Edmunds dealer partner, Bayway Leasing, is now offering transparent lease deals via these forums. Click here to see the latest vehicles!
Options
Popular New Cars
Popular Used Sedans
Popular Used SUVs
Popular Used Pickup Trucks
Popular Used Hatchbacks
Popular Used Minivans
Popular Used Coupes
Popular Used Wagons
Comments
I repeat: I have seen the exact problem I describe and my solution involved removing a defective alternator that was emitting substantial high frequency voltage spikes.
Also, of course, if you've ever worked on a power buss for a fighter aircraft or a helicopter, you would never dream of stating that a commercial-grade computer system (even one ruggedized for a car) would handle what can happen. Making military avionics handle it is both expensive and difficult; we don't spend nearly as much on our cars to make them reliable.
Not necessarily a true statement. The cost of "certifying" military aircraft components, or even commercial aircraft, must be spread over a relatively few number of annual sales vs the automotive industries millions of sales annually.
A GOOD AUTOMOTIVE TECHNICIAN NEEDS THIS INFO TO FIND ANY PROBLEM WITH A CAR. DO NOT LEAVE THIS TO ENGINEERS.....THEY ONLY KNOW HOW TO DESIGN IT NOT TROUBLESHOOT IT. BY THE WAY WELCOME TO "DRIVE BY WIRE" A THROTTLE CABLE AND SPRING PROBABLY WOULD HAVE AVOIDED THIS PROBLEM
With regard to Edmunds' $1 million prize for demonstrating Toyota's unintended acceleration, I think it will be difficult if not impossible to duplicate this problem in a controlled environment unless Toyota allows its ECU source code to be analyzed by contestants in real-time during EMI / RFI interrogation. The software code will have to be modified anyhow, and the more minds working on this problem, the better. If Toyota declines to allow analysis / examination of their ECU source code, that will speak volumes about the company mindset / philosophy as well as making the contestants job of solving Toyota's problems much more difficult.
I bet accelerators were stuck a lot more when using cables and return springs then they do in the newer cars. Springs broke or came undone and cables were covered in dirt and grime so they didn't return fast enough. You didn't have the media or computers broadcasting the failure all over the world, and people knew enough to stick your toe under the pedal and pull it back up to slow down so you could fix it.
However, there is a common algorithmic mistake often found in the software or microcode that can result in bad calculations. This would be much more easily checked by examination of the source code rather than your proposed reproducible experiment prize. Furthermore, even extremely good engineering and programming staff make this mistake from time to time and once made it can be very difficult to notice in routine software testing.
The mistake, which varies in detail by computer language, is essentially this: A variable is declared as a signed variable. That is, it has a bit in its numerical format that indicates whether it is a positive or negative value. Then signed variable is used as an accumulator or is otherwise involved in unsigned operations that overflow into the sign bit. Once this happens many algorithms involving the variable may become unstable and generate nonsense signals from otherwise non-defective electronic components.
Now I know you haven’t even written up your prize rules yet. I’d enjoy the prize if this turns out to be the problem, but I am much more concerned that someone, like you, with market clout floats the idea to Toyota so they hear it. Once focused on the potential for this error in microcode or software, it should be easy for them to see whether or not it is true. That could possibly save lives. So whether this fits your prize model or not, please do what you can to get this idea to them. Save one life and I’ll be happy enough.
And if this is not the problem not much will be wasted in taking a look.
Regards,
Mark W. Farnham
Frank Kilbourne
Palm Desert, CA
Some kind of external signal is triggering the acceleration sensor even though the driver is not accelerating or stepping brakes. (even with hand brake is on)
External signal might be any radio signal, cell-phone signal, RF transmission signal whatever.
Here is my suggestion to find out the cause.
1. gather video clips of unintended acceleration
2. try to find out all possible wireless signal sources in the scene environment.
3. test all the signal with the acceleration sensor, one at a time, 2 at a same time, 3-4 sources at a same time.
Then there will be a clue to answer.
The dealer I took it to was unable to duplicate the issue, but after about 2 years I finally figured out how to duplicate the problem. If I turned on the A/C to full blast, then the 3 second acceleration delay happens. But even this method did not recreate the issue all the time, then I noticed that each time I had the issue, the cars Radiator Fan was on. Then about 6 months ago my car died and the dealer determined that my battery died after only 2 1/2 years.
