It appears that the ECM is triggered into a momentary latched condition where it commands full throttle. Margin testing the ECM voltages and doing injected noise tests would seem appropriate.
You bring up an interesting point. Is there some sort of EMC (Electromagnetic Compatibility) standard for automobiles, similar to Mil-Std-461 for military gear?
There is one thing about noise trouble shooting: no cases are the same and satisfying low probability failure rates is extremely difficult. It is statistical problem where one has to accelerate the failures by stress testing so cause and effect improvements can be made
Exactly. Consider how many Toyota are out there, being driven day in and day out with no SUA incidents at all. What's that, hundreds of millions or billions of throttle operations in a year, yet only a handful of SUA incidents. That's a lot better than 6 nine's reliability.
Do you and bnet rank RFI/EMI lower than other more probable causes?
I think that with nothing known about the throttle circuitry it is hard to favor anything because the options are infinite. But since I had 6 identical incidents in a different environment each time I would have to guess that in my case the trigger was internal to my car. In trouble shooting such a low probability event it has been my experience that you have to make a list of various causes and try and define a stress test to see if you could invoke the problem for each item on the list. It could be triggered by voltage noise due to some unknown condition, a ground loop resonance, as mentioned before software glitches --- noise doesn't have to be only voltage.
And your right the SUA incident would be very difficult to replicate because you would be working your way through a long list of suspects hoping you could stress the electronic throttle system in a way that would show the problem. You might be wrong on a list of 100 suspects and find it is the 100th idea tested.
I cut my teeth on the magnetic drum system on the Univac 1107 in the 60's. The system would run intermittently and then quite. A sudden rush of cold air on my neck when the air conditioner turned on that coincided with the failure gave the idea that it might be temperature related. Running a hair dryer on the drum intake showed that the clock pulse interrogating data would move over 4 data pulses with the hot air and move back when it cooled. An overnite fix with a drum temperature control solved the problem. It took about 6 months to nail. At the time we were running 7/24 with management covering shifts trying to push out the 20 first computers systems on the floor that weren't running because the drum didn't work. Tremendous pressure to say the least ---- but the most fun in trying to figure it all out.
For this problem one would have to have a lot of information for the Toyota electronic throttle which I would question the feasibility of getting. Does anyone know if it is possible to get this type of information?
I've read a lot of the posts and there are a lot of good ideas out there
You read it here first. It is quite possible the Toyota's ECU modules are being upset by EMI noise generated by an arc welder in the vicinity or a high power source. Arc welders run typically at 220 VAC and can provide hundreds of amps of current (very high power) to fuse metals together at thousands of degrees. When an arc is first generated, a huge EMI power pulse / spike is emitted and this pulse could upset a vehicle's ECU computer IF it was not properly filtered. The fact that the problem is so elusive lends itself to this kind of theory. However our '01 VW Beetle Turbo loses power for 2 to 3 seconds along a certain stretch of freeway where we live - not always, but it has occurred enough times that I've taken notice when I driver the car. So it's not just Toyota that suffers from ECU interference. In closing, it is INSANE that any auto manufacturer would implement a throttle-by-wire system controlled by software. These outfits deserve what they get. I say: go back to the tried and true: cable actuated throttle system - and JUNK the software.
Below is a link to a story relating to my previous blog today about media hysteria, and possible story "planting" by some third party concerning the Toyota unintended acceleration issue.
And in other news, Toyota is flatly denying that ANY problem with its software exists. Looks like that award will never come at this rate short of some miracle, because Toyota is not letting anyone into their code(or that block box either)
The sure fix for this problem is to go back to the cable and spring and get rid of the the electronic accelerator. Why beat your head against the wall looking for a cause, just fix the problem before any one else dies!!!!!!!!!!!!
I agree with you my friend but I go a bit further => cellphone radiation is everywhere in the air.and one thing NO ONE KNOWS to this day is whether in all that airborne moving mass of invisible cellphone radiation sooner or later you're going to have some COMBINING OF ENERGIES into a random and stronger energy.
Extrapolating on that is your idea because these sensitive electronic systems in Toyota's vehicles are like mini-antenna systems giving all that cellphone energy a place to call home. Look a second please at how many of these incidents are happening as the vehicle is in motion traveling horizontal to the cellphone energy, meaning their electronics systems are in effect cutting across artificially-created gravitational lines of flux.
Cellphone radiation may affect humans less since we don't walk 60 miles per hour everywhere we go. Right now we are still using Internet Protocol #4 (IPv4) but there are worldwide communications companies pressing hard to get people to accept Free or Low Cost international communications and phones which extra usage would then force IPv6 to be needed.
Why? => Because they are pushing the limits of IPv4 to force the world community to turn the dial up to Internet Protocol #6 (IPv6). IPv6 may very well cause more Toyota crashes from runaways. I heard a report this morning the same thing happened to a guy driving a Prius so this is no longer Toyota's problem alone. ** The wonderful Toyota Corporation cannot be held responsible for the communications industry hot-wiring the earth's atmosphere stuffed full of stray EMF radiation. **
This problem can be traced back to a single common element: the new cruise control wiring harness design, specifically the contact points. It is simply too small and shorts out without any evidence of a failure because the contacts are too compact. This is not "rocket surgery", just basic analog electronics failure.
I have several angles regarding this sudden acceleration problem. Here is the first.
CORRUPTED SOFTWARE OR MALICIOUS CODE INJECTED The firmware has either become corrupted or contains a malicious code.
When the various sensor inputs present to the computer their signals in a particular order and value, the software generates a full-throttle command.
Causes: 1. This could be a result of an intentional or unintentional error in the computer's code as it was written but unlikely as Toyota has repeatedly tested their software.
2. This may also be caused by the firmware having become corrupted by any number of sources.
3. A malicious code has been unintentionally loaded into the computer's memory after the vehicle has left he factory. There are several possibilities for this scenario: a. Routine service of the vehicle by an authorized dealer. When the technician connects his/her OBD test equipment to the vehicle's port, the vehicle's software is unknowingly re-written. b. Annual inspections also require a technician to read the vehicles codes via the OBD port. Again, a malicious code could be written to the vehicle's computer.
