, if people are crashing the 2010 Prius in cases where they didn't crash the '04-'09, that this is the reappearance of the ABS over-reaction condition that has appeared in other ABS-equipped cars over the years, and not just the less-than-graceful transition from regenerative braking to friction braking that the earlier cars exhibit. If that's the case, this would not have been a recall without the current panic. I'm not aware of any other car that's been recalled for this issue, and that Taurus certainly tried to crash itself a few times on me!
Posted by: Wes | 05 February 2010 at 06:15 AM
.
Round three of Toyota vs. Obama Administration. Since GM is now a government agency it's high time that the government go after it's competitors, right?!?!
Toyota you're in for some of the most hateful rhetoric you have ever heard. You are not properly unionised.
.
Posted by: The Goracle | 05 February 2010 at 06:27 AM
Wes & Goracle:
You may both be right on this one.
Ford's HEVs have the very same problem but it is kept under the carpet. Is it because it is one of the local Big 3? Our politicians and media people are playing a dangerous game on this one.
My wife and I have been driving trouble free Toyotas for 20+ years and feel very secure. However, our Toyotas are built in Japan with Japanes parts. That may be the real difference in some cases.
Posted by: HarveyD | 05 February 2010 at 09:32 AM
This is the downside of computer control. I helped pioneer microprocessor design in the early 70s for engine computers, now they are used for ABS and other things. They should assist and NOT take over.
Maybe the pothole problem could be corrected with a suspension sensor that can tell a pothole from an icy road. At any rate, the problems may be fixed by a software upgrade and not a mechanical fix. You are judged by the quality of your software as much as hardware now.
Posted by: SJC | 05 February 2010 at 12:14 PM
Agree with SJC. Microprocessor should only be allowed to make small changes but not taking over completely with respect to the throttle and the braking function. There should always be a direct mechanical linkage from the driver to the throttle plate.
I wonder why the 2010 Prius can't just inherit from the 2007 Prius the proven hardware, microprocessor, and software with respect to braking function. My 2007 Prius functions flawlessly with respect to all functions. Perhaps a very small percentage of cars has defective computer chips and only malfunction intermittently. May be better quality control with respect to all the IC chips. May be all recalled models of Toyotas should have their computer chips thoroughly checked at the dealers. Technology exists for comprehensive testing of IC chips for weaknesses that may lead to intermittent malfunctions.
Posted by: Roger Pham | 05 February 2010 at 02:20 PM
It can be a function of complete debugging and how many test vectors they had. You can not test EVERY possible case, but you have to have enough of them to cover a lot when it comes to something this critical.
Posted by: SJC | 05 February 2010 at 05:12 PM
Regardless of how thorough you think you have tested software and applied coding standards (such as MISRA or validation per DO-178 for example) there will always be some type of error as the code complexity grows.
The more you try to do and the more capable you make a piece of software the more likely you are to inject a bug. If you are using a compiler for a high level language you have to add in the possibility that the compiler might create strange assembly code (regardless of optimization settings).
So now we see what happens with throttle by wire...I can't wait to see what happens when we are trying to control multiple individual electric motors with brakes that are mostly integrated into the electronic system.
For an idea of what it is like to create a piece of software (disregarding the more crucial errors caused by improper requirements specification or architecture) picture yourself learning a 2nd language and then writing 1, 2, or even 5 million lines in a book in that 2nd language without making a single error along the way. Sure, the compiler checks basic syntax and you can apply static analysis and dynamic analysis tools, but that still doesn't guarantee fault free software...
Posted by: Patrick | 06 February 2010 at 02:14 AM
Roger said,
Perhaps a very small percentage of cars has defective computer chips and only malfunction intermittently. May be better quality control with respect to all the IC chips.
I would say that is unlikely. A defective micro is relatively easy to detect and probably wouldn't malfunction "intermittently". It would be just as likely to say that a stray blob of solder is bouncing around in the ECU and causing random intermittent shorts.
There are several classes of non-deterministic software errors that are nearly impossible to track down with standard testing (though the experienced person may look at the symptoms and have a few guesses as what may be happening). You could have a race condition with interrupt vectors and a shared resource because the RTOS is faulty, or you are setting up faulty methods for accessing resources (this would be similar to problems with mutexes and semaphores in desktop programming). You could have memory management issues (allocation, deallocation, fragmentation). You could have implemented a function in a non-standard way with undefined behavior where the processor is just "guessing".
Posted by: Patrick | 06 February 2010 at 02:22 AM
Gentlemen;
Don't confuse designing s/w with writing or coding it. Double and triple reduncy is also implementable in s/w as it is in h/w.
Toyota chose to use its old mechanical throttle control as the redundant path, and then chose to disconect and lter remove it, leaving no redundancy. That choice of false economy is a software engineering design mistake, not a coding error, or bug.
Toyota and the other automakers, who say "There but for the grace, go I..." can ponder and ruminate at liesure, the consequences of such cost-saving mistakes, in critical systems.
Posted by: Stan Peterson | 06 February 2010 at 07:26 AM
It could be RTOS, but more likely a logical error. A case may not have been tested and when it occurs the state diagram goes to a default condition or meta state.
Imagine a pothole, with icy road and 40 mph crosswinds with 4 passengers. The ABS, traction and stability controls are all trying to deal with each condition, but do not deal well with ALL conditions at the same time well.
Posted by: SJC | 06 February 2010 at 11:55 AM
@Patrick,
Thanks for your input. However, these intemittent glitches happened more than once in the same cars while not at all on the vast majority of vehicles. The following is a quote from:
http://www.hybridcars.com/safety/hybrid-owners-question-toyota-acceleration-explanation-26579.html
"KIM 3 days ago
we have a 2010 Prius and on several occassions the car has accelerated when the brake is depressed. This has happened mostly at startup. It was pretty scary to the driver. We took it to Toyota but they said they couldn't find anything wrong with it. We left it a couple of days to see if they could replicate in the morning but to no avail. ...
This points to intermittent computer hardware problem, may be due to temperature changes, moistures, radio interference, etc on IC chips that may have too slight a defect to be detected on routine testing. If it has been a random software problem, it would not have recurred in the same car and not at all on other cars having exactly the same softwares and hardwares. If the dealer found no mechanical defect, then the responsible electronics in the malfunctioned car should have been replaced to see if the intermittent problem is resolved.
Posted by: Roger Pham | 06 February 2010 at 10:39 PM