Which leads to my humble conclusion, the Base model Camry's are underpowered for one, which causes the Battery to die quickly, as the alternator can't keep up with the energy requirements. The CPU of the vehicle may not be getting enough power, so it is unable to process the input quickly enough; alternatively, it could be that the Radiator running creates multiple EMI fields throughout the vehicle, which cause signal data loss, so the CPU does not have the information it needs to know what to do. Since these vehicles are underpowered, it keeps doing what it was doing before, which is not accelerating.
How does this related to the sudden acceleration issues? Well, for one, I would be interested in finding out if the people that have experienced this issue had the A/C running and whether the Radiator was running, if so, perhaps there is an EMI issue or some other issue related to the Radiator. Second, are most of the vehicles that experience these issues Lexus/Toyotas with V6's. If so, the alternator and engine are much stronger and are able to handle the power requirement of the CPU, but if the CPU experiences EMI interference, instead of defaulting to slow behavior, as in my underpowered Camry CE, they experience the opposite issue, and have sudden acceleration, as the CPU is getting high voltage data that tells it the last thing the customer did is press hard on the pedal.
This would coincide neatly with what I read from someone who had the brake override firmware fix installed but still experienced sudden acceleration, but after 3 seconds, the car slowed down (brake override reached through the CPU after 3 second delay and cut fuel to the engine).
This is all just a theory but this issue should be reproducible given some intelligent engineers at Toyota that keep an open mind. My guess is they already know what the issue is but don't want to fix it, due to the expense, thinking it is better to just install a brake override firmware that masks the issue a bit, but does not completely resolve issue. An idea on how they could fix the issue if they wanted to spend the money would be to install a more powerful CPU that can multitask better and encase it in an EMI protected enclosure, but even this may not completely fix the issue.
Personally, I've always hated drive by wire (HATE!!!). I remember almost buying an Accord a couple years back, until I found out they use drive by wire also (almost all cars use it now). I work with computers everyday and know how ghosts in the machine can cause problems, even between identical hardware (maybe it's a quantum fluctuation issue, but now I'm babbling). I loved my old Corolla that wasn't drive by wire, too bad they don't make them like that anymore
A common assumption, but not necessarily true. It's possible to have a subtle bug that only surfaces under certain specific conditions. If those conditions are rare, the bug manifests itself only once in awhile. Look up the THERAC-25 case for an example of this. That was a medical device that had rare, but deadly malfunctions that were very difficult to reproduce because a very specific set of circumstances triggered the bug.
In fact, I would say everyone interested in the Toyota acceleration problem should read the THERAC-25 report. It's both technically interesting, and an eye-opener for the human factor in these kinds of situations.
-Andrew L
In a car with an automatic transmission, when the driver presses the brake pedal, the software automatically closes the throttle.
Simple and testable. Should be easy to verify and validate that it works as designed.
BTW- no Japanese cars currently have this feature in their software. One could/should ask why not?
This IS a programming problem and I can reproduce it about 80% of the time: 1. Vehicle has to be in cruise control. 2. Results are most consistent if the speed is over 60mph. 3.The vehicle has to be climbing a hill or long overpass. 4.Tap the cruise control stalk down once. 5. Once the car slows a bit tap the stalk up a couple of times. This will result in a hard downshift and a burst of acceleration that is equal to a one-third to a one-half-to-the-floor punch on the accelerator. The vehicle shoots three to five miles per hour past the desired speed (one tap should result in a one mph gain) then it drops more slowly to the set speed.
It seems the speed/tranny/cruise information takes a couple of seconds to get its act together. This vehicle is rated to tow 5,000 lbs. It’s just a guess, but I’ll bet that figures into the equation.
I have a 2008 4wd Limited with 31,000 miles on it. Even though it is not on the recall list, it has the problem. When it was new, it would downshift at the slightest rise or overpass, but without the acceleration problem. I voiced my concern at the time that the transmission would wear out prematurely. After a while, Toyota had a programming fix for it. It would appear that they just need another one.
Thank you for contacting Toyota Motor Sales, U.S.A., Inc. We appreciate your consideration and hope to have your email addressed as quickly as possible. Our current office hours are Monday through Friday from 5 AM to 6 PM and Saturday 7 AM to 4 PM Pacific Time. If you need immediate assistance, we recommend you contact the Customer Relations Manager at your local Toyota dealership.