The equipment used to read the vehicle's OBD port has changes or upgrades loaded onto it on a regular basis. The source of the malicious code: a. Updates downloaded from the Toyota website b. Updates sent by Toyota via CD-ROM c. The manufacturer of the equipment's website d. Updates sent by the manufacturer of the equipment via CD-ROM Any of these update sources could had been hacked.
Check: 1) Satellite Radio/GPS for intended or unintended computer intrusion into drive by wire 2) for Cell or PDA interference with drive by wire functions
How does one 'inject malicious code' into the Toyota ECU processor to cause the problems you describe? Answer: It doesn't happen that way. Embedded code cannot be modified 'on-the-fly' while the car is being driven; that is to say that that particular task requires a (laptop) computer with Toyota's proprietary interface software, a suitable cable interface to connect (OBD II) with the Toyota vehicle to update / modify the ECU's code. What is more than likely happening is that EMI / RFI noise is being radiated into the Toyota ECU and causing the computer to go into an open-loop mode. A device such as an arc welder could be causing the problem; they run at 220 VAC and provide 100's of amps of current instantaneously. The high-power spikes / wideband RF noise emitted by the arc-welder or similar high-power source is called EMI radiation and can affect computers and other electronic devices - such as a pacemaker. Once the Toyota computer is locked up there is no amount of installed software patches that will solve the problem. Toyota has three serious design oversights to deal with: a) ECU software, b) INSANE fly-by-wire throttle / brake design, c) Insufficient electronic filtering on its ECU's I/O ports.
Whether Toyota "good" or "no good" remains to be verified to the satisfaction of everyone's curiosity. The difference in the link to the news story that I pasted here, and the news link that you pasted was, in my link, the guilty party was caught red-handed with verifiable physical proof of the story tampering. In the link you pasted, it's simply another Toyota owner claiming unintended acceleration - no proof, physical or otherwise, other than the claim itself, and the car obviously speeding down the highway.
One thing that I think needs to be corrected after the claims in this latest news story, and others is, the term "unintended acceleration." I don't believe it is any longer the proper term to match the claims being made. If my vehicle accelerates "unintendedly" that term connotates an inconvenient situation that I can stop or control to some degree at any point if I wish too. I think the proper term to match the claims that are being made is "uncontrollable acceleration." But, if the descriptive term is upgraded to the adjective uncontrollable, then that raises the question as to why, with all these uncontrollable acceleration incidents, has there been no collisions, injuries, or even worse, fatalities.
If multiple vehicles, the number of which has been described in the claims, begin accelerating uncontrollably, it is a matter of simple odds, with the heavy road traffic, non-linear roads, and multiple stop signs and red lights, that the above consequences would result. Yet, none seemed to have happened so far. Of course, odds do allow for the possibility that a car could go careening down the road uncontrollably without colliding with something, or injuring or killing someone, but I believe astronomical would be the term to best describe them.
Again, I am not saying an uncontrollable acceleration problem absolutely does not exist in Toyota vehicles. But, there could be many reasons that a Toyota owner or driver, could or would make a claim like this. I'll let your imagination decide what those reasons could be. For myself, I am saying that many things about this situation simply do not add up, and I offered a possible theory as to what could be behind this issue. But ultimately, I will reserve my final judgement for when all the pieces of this puzzle fall into place, as I'm sure we will all be compelled to do.
My 2007 Corolla does it so often, I've become used to holding down the brake a bit harder at every stop. What is happening is that the engine rpm can pick up unexpectedly while at a stop at the same moment that the radiator fan kicks in. What is occuring is a rapid power load situation is causing the alternator line voltage to try and stabelize current draw by increasing the engine speed. This increase will vary according to how many other electricals are in demand and the battery's ability to supply current. Rpm increases are sometimes light, while other times can increase eneough to cause the car to start rolling forward in Drive. If your not paying attention, you can move straight into an intersection. The fix according to me would be to modify the firmware such that when the car speed is less than the speed which automatically engages the front seat seatbelt 'unattached' alarm, the switched voltage to the radiator fan can be half-split in order to drive the fan at half-speed and therefore draw a fraction of the current from the electrical system. This should relieve the alternator from signaling the engine rpm to compensate with a sudden rpm increase and also provide the engine with the necessary cooling while at idle speed. I would be happy if Toyota could do that for me.
I haven't gone through every posting on this entire blog, but in case someone hasn't already done so, here is a possible course of action if any automobile begins accelerating uncontrollably and other measures to bring it to a stop don't work. If the vehicle has an automatic transmission, immediately shift the transmission into neutral. After doing this, depress the brake pedal until the vehicle comes to a complete stop. Then, turn the ignition off, and apply the parking brake. If the vehicle is manual transmission, immediately engage the clutch, and place the transmission into the neutral position. After this, depress the brake pedal until the vehicle comes to a complete stop, then, turn the ignition off, and apply the parking brake. Obviously, if reasonably possible, you would attempt to bring the vehicle to a stop out of the way of other motorists. There is the chance of engine damage from a high rpm in neutral, but the safety to the occupants of the vehicle is obviously of more importance than any possible damage to the vehicle. Also, if possible, and you are able to keep your wits about you, it would probably be a good idea to engage the emergency lights as soon as reasonably possible as well.
I am Nguyen Ngoc Bich - engineer from Viet nam. I am interested in Toyota problems about unintended acceleration of Toyota vehicles and have red many articles about the problems on www.toyota.com, www.edmuns.com and other relevant websites. I decide to make a proposal to solve the problems for TOYOTA as follows: In my mind, root cause of all incidents of Toyota vehicles relevant with unintended acceleration problem is not in gas pedal but in electronic equipment of the throttle explored in high humidity condition. I did see any ECM in fact but see on the picture and believe this ECM is not too water proof in place of I/O outlet!!!
