Help Diagnosing Unintended/Uncommanded Acceleration Issue

Yes, those can be reasonable solutions. If you have time, you can put it all together into a smallish box, complete with power supply. You don't need much power, they run a long time on batteries

 

Correct, they are redundant electrical and mechanical brakes. In any modern car, total brake failure outside of severe accident damage or extreme maintenance neglect is not credible.

 

What if there is a failure inside the accelerator pedal assembly?



It wouldn't be any different than a broken return spring on a carburetor. According to a neighbor of mine, who runs a small fleet of Kenworth highway tractors, when the pedal sensor fails, what *usually* happens is the engine remains at idle, and you're stranded on the side of the road

He did have one truck that would sometimes accelerate on its own. Flooring the accelerator in a Kenworth gives you *plenty* of time to think about what is going on, and how to fix it.

There is only so much redundancy one can build into a system. I've done software engineering using secure, embedded RTOS's, on fly-by-wire military aircraft. There is a point of diminishing returns going past 3 individual sensors for the same control input

I do recall being copied to one notice that was widely circulated within the industry. A NW Airlines Airbus A320 would sometimes experience uncommanded roll, which was easily corrected but frightening for all aboard

The sidestick was found to have a gunky residue inside, causing the contacts in the sidestick to sometimes stick. Further analysis of the residue indicated coffee and sugar. S*** happens, and in that case a replacement of the sidestick cured the problem

Prius owners with HID headlights have had corrosion-related failures of the rear axle height sensor assembly, and you don't always get a code. What happens is that the headlights don't point properly, usually way too low

The point to all of this, is that even if I were to hook up my $50,000 MSO to record constantly, it would *not* be recorded as a "glitch." To the car, it will seem as if you really *did* floor the gas pedal. Much like that Airbus, the Quick Access Recorder verified receiving an input from the PIC sidestick, but the pilot never touched the sidestick

I suppose if such an event would happen to me, I'd just replace the pedal sensor assembly

 

In normal use, the pedal is bypassed from the hydraulic system. There is a "stroke simulator" valve



Hold on a second. BooBoo walked across the keyboard here, I had to retype everything. Bad cat, bad cat

So, if the system senses a failure, the braking system will enter a failure mode. This is one such failure mode, the solenoid valves are switched in such a way to provide direct pedal to the front brakes only



One of our forum members, Patrick Wong, his fathers Prius had a brake failure that did not correctly enter the failure mode. This was a mechanical issue, I can't recall the exact part replaced but I'm sure Patrick will know

 

I had one NHW11 Prius that came in for refurbishment with a problem:

Notice that in this 2001-03 accelerator, the nylon cam follower was actually jamming in the "U" channel slide. It would not freeze but I could feel a mechanical 'spot.'

This out-of-focus photo more clearly shows the problem:



I really didn't see it until I was looking at the resistance curves:


What I've never seen is the internal structure of the NHW20 accelerator assembly. I don't know if a similar 'cam-follower' mechanism is used or something that does not involve additional moving parts.

Bob Wilson

 

BTW, nice job in changing the Title to this thread.

Please let me know how you did it! I was hoping to change the title to an earlier thread of mine, but I couldn't accomplish the feat. Is it a moderator-only option?

 

Another thought on this:

The accelerator pedal does not control the throttle position - the HV ECU controls the throttle position. The accelerator pedal position is just an input to the HV ECU. It is your vote on how fast the car should move. The HV ECU is busy with other powertrain functions even when your foot is off the accelerator pedal. The HV ECU may decide to run the ICE to generate charging current using MG1 - that has nothing to do with the accelerator position.

The HV ECU might command a certain throttle opening to run the ICE to charge the battery and then notice that you have reduced pressure on the brake pedal. It should then simulate "creep" of the transaxle and adjust the throttle position to compensate. What if the throttle plate sticks momentarily? Theoretically, it should set a code when the throttle position doesn't match the commanded position. What if the error isn't long enough in duration or doesn't occur often enough to set the code?

 

I'm thankful the moderators did as well as the user staying around to help us work on problem diagnosis. We've seen too many "hit and run" posters about this subject who lack the patience and insight to work with us on the problem.

It is a hard problem combining both elements safety as well as being intermittent. But it looks like folks are bringing their "A" game and treating it seriously.

Bob Wilson

 

You wouldn't have to touch the gas pedal. If there is an internal problem with the sensor, as Bob pointed out with the bad Gen I pedal assembly, just normal road vibration could cause a change in resistance values.

A similar issue is seen with valves used in chemical plants. The fancy term is "stiction" which means Sticky Friction. With the valve, nothing happens with the original control input. The system commands more OP to the valve. Nothing happens. By this time, a "command disagree" alarm is generated, as the valve isn't responding. Suddenly the valve positioner makes a LARGE adjustment

With the sidestick issue on the Airbus, neither the PIC or the co-pilot touched their sidesticks when the uncommanded roll happened. They had already had a positive rate of climb, and were on autopilot

With the Airbus, a brief bump of the sidestick will not disconnect the autoflight system. IIRC a 2 second control input *will* disengage the autoflight, and the master warning wailer goes off to tell the PIC the autoflight is dumped

When both pilots - highly trained and ex military - swear up and down they never touched the sidestick, and the CVR has verbal confirmation to the effect of "what the f*** just happened?" then one must assume something happened somewhere in the control system

As a result of that uncommanded roll issue, every Airbus operator received notification. There were actually two separate issues logged, at first it was thought to be a software glitch

CHI95IA138

CHI95IA342

I would not be surprised if a similar issue isn't happening to your car. There could also be a wiring issue between the pedal and the ECM. Eg something that has chaffed the harness

 

Bob

A very plausible theory. I would not be surprised if a similar structure is used for the Gen II. We need somebody to confirm this. Eg a similar throttle body is used for the Gen II

 

Another very plausible theory