Discussion Thread
Customer (Patrick) 03/03/2010 02:22 PM
I do not have a question. I have more of suggestions or possible solution to the issue of out of control acceleration. I could be totally off base, but I believe the issues lies in the Cruise Control system. The is an acceleration button on many steering wheels on the toyotas that are having issues. I believe that the out of control acceleration is starting in that electrical system or signal. I believe it is activating some how and the signal that is sent to cause the acceleration is continously being sent or the system is interpreting something as being that signal, therefore accelarating. What ever that signal is, it is also overriding the brakes. I would check the Cruise control system acceleration module to see if there are any issues in that system. One more thing. I love my Toyota FJ Cruiser. I dont care how many safety issues or recalls Toyota has. A recalled Toyota is better than an American Vehicle any day of the week. Just remember what you got you to where you are, go back to that and you'll be alright.
In December when reading about Toyota surges I found that Toyota owners described their surges much the same as I describe those in my Tiguan. So - I messaged VW, Toyota and NHTSA asking them to compare car parts and find something "in common." No one would listen to me. Last month, quite by accident, I discovered that my Tiguan transmission and those in the recall Toyotas were made by the same company in Japan. I wonder if there could be a common denominator here. I do not know enough about transmissions to figure this out - but maybe someone else could get a parts list for 2009 Tiguans and the recalled Toyotas and see if there is something in common that could be causing these very scary surges. I have driven my Tiguan only 8,000 miles and have experience 14 surges.
For instance a race condition that results EVERY time the brakes are applied absolutely simultaneously with the CC set/accel function. EVERY vehicle might have software subject to the condition but in the real world how often would those exact conditions occur.
Maybe because that it's a patently FOOLISH effort.
If the "software" happens to be "out-to-lunch"...??
Which may very well prove to be the case with Toyota/etc's UA.
However, this is under "normal" conditions not a failure mode. The fact that the car continues to accelerate even with the brake pushed makes it doubtful that sticky accelerators or floor mats are causing these problems.
Also, whoever said to add a "tps 1 and tps 2" are you sure that these cars don't already have these? I have NEVER seen a DBW system that didn't have two APPs and two TPSs. The entire system is built to be redundent for this exact purpose.
also, if the problem was related to the cruise control wouldn't stepping on the break cause the cruse control to shut off and thus ending the acceleration?
Finally, DBW is important for one main reason - traction control. Traction control saves lives and thats the bottom line. For this reason alone the big manufacturers won't ever go back to a cable system.
I can guarantee, though, that it is not. Cruise control when it does this, isn't sticking open and the system always turns off if you hit the brakes. You're confusing a few second surge in power due to the system kicking in suddenly (unexpected but normal function of the resume feature) versus flat-out 100% throttle until you physically turn the car off or it grinds to a full stop(engine still racing at full throttle while stopped).
Cruise control might give you a few scary moments like you described, but this really is something causing the ETC to jam wide open and stay there. And that's just not the same thing.
Note - this problem is really only dangerous in automatic equipped vehicles. With a manual, you just put the clutch in in such a scenario and brake with your right foot. Or turn off the engine. Unlike automatics that often lock out removing the key or turning off the engine unless it is in park (or maybe neutral sometimes), manuals can always be just simply turned off like this.
Also, many cars with manual transmissions(not sure for Toyota though) have manual throttle cables since the technologies don't work well together. Yet another reason manuals are better. You get full control of your potential deathtrap in a worst case scenario.
Inverse of ABS functionality.
The problem must be solved in 2 steps, what's built is built, first is a software
upgrade to the current with a safety feedback this will combine current
and existing sensors. Stage 1, update the software to monitor the BOO
(Brake on off switch) and the engine RPM feedback, If the BOO signal is
true then engage the converter clutch if the RPM and Speedsensor do not
change state, the engine will stall, although it is not the total fix it will
be a redundant safety. Step 2, we get away from the pulse width modulation
of engine control, go to analog, use 4-20ma and set a range monitor
signal out, BOO, and RPM, Use the PCM to modify the signal with respect
to the inputs, if 4-20 does not modify, place in FMEM and turn on check
engine lamp. Customer gets home, to dealer, and family is safe. Automotive
needs to become positive side switching vs. ground side switching, the controls
are more robust and it prevents undesired activation due to an unexpected
ground.
It is the cruise control computer. It could be a defective sending unit or defective chip or other component in the computer itself.