As I see on this picture, I/O of this ECM are likely to Ethernet cable outlets that their pins are too close to each other. In high humidity condition vapor can enter clearance between cable plug and outlets. When ambient temperature of the ECM reduce due weather or even due running conditioner of car, small drops of water may appear and connect pins and short circuit will occur that situation as professor’s Gilbert experiment will occur in real world! Because water is reason of unintended acceleration what will be evaporated quickly that very difficultly to recreate the situation in after incident investigations. To avoid this problem some follow measures will be done: 1. Make ECM more waterproof, especially where I/O of ECM and gas pedal. 2. A simple and safest method to come out from unintended acceleration situation is that Toyota can supply a switch to switch of voltage (< 5VDC) from ECM to throttle’s step motor when press on brake pedal and engine immediately come to idle condition. This method can make vehicle safe when wrong press both gas and brake pedals at same time too.
More explanation: As I have noted above that water is main culprit of suddenly unintended acceleration of cars with electronic throttle. But where water from? Body of ECM is made with stainless steel to protect ECM from RF. If ECM is not dressed with plastic cover when ambient temperature reduced by air conditioner humidity in air can create water drops on surface of ECM, these drops of water may join each other to create small flows and can enter outlets of ECM.
In my line of work condensation is a big problem and it could be the problem with your acceleration problem. I think your electrical box or computer is condensating on the inside and when you look for the problem the condensation has already dried up leaving no evidence. This is my million dollar shot, worth a try anyway.
When the various sensor inputs present to the computer their signals in a particular order and value, the software generates a full-throttle command.
That happens to my favorite theory also. While the software in the ECU was tested, thoroughly, I'm sure, it is almost impossible, except by using formal methods, to guarantee that the SW is bug free.
Here you have a piece of code that takes inputs from dozens or more sensors (including the throttle position sensor), processes those inputs, and uses the results to control fuel injector and spark timing, evaluate whether emissions are being met or not, and also, positions the throttle. It is impossible to test for all possible combination of inputs, over all ranges of values (both in range and out of range), over all possible input sequences, over all environmental conditions, ... (well, you get my drift). So, while the SW may very have been extensively tested, I'm sure many people could come up with a scenario that was not tested.
No amount of testing is going to guarantee that the SW is bug free. That merely reduces the probability that one or more bugs exist. So, given millions of vehicles driven thousands of miles a year, it should not surprise any of us that certain combination of inputs/values/sequences/timing, etc that may cause the UA problem will occur.
Not enough Shielding around Components relying on hundreds of sensors through out cars today,it is a location thing Being over head Electric wires ,relay of fire ambulance signal change Tech. This would be a guess but a fault not unheard of in todays driving world. The relay phone towers for billions of cell phones will boost signal what might just interrupt input unshielded components?Hydro Ontario charges all of its customers for Hydro leakage or stray voltage. This is to say it would not be showing in the car after the interrupt but at the time of assault. Toyota needs to look at environment causes.With high output high Def Digital satellites who knows what Shielding is required.
In the 90's I was installing cruise controls at a small shop here in Texas when one Jeep was going full throttle either setting still or going down the road. it was the interface or the cruise and the engine control module. with the cruise disconnected the vehicle wouldn't act up with the cruise wired-up it would go throttle-up. you can go so deep in the subject that you miss the simple answer. of all the cars that has throttle problems just tell me how many has cruise controls on board.
I believe the problem is in either the cruise control system suddenly kicking on and accelerating and not turning off after the brake is pressed.
If this is not the problem then the runaway accelerator could be forced to disconnect using the vehicles speed sensor to tell when the vehicle is moving and a momentary switch to turn off the throttle to prevent a runaway situation.
Barring this, going back to a cable controlled throttle would more than likely prevent the problem.
Nguyen, your posting/theory is my "favorite" so far. Very well-reasoned and well-presented. (And I'm not just saying that because I'm hoping you will give me part of your pending $1M prize!)
For those who surmise that EMI from 30KV power lines could be the cause - I have big doubts about that idea. We could possibly consider to do the math using 1/R^2 law, google/pick some reasonable numbers for current flow, and compare EMI/flux from 30KV power lines 200 feet from the ECM, and ignition spark/current just a few inches from the ECM. Then just for comparison, compare the flux from a 12V/low-current electric razor when at a distance of a few millimeters. May as well throw in a comparison of cellphone EMI flux too... Also we might want to consider that the ECM is presumably inside a faraday-cage and thus immune to RFI/EMI. The sensors feeding the ECM can't all be inside faraday cages, can they? maybe they could be affected.
One reason I like Nguyen's water idea is that it appears to be one of the simplest explanations that can explain the behavior. Bugs almost always seem mysterious before the root-cause is found. After they are discovered and the engineers say, "wow that seems so obvious in hindsight - how did we ever miss that?!"
A possible way to test the humidity-theory against current data is to correlate the humidity level for the time/location of each accident. Another approach could be to look for regional variations in the UA reports - are they happening mostly in humid areas of USA? (
"The sure fix for this problem is to go back to the cable and spring and get rid of the the electronic accelerator. Why beat your head against the wall looking for a cause, just fix the problem before any one else dies!!!!!!!!!!!!"
I agree with this 100%. Last week we bought a Kia Soul and while checking under the hood I noticed it still has the traditional cable going to the throttle body from the gas pedal. I became used to driving cars with the drive-by-wire system, a Honda Accord and a VW GTI, and I was really pleased with the Kia's acceleration from a full stop, it is much smoother (not faster, smoother) than cars with the electronic throttle. I'm sure it costs less as well.
I am a semiconductor engineer with a MSEE, I previously designed ICs at Motorola and I've worked on auto electronic ICs. I do not have a schematic for the circuits in this application and I expect Toyota would not let them out so my hypothesis is based on my professional conjecture. I presume the throttle is controlled by a stepper motor which is powered by a high current device like a bipolar NPN driver. It is possible that this device is contained in an IC which has multiple devices on one chip. It is possible that the layout of the chip was done in such a manner that a parasitic SCR was implemented without knowing about it. This was a too common mistake in the early days of IC design and the knowledge of how to avoid it may not exist at Toyota's supplier. As such, the device acts as a latch. It can be triggered by a negative voltage spike on the epi (base region) that pulls it below ground. The result would be the device saturates and the output is driven to the collector voltage and it stays there until the power is turned off. That would presumably create the full throttle condition which would not end until the power was shut off.
It is also possible that a layout mistake creates a parasitic MOS device that has the same effect. To analyze this I would need the schematic, device layout, and semiconductor process parameters. Unfortunately, I can see no way to get this information unless Toyota cooperates.
If you would like to proceed, my e-mail is williamtduffy@gmail.com
Thank Elias' comment to my posted theory. I think my proposed can cover reason of almost incidents of sudden unintended acceleration, but not all. Even incidents may occur with different reason: failure of gas pedal, failure of ECM due water or bad quality of semiconductor components,... but simplist and easist measure to help drivers quickly go out from dangerous situation is switch (may be as separate emergency knob or attached to brake pedal as I have proposed in #282) adding to power cable from ECM to the step motor of throttle. Day to day, more and more information about incidents caused by unintended acceleration but drivers can not go out from situation because in all car without this emergency knob!!! I hope TOYOTA corporation have red my comments!
This comment is posted to reinforce my proposal in comment #282 After reference of evaluation of Gilbirt demonstration by the Exponent I have finding that disign of the socket on gas pedal is not so good: • Orientation of the socket for plug in from top on the gas pedal is good path for water entering in case of damage of gasket. • Plastic shell surrounds housing of male portion to prevent direct moisture contact: but the plastic as a cup for keeping water inside the socket. • Wire are individually sealed with cylindrical polymeric grommets: water follows wires can fill up the grommets and under vibration generated by running engine and running car step by step may enter the socket through the poor elastic grommets and accumulated inside. Due these wicknesses water may enter into socket and cause short circuit between wires. The best design for this socket is it's orientation to be rotated down!
parasitic SCR was implemented without knowing about it. This was a too common mistake in the early days of IC design and the knowledge of how to avoid it may not exist at Toyota's supplier. As such, the device acts as a latch. It can be triggered by a negative voltage spike on the epi (base region) that pulls it below ground
If that happened, wouldn't the IC likely fail permanently because of high current? And, even if it didn't fail catastrophically, wouldn't there be some indication of this condition that would show up during a failure analysis - either a visual inspection or SEM?
And you're right - Toyota or it's IC vendor is not going to willingly give up the information you mention. IP rights, trade secrets, and aall that sort of stuff. However, there are companies out there that could reverse engineer such a chip,
However, there are companies out there that could reverse engineer such a chip, They would likely be threatened with the DMCA by Toyota if they tried that tactic to fix this problem!
They would likely be threatened with the DMCA by Toyota if they tried that tactic to fix this problem!
Probably right. What a POS piece of legislation that is. Topic for another thread...
Though, I wonder if that statute would apply to a defective product???
I was also figuring that maybe the feds would reverse engineer, or have a company they hire do it. That would leave the private company off the hook for any kind of IP violation. Then, let Toyota go after the feds.
If that happened, wouldn't the IC likely fail permanently because of high current? And, even if it didn't fail catastrophically, wouldn't there be some indication of this condition that would show up during a failure analysis - either a visual inspection or SEM?
It depends. If the current had more than one path to ground, it might burn out the metal links on the chip. But it is possible that the stepper motor has a high enough impedance to limit the current and preserve the chip. And indeed, if the chip burned out, an inspection would reveal the fault.
In any electronic control circuitry, thermal issues can be a problem. Thermal imaging of all silicon based components should be monitored in test vehicles until unintended acceleration occurs. In integrated circuits or discrete silicon based components, a phenomenon known as thermal runaway can occur in transistors (both discrete and those contained in I.C. components). This will throw the transistor into a "saturated" state (full on) even though it is intended to be off. This would change the state of logic in the system. When in thermal runaway, the component will remain saturated until the temperature lowers or power is removed from the component(s). The thermal source could be external (i.e. engine/exhaust related) or internal (i.e. intermittent circuit anomaly). The intermittent ones are very difficult to predict and or locate... thermal imaging of all components in this system will at least lower the number of possible locations and reduce time to discovery of the root cause. Just a possibility.
Software components (algorithms) that can affect the operation of the electronic throttle control: engine temperature algorithm affects throttle opening during engine warm up, cruise control resume algorithm affects the throttle opening to increase engine speed to produce the cruise set speed, traction control affects the throttle opening to reduce wheel spin. The input from the brake switch should also be included to affect the throttle opening in the same manner as the traction control. These algorithms affect pedal input independently to the drivers input to the throttle pedal position sensor. The powertrain control module uses the throttle position sensor on the throttle body to verify the throttle valve position and compares it to the input of the accelerator pedal position senor and engine speed (RPM) to produce the vehicle speed intended by the driver. The algorithms modify that input to the engine control module to produce the intended affect on engine speed. A bug or glitch in the algorithms could cause an unintended acceleration with an out put error to the electronic throttle body. An event data recorder could capture the anomaly that caused the event.
Vehicles most probably to exhibit a glitch: Toyota Camry SE or XLE with 3.5 V6 engine and smart key, and Lexus ES 350. Test 1 for software glitch: Cruise Control ON and engaged at set speed of 60 MPH or greater. Use the accelerator pedal to accelerate the vehicle speed to 15 MPH greater than the set speed. Test 2 for a software glitch Engine Coolant temperature is at or below 50º C, Cruise Control ON. Note engine idle RPM. place vehicle in Drive and accelerate to 1200 RPM above the noted engine idle RPM.
It is of interest that if the current data recording devises installed in the vehicles in question have not been interrogated as to the operation of the vehicle during the crash event.
.... now here is a cultural problem we should also look at. The fact that Toyota keeps insisting they do not have software / electronic problems, could be because these designer insist it is not their problem! "You can't ask a designer to bust his own design. That's like asking the elephant to watch the peanuts." -- quote from Doug Shade, Motorola -- If one goes back a few years, Mitsubishi had a similar problem in Japan, where they were (hiding) a lot of issues with their cars. Finally it came out. Embarrassment is a cultural problem in japan, where one does not easily admit to their faults. Mitsubishi was slammed hard by the consumer and suffered years of poor sales. Not sure if they ever recovered. Toyota seems to be going in the same direction. Everything seems to point to electronics, if its firmware or cruise control related, it does not seem to be hardware related. Anyway, it's time the Japanese culture admits making errors is actually a good thing....
It's too bad Toyota is so insular. I could solve their problem. I have years of experience as head of engineering for a major electronics firm. I am confident I could figure this out and I have many technical resources upon which to draw. For them, saving face is more important than discovering the cause of the problem. I know the laws of physics. Unfortunately, I'm no good when it comes to politics. The tragic result is that more Toyota customers will die before they fix this.
Many years ago a good friend of mine told me his wife was complaining to him that her Pontiac would accelerate suddenly for no apparent reason. It was an intermittent condition The car would run normally most of the time, then act up and accelerate for some reason. In some cases hitting the gas pedal returned the engine to normal speed-other times not. He had the car in the garage half a dozen times, but they couldn't find anything wrong. He asked me to look at the car since he knew I liked things mechanical, and enjoyed working on cars, so I spent a Saturday morning checking the car's throttle system. Couldn't find anything wrong with the gas pedal, linkage, cabling, etc.. Finally did a second check of the carburetor and then noted that the butterfly valve appeared to be loose- ie, the screws that secure the butterfly valve (plate) to the cross shaft hadn't been staked properly allowing the screws to loosen and permitting the valve to float up and down in the cross shaft Since the butterfly valve was loose and could move within the cross shaft this would allow the valve to sometimes jam in the carburetor housing- other times not. I duplicated this jamming repeatedly-if the valve slid down when the throttle was opened fully it would jam in the housing and not return to idle speed. I disassembled the carburetor, tightened and staked screws and the problem disappeared. Not sure if this condition can apply to the Toyota- but one never knows.
somewhere I had seen on tv or something that the signals from the ecm were like coax cable tv wire, and this is similar to strange interference we have in like tv problems. ingress, interference egress, signal leakage. good luck... tiasabaki@mchsi.com
I have been a certified EMC engineer for over 20 years. During the 80's I was the Principal Investigator for the US Navy for EMC problems aboard nuclear and diesel submarines. We had over 40 engineers and technicians making measurements aboard submarine all over the world. I have a patent on an EMC sensor and I currently consult (Technical Expert) to the Department of Commerce to determine if laboratories are making good EMC measurements.
That said, my approach would be as follows: put a vehicle that has exhibited the acceleration problem on a dynamometer, run the car at about 40 mph. Radiate the car in accordance with RS101 and RS103, of MIL-STD-461 (REQUIREMENTS FOR THE CONTROL OF ELECTROMAGNETIC INTERFERENCE CHARACTERISTICS OF SUBSYSTEMS AND EQUIPMENT). The military tests are more stringent than national/commercial tests (EN/TUV/CISPR, etc.).
RS101and RS103 are used to emulate RF magnetic and electric fields and their effects on electronics to be installed on military tanks, ships, planes, etc. They emulate emissions from car electronic components, game boys, security devices (driving past banks), etc. in asystematic way. Monitor the car for acceleration/reaction. If there are no reactions to the radiated fields (30Hz-40GHZ), then it is unlikely this is an EMC problem, and is mostly mechanical in nature (something jammed).
Once these tests are completed, hand held antennas can be used to determine which component is susceptible. Then the coupling method will need to be defined, and a resolution determined. This type of testing could cost $20,000 - $50,000, but provides the data needed to define the problem.
I've done this method with dozens of problems, and always defined the problem. If a chip is the culprit, it need not be redesigned. Good Shielding Practices can easily resolve the problem.
I know why this is happening and its very simple. I will wait for the rules of the contest before I let it be known. I am a proffesional driver and journyman mechanic with over 30 yrs experiance
As read on Wikipidia (feel free to distrust this information...): "A majority of motorcycles still use carburetors due to lower cost and throttle response problems with early injection setups, but as of 2005 many new models are now being introduced with fuel injection. Carburetors are still found in small engines and in older or specialized automobiles, such as those designed for stock car racing." And then there's this: "EFI could cause big problems. The sophisticated microprocessors used with EFI systems can do more than just meter fuel. The technology already exists to drive the car by itself.". Read it yourself "here". Just wondering if there's more problems up ahead, than the issues known to us today...
I have referred primary report of professor Gilbert’s about Toyota SUA and have saw a picture of APP sensor board located on the gas pedal. The printed board consist a 14-legs chip and some capacitors and resistors. Whole board is free for moisture contact because without any isolation layer. Legs of chip are very close to each other (<0,5mm) that water easily to cause short circuit. In case of moisture exist inside space for the board location, when cabin’s temperature is reduced with air conditioner, water may be condensate and create drop on the board, cause short circuit and sensors send wrong signal to ECM what may follow to a SUA. This situation may easily occur when gasket has damage that water or moisture can enter the board's location. To eliminate this kind of failure need to cover the board with isolation waterproof glue.
I would like to know, and I wonder if Toyota could tell me, did every run away car have cruise control on it? If one didn't have cruise, then that would rule out the cruise control causing the problem. I would be willing to bet the issue is the cruise control going nuts and causing the problem.
You're assuming there is only 1 root source of the problem. As you can see from other posts other brands and vehicles are intermittently affected by UA. There is no reason with all these components and systems trying to work together that you couldn't have multiple reasons - a few bad sensors, bad software, and a bad ECU chip in different individual vehicles, as examples.
And how do you know that the board is not coated with something? A number of the moisture resistant products out there are very thin and clear, and it is all but impossible to tell from a photo whether a PCB (printed circuit board) is so treated or not. Parylene(tm) coating that are put on by vapor deposition are almost impossible to see, even when you have the board in your hands. Humiseal(tm) is an acrylic that, if sprayed on thinly (as opposed to dipping the board) is difficult to tell if it's there or not from a photo.
He's looking for his 15 minutes, and a little cash Period. Bankrupt and a Car Guy [Corvettes] I hope CHIP and Toyota Criminally Charge him when it's proven he faked the whole event. That will Cut Down on Unintended Acceleration Events.
I used worked with electronic boards for long period that I believe this PCB is coated only printed board, soldered connections of componets with printed board are not coated. I just have found other error of the gas pedal when watching video "“An important update: 02.11.2010” from www.toyota.com, that a metal back lid of the gas pedal was assembled without gasket! That means moisture or water may enter the gas pedal easily.
Comments
You bring up an interesting point. Is there some sort of EMC (Electromagnetic Compatibility) standard for automobiles, similar to Mil-Std-461 for military gear?
There is one thing about noise trouble shooting: no cases are the same and satisfying low probability failure rates is extremely difficult. It is statistical problem where one has to accelerate the failures by stress testing so cause and effect improvements can be made
Exactly. Consider how many Toyota are out there, being driven day in and day out with no SUA incidents at all. What's that, hundreds of millions or billions of throttle operations in a year, yet only a handful of SUA incidents. That's a lot better than 6 nine's reliability.
I think that with nothing known about the throttle circuitry it is hard to favor anything because the options are infinite. But since I had 6 identical incidents in a different environment each time I would have to guess that in my case the trigger was internal to my car. In trouble shooting such a low probability event it has been my experience that you have to make a list of various causes and try and define a stress test to see if you could invoke the problem for each item on the list. It could be triggered by voltage noise due to some unknown condition, a ground loop resonance, as mentioned before software glitches --- noise doesn't have to be only voltage.
And your right the SUA incident would be very difficult to replicate because you would be working your way through a long list of suspects hoping you could stress the electronic throttle system in a way that would show the problem. You might be wrong on a list of 100 suspects and find it is the 100th idea tested.
I cut my teeth on the magnetic drum system on the Univac 1107 in the 60's. The system would run intermittently and then quite. A sudden rush of cold air on my neck when the air conditioner turned on that coincided with the failure gave the idea that it might be temperature related. Running a hair dryer on the drum intake showed that the clock pulse interrogating data would move over 4 data pulses with the hot air and move back when it cooled. An overnite fix with a drum temperature control solved the problem. It took about 6 months to nail. At the time we were running 7/24 with management covering shifts trying to push out the 20 first computers systems on the floor that weren't running because the drum didn't work. Tremendous pressure to say the least ---- but the most fun in trying to figure it all out.
For this problem one would have to have a lot of information for the Toyota electronic throttle which I would question the feasibility of getting. Does anyone know if it is possible to get this type of information?
I've read a lot of the posts and there are a lot of good ideas out there
http://www.autoblog.com/2010/03/08/report-abc-news-faked-at-least-one-part-of-ru- naway-toyota-repor/?icid=main|htmlws-main-n|dl9|link5|http%3A%2F%2Fwww.autoblog.- com%2F2010%2F03%2F08%2Freport-abc-news-faked-at-least-one-part-of-runaway-toyota- -repor%2F
Extrapolating on that is your idea because these sensitive electronic systems in Toyota's vehicles are like mini-antenna systems giving all that cellphone energy a place to call home. Look a second please at how many of these incidents are happening as the vehicle is in motion traveling horizontal to the cellphone energy, meaning their electronics systems are in effect cutting across artificially-created gravitational lines of flux.
Cellphone radiation may affect humans less since we don't walk 60 miles per hour everywhere we go. Right now we are still using Internet Protocol #4 (IPv4) but there are worldwide communications companies pressing hard to get people to accept Free or Low Cost international communications and phones which extra usage would then force IPv6 to be needed.
Why? => Because they are pushing the limits of IPv4 to force the world community to turn the dial up to Internet Protocol #6 (IPv6). IPv6 may very well cause more Toyota crashes from runaways. I heard a report this morning the same thing happened to a guy driving a Prius so this is no longer Toyota's problem alone. ** The wonderful Toyota Corporation cannot be held responsible for the communications industry hot-wiring the earth's atmosphere stuffed full of stray EMF radiation. **
My submission for the Toyota unintended acceleration problem by huanito
Mar 08, 2010 (1:53 pm)
March 8, 2010, at 1:30 pm, a Prius accelerated ...
http://news.yahoo.com/s/nm/20100309/us_nm/us_toyota_california_5
huanito said toyota good, the news said toyota no good.
CORRUPTED SOFTWARE OR MALICIOUS CODE INJECTED
The firmware has either become corrupted or contains a malicious code.
When the various sensor inputs present to the computer their signals in a particular order and value, the software generates a full-throttle command.
Causes:
1. This could be a result of an intentional or unintentional error in the computer's code as it was written but unlikely as Toyota has repeatedly tested their software.
2. This may also be caused by the firmware having become corrupted by any number of sources.
3. A malicious code has been unintentionally loaded into the computer's memory after the vehicle has left he factory. There are several possibilities for this scenario:
a. Routine service of the vehicle by an authorized dealer. When the technician connects his/her OBD test equipment to the vehicle's port, the vehicle's software is unknowingly re-written.
b. Annual inspections also require a technician to read the vehicles codes via the OBD port. Again, a malicious code could be written to the vehicle's computer.
The equipment used to read the vehicle's OBD port has changes or upgrades loaded onto it on a regular basis. The source of the malicious code:
a. Updates downloaded from the Toyota website
b. Updates sent by Toyota via CD-ROM
c. The manufacturer of the equipment's website
d. Updates sent by the manufacturer of the equipment via CD-ROM
Any of these update sources could had been hacked.
1) Satellite Radio/GPS for intended or unintended computer intrusion into drive by wire
2) for Cell or PDA interference with drive by wire functions
One thing that I think needs to be corrected after the claims in this latest news story, and others is, the term "unintended acceleration." I don't believe it is any longer the proper term to match the claims being made. If my vehicle accelerates "unintendedly" that term connotates an inconvenient situation that I can stop or control to some degree at any point if I wish too. I think the proper term to match the claims that are being made is "uncontrollable acceleration." But, if the descriptive term is upgraded to the adjective uncontrollable, then that raises the question as to why, with all these uncontrollable acceleration incidents, has there been no collisions, injuries, or even worse, fatalities.
If multiple vehicles, the number of which has been described in the claims, begin accelerating uncontrollably, it is a matter of simple odds, with the heavy road traffic, non-linear roads, and multiple stop signs and red lights, that the above consequences would result. Yet, none seemed to have happened so far. Of course, odds do allow for the possibility that a car could go careening down the road uncontrollably without colliding with something, or injuring or killing someone, but I believe astronomical would be the term to best describe them.
Again, I am not saying an uncontrollable acceleration problem absolutely does not exist in Toyota vehicles. But, there could be many reasons that a Toyota owner or driver, could or would make a claim like this. I'll let your imagination decide what those reasons could be. For myself, I am saying that many things about this situation simply do not add up, and I offered a possible theory as to what could be behind this issue. But ultimately, I will reserve my final judgement for when all the pieces of this puzzle fall into place, as I'm sure we will all be compelled to do.
The fix according to me would be to modify the firmware such that when the car speed is less than the speed which automatically engages the front seat seatbelt 'unattached' alarm, the switched voltage to the radiator fan can be half-split in order to drive the fan at half-speed and therefore draw a fraction of the current from the electrical system. This should relieve the alternator from signaling the engine rpm to compensate with a sudden rpm increase and also provide the engine with the necessary cooling while at idle speed.
I would be happy if Toyota could do that for me.
I am interested in Toyota problems about unintended acceleration of Toyota vehicles and have red many articles about the problems on www.toyota.com, www.edmuns.com and other relevant websites. I decide to make a proposal to solve the problems for TOYOTA as follows:
In my mind, root cause of all incidents of Toyota vehicles relevant with unintended acceleration problem is not in gas pedal but in electronic equipment of the throttle explored in high humidity condition. I did see any ECM in fact but see on the picture and believe this ECM is not too water proof in place of I/O outlet!!!
As I see on this picture, I/O of this ECM are likely to Ethernet cable outlets that their pins are too close to each other. In high humidity condition vapor can enter clearance between cable plug and outlets. When ambient temperature of the ECM reduce due weather or even due running conditioner of car, small drops of water may appear and connect pins and short circuit will occur that situation as professor’s Gilbert experiment will occur in real world! Because water is reason of unintended acceleration what will be evaporated quickly that very difficultly to recreate the situation in after incident investigations.
To avoid this problem some follow measures will be done:
1. Make ECM more waterproof, especially where I/O of ECM and gas pedal.
2. A simple and safest method to come out from unintended acceleration situation is that Toyota can supply a switch to switch of voltage (< 5VDC) from ECM to throttle’s step motor when press on brake pedal and engine immediately come to idle condition. This method can make vehicle safe when wrong press both gas and brake pedals at same time too.
More explanation: As I have noted above that water is main culprit of suddenly unintended acceleration of cars with electronic throttle. But where water from?
Body of ECM is made with stainless steel to protect ECM from RF. If ECM is not dressed with plastic cover when ambient temperature reduced by air conditioner humidity in air can create water drops on surface of ECM, these drops of water may join each other to create small flows and can enter outlets of ECM.
That happens to my favorite theory also. While the software in the ECU was tested, thoroughly, I'm sure, it is almost impossible, except by using formal methods, to guarantee that the SW is bug free.
Here you have a piece of code that takes inputs from dozens or more sensors (including the throttle position sensor), processes those inputs, and uses the results to control fuel injector and spark timing, evaluate whether emissions are being met or not, and also, positions the throttle. It is impossible to test for all possible combination of inputs, over all ranges of values (both in range and out of range), over all possible input sequences, over all environmental conditions, ... (well, you get my drift). So, while the SW may very have been extensively tested, I'm sure many people could come up with a scenario that was not tested.
No amount of testing is going to guarantee that the SW is bug free. That merely reduces the probability that one or more bugs exist. So, given millions of vehicles driven thousands of miles a year, it should not surprise any of us that certain combination of inputs/values/sequences/timing, etc that may cause the UA problem will occur.
If this is not the problem then the runaway accelerator could be forced to disconnect using the vehicles speed sensor to tell when the vehicle is moving and a momentary switch to turn off the throttle to prevent a runaway situation.
Barring this, going back to a cable controlled throttle would more than likely prevent the problem.
Very well-reasoned and well-presented.
(And I'm not just saying that because I'm hoping you will give me part of your pending $1M prize!)
For those who surmise that EMI from 30KV power lines could be the cause - I have big doubts about that idea.
We could possibly consider to do the math using 1/R^2 law, google/pick some reasonable numbers for current flow, and compare EMI/flux from 30KV power lines 200 feet from the ECM, and ignition spark/current just a few inches from the ECM.
Then just for comparison, compare the flux from a 12V/low-current electric razor when at a distance of a few millimeters. May as well throw in a comparison of cellphone EMI flux too... Also we might want to consider that the ECM is presumably inside a faraday-cage and thus immune to RFI/EMI. The sensors feeding the ECM can't all be inside faraday cages, can they? maybe they could be affected.
One reason I like Nguyen's water idea is that it appears to be one of the simplest explanations that can explain the behavior. Bugs almost always seem mysterious before the root-cause is found. After they are discovered and the engineers say, "wow that seems so obvious in hindsight - how did we ever miss that?!"
A possible way to test the humidity-theory against current data is to correlate the humidity level for the time/location of each accident. Another approach could be to look for regional variations in the UA reports - are they happening mostly in humid areas of USA? (
a bientot...
I agree with this 100%. Last week we bought a Kia Soul and while checking under the hood I noticed it still has the traditional cable going to the throttle body from the gas pedal. I became used to driving cars with the drive-by-wire system, a Honda Accord and a VW GTI, and I was really pleased with the Kia's acceleration from a full stop, it is much smoother (not faster, smoother) than cars with the electronic throttle. I'm sure it costs less as well.
It is also possible that a layout mistake creates a parasitic MOS device that has the same effect. To analyze this I would need the schematic, device layout, and semiconductor process parameters. Unfortunately, I can see no way to get this information unless Toyota cooperates.
If you would like to proceed, my e-mail is williamtduffy@gmail.com
After reference of evaluation of Gilbirt demonstration by the Exponent I have finding that disign of the socket on gas pedal is not so good:
• Orientation of the socket for plug in from top on the gas pedal is good path for water entering in case of damage of gasket.
• Plastic shell surrounds housing of male portion to prevent direct moisture contact: but the plastic as a cup for keeping water inside the socket.
• Wire are individually sealed with cylindrical polymeric grommets: water follows wires can fill up the grommets and under vibration generated by running engine and running car step by step may enter the socket through the poor elastic grommets and accumulated inside.
Due these wicknesses water may enter into socket and cause short circuit between wires.
The best design for this socket is it's orientation to be rotated down!
If that happened, wouldn't the IC likely fail permanently because of high current? And, even if it didn't fail catastrophically, wouldn't there be some indication of this condition that would show up during a failure analysis - either a visual inspection or SEM?
And you're right - Toyota or it's IC vendor is not going to willingly give up the information you mention. IP rights, trade secrets, and aall that sort of stuff. However, there are companies out there that could reverse engineer such a chip,
They would likely be threatened with the DMCA by Toyota if they tried that tactic to fix this problem!
Probably right. What a POS piece of legislation that is. Topic for another thread...
Though, I wonder if that statute would apply to a defective product???
I was also figuring that maybe the feds would reverse engineer, or have a company they hire do it. That would leave the private company off the hook for any kind of IP violation. Then, let Toyota go after the feds.
It depends. If the current had more than one path to ground, it might burn out the metal links on the chip. But it is possible that the stepper motor has a high enough impedance to limit the current and preserve the chip. And indeed, if the chip burned out, an inspection would reveal the fault.
Vehicles most probably to exhibit a glitch: Toyota Camry SE or XLE with 3.5 V6 engine and smart key, and Lexus ES 350.
Test 1 for software glitch: Cruise Control ON and engaged at set speed of 60 MPH or greater. Use the accelerator pedal to accelerate the vehicle speed to 15 MPH greater than the set speed. Test 2 for a software glitch Engine Coolant temperature is at or below 50º C, Cruise Control ON. Note engine idle RPM. place vehicle in Drive and accelerate to 1200 RPM above the noted engine idle RPM.
It is of interest that if the current data recording devises installed in the vehicles in question have not been interrogated as to the operation of the vehicle during the crash event.
If one goes back a few years, Mitsubishi had a similar problem in Japan, where they were (hiding) a lot of issues with their cars. Finally it came out. Embarrassment is a cultural problem in japan, where one does not easily admit to their faults. Mitsubishi was slammed hard by the consumer and suffered years of poor sales. Not sure if they ever recovered.
Toyota seems to be going in the same direction. Everything seems to point to electronics, if its firmware or cruise control related, it does not seem to be hardware related.
Anyway, it's time the Japanese culture admits making errors is actually a good thing....
That said, my approach would be as follows: put a vehicle that has exhibited the acceleration problem on a dynamometer, run the car at about 40 mph. Radiate the car in accordance with RS101 and RS103, of MIL-STD-461 (REQUIREMENTS FOR THE CONTROL OF ELECTROMAGNETIC INTERFERENCE CHARACTERISTICS OF SUBSYSTEMS AND EQUIPMENT). The military tests are more stringent than national/commercial tests (EN/TUV/CISPR, etc.).
RS101and RS103 are used to emulate RF magnetic and electric fields and their effects on electronics to be installed on military tanks, ships, planes, etc. They emulate emissions from car electronic components, game boys, security devices (driving past banks), etc. in asystematic way. Monitor the car for acceleration/reaction. If there are no reactions to the radiated fields (30Hz-40GHZ), then it is unlikely this is an EMC problem, and is mostly mechanical in nature (something jammed).
There are laboratories across the nation that do this type of testing. A list is available here: http://ts.nist.gov/Standards/scopes/ect.htm
Once these tests are completed, hand held antennas can be used to determine which component is susceptible. Then the coupling method will need to be defined, and a resolution determined. This type of testing could cost $20,000 - $50,000, but provides the data needed to define the problem.
I've done this method with dozens of problems, and always defined the problem. If a chip is the culprit, it need not be redesigned. Good Shielding Practices can easily resolve the problem.
"A majority of motorcycles still use carburetors due to lower cost and throttle response problems with early injection setups, but as of 2005 many new models are now being introduced with fuel injection. Carburetors are still found in small engines and in older or specialized automobiles, such as those designed for stock car racing."
And then there's this: "EFI could cause big problems. The sophisticated microprocessors used with EFI systems can do more than just meter fuel. The technology already exists to drive the car by itself.". Read it yourself "here".
Just wondering if there's more problems up ahead, than the issues known to us today...
Step 1 - Did all of the cars have cruise control.
If not, what did all of the cars have in common?
Just my 2 bits.
Mike G
Is there an automotive equivalent to Mil-Std-461? Something put out by SAE, perhaps?
I am sure this is in cars today, but not offered as a service. They would have been doing testing.
Has anyone looked into the idea that hackers may get into this system to effect individual cars.
I know this is a bit "out there" but technically, it could be possible.
Bankrupt and a Car Guy [Corvettes]
I hope CHIP and Toyota Criminally Charge him when it's proven he faked the whole event. That will Cut Down on Unintended Acceleration Events.