I believe this could be the cause due to my own experience with cruise control. I had the car in cruise control going 60 on the highway and as I came into town and up to a line of cars at a stop light I accidently bumped :"resume", causing the car to lurch forward, nearly creaming myself & 3 cars in front of me! (I braked in time).
Thank you,
Deb
i was a tech for about 16 year and know that the internal combustion engine principle hasn't changed. what has changed is the way that it's controlled or managed. instead of mechanical controls today we have electronic management.
something has changed in the models affected. new vendors for OEM parts (throttle bodies, electronic controls, routing of electronic wiring, proper shielded wiring etc). take a closer look at the basics could a throttle body be binding up with temp change throttle position, premature wear, warpage? watch for voltage spikes? i cant remember what car company it was but a few years ago they where the first to use throttle by wire and had the same problem.
i know today it's so easy to hook up a scanner and say"nope nothing wrong with your car the computer says so" and the customer drives off and comes back a few minutes later and screams "it's acting up again" !
i am not an engineer but sometimes you just have to take another route to find a problem and most times it's not what it seems to be.
i hope that toyota will read these postings and maybe find a fix for this problem, when the problem wont go away look at other factors that could be.
But history can show us the answer to this seemingly unsolvable dilemma: Alexander the Great was once faced with the challenge of untying a knot called the Gordian Knot; which was so complex that the wisest and most skilled people in the kingdom had been unable to untie it. After considering the situation; he drew his sword, and cut through the knot. And that simple solution was clearly successful, in a situation where all the best minds of the era had failed.
OK; suppose the unintended acceleration problem is sometimes caused by solder whiskers; sometimes by data overload from simultaneous contradictory inputs, sometimes from drivers whose boots are too big, sometimes from a sticking throttle butterfly or throttle pedal sensor, and sometimes from the RFI generated by a passing illegal CB installation. That would be one intimidating issue to deal with by a single solution. But I suggest that there is a common factor in ALL of those possibilities (as well as any other possibility which can be thought of): ALL causes for unintended acceleration must use the vehicle computer(s) to open the throttle valve, lock the transmission from being shifted into neutral, prevent the brakes from being applied, and block this out of control process from being shut off with the "stop" button.
So, I propose mounting a big, red panic button on the dashboard; which is connected so that when activated, it cuts off the power supply to the ECU, and every other pertinent computer on the vehicle. When the power to the computer is cut off; the power which opens the throttle valve is cut; the power which runs the electric fuel pump and triggers the fuel injectors is cut; and the power which creates the ignition sparks that fire the fuel in the cylinders is cut. So the car cannot possibly continue accelerating; regardless of why it started doing so in the first place.
For those of you who prefer an automatic mechanism (to cover the case where the driver panics and freezes to the extent that they cannot push a "save me" button); this can also be done with an accelerometer (which detects vehicle acceleration) and a switch which is mechanically connected to the brake pedal (and/or to the vehicle speed sensor). If acceleration is detected while the brake pedal is being pressed for a preset length of time; it then triggers a relay which shuts off power to the computer(s).
Please send the check to my address of record. Thanks; (and I will keep posting on the public Answers forum regardless). Joel
Much seems to be written already about possible cruise controls. This may be a good start. The question should start at "What affects a car to accelerator?" One, two or many things? I would hazard a guess that there are two (maybe three) electronic segments that may initiate the run away problem.
Thinking simply, it appears the two primary contributors to the run away problem are 1) the accelerator and 2) the cruise control. Either could run uncontrolably and cause an uncontrolable acceleration.
I am not sure of the actual circuitry but consider globaly how the two may be the root cause. Simply put, each system must introduce or sense a difference voltage and tries to zero out this voltage to cause a reaction. In the case of the accelerator, when at rest, there is no error voltage introduced to require acceleration. When the accelerator is pushed down, the system senses a difference from the accelerator at rest, and an error voltage is caused, which is inroduced to a servo system. The servo system attempts to null this voltage and once the accelerator stops moving, this position will re-establish the driver's selected speed. The system is in equillibrium when the error voltage becomes zero. The car is now traveling at the driver's accelerating setting. When the driver removes his foot from the accelerator, the system senses an opposite voltage difference and the servo system returns to idle setting.
In the case of the cruise control, the primary cruise speed is established when the cruise control is turned on and "set". The error voltage is at zero and no signal change given to the servo except maintain position. The accelerator can now be released and speed maintained. The controls of the servos normally permit various changes, like retard, accelerate, or disconnect.
Interestingly, the accelerator can be used to over ride the cruise control by stepping on the accelerator to inject and error voltage and cause the car to accelerate to a higher speed. If the driver removes his foot, the system defaults to the established cruise control setting.
Of not should be that both systems, accelerator and cruise, feed into the servo system that is the actual speed interface for engine control. The servo normally need an input to change engine speed.
This assemption then would result in trying to identify which of these entities are the root cause of random acceleration.
Some discussion of early control instability makes my return to my vacuum tube days. Back then an amplifier circuit, if pushed too hard, could be caused to oscillate and run away. The early designers learned to control the tube from oscillating by introducing fee-back out of phase to control the undesired effect and result in a stable amplifier. In our case were, the servo (or what ever you name it) appears to have in certain cars a capability to run wild (oscillate) and once started - no input, brake, releasing the accelerator or turning off the cruise control would affect the wild running servo . . . and continual increase in speed resulting.
The servo in your system's design is such that a protective feed-back appears missing. Back in the 50s we used a bode plot (zeros and poles) to extablish this affect and prevent it. I am in my 80s and could not remember the mechanics of design - now.
As a side note, I know we are in the digital world, but the theory of oscillating amplifier in this, once it runs awy can only be stopped by removing power - or introducing feed back.
Why certain cars? Simple - you haven't characterized each "servo system" and ensured that it can't run wild.
Not to mention that anything can and will fail given the circumstances are 'right'.
Those circumstances are by no means confined to the end product nor the circumstances under which that product is being used mind you.
In my case our Toyota Matrix developed a grunting/rattling sound when the cruise control was activated about a year ago. Which problem has been fixed by Toyota swiftly, with the remark that this was something they had never experienced before.
Fact is that recreating just the right circumstances, including that of the production-line on Monday morning for instance is nearly impossible and therefor there's very little anybody can do about a minor percentage of products that fail under a minor percentage of circumstances, hence out of millions of vehicles or parts thereof being operated and produced under countless circumstances a 'few' will fail.
In my case disengaging the cruise control when the problem occurred didn't help, yet first after pulling off the high way, turning off the engine (hence resetting the systems) 'it' was solved, needless to say I opted out of the lazy drivers function until Toyota replaced suspicious parts....what went wrong we'll never know.
Of course everything needs to be done to make these cases as few as possible at any given time and circumstances....yet fail proof doesn't exist.
It's why NASA had a back-up for everything including the back-up and a triple check on even the parts and humans that never see space at all.
And still then.....
Until the problem(s) in Toyota's (and other carmakers) case are clear it might be smart to bypass the electronics and mechanicals that might be at fault all together and even have an option to opt out of such devices on new products.
After all, install an automatic brake system and some vehicles will come to an abrupt stand still when least favorable, given the circumstances are 'right'.
In the meanwhile keep up the good work and hope for very few 'Mondays" in the process of doing so.
P.s., thee more gadgets, the more 'Mondays' will pass the quality check, even on any other day of the week.
Bottom line in this is, do we need all the gadgets (they have been failing ever since they came to market, even if it's 'just' one out of countless) and if so where is the back-up or the by-pass?...and even those can and will fail eventually under the "right" circumstances)...too expensive right?
I'm convinced this particular problem can be 'solved' by overhauling the cruise control systems, let's leave 'taking off' to NASA.
Yet still then, Murphy's law doesn't bypass the automotive industry and therefor one can't put any blame on the brand Toyota (or any other) because Murphy pays a 'visit'.
Especially when electronic circuitry is part of any given system it will act up sooner or later, at times a lightning strike nearby is all it takes, leaving no traces not even a hint of a micro surge that triggers a 01100101 input rather than the 01110101 which was expected.
Now include a little moisture, some dirt and wear and tear, a path for a leaking single electron and voila...
Any system is only fool proof when not installed or used at all.
All "we" can do is make sure the system-designers and producers are taking their jobs as serious as "we" trust they do.
Now if only all tweeter users would refrain from doing so until they are in "P" the roads will become way safer even with the odd take off of any make of car.
Maybe an unwanted signal-shield solves most of these problems?
In the meanwhile I personally urge everybody to refrain from using cruise and tweeter controls...there's no need for either of them after all.
sources like garage door openers.
Driving this morning, I noticed I normally leave my cruise control on and got to thinking what would happen if the acc/resume shorted and locked out all other systems... a bad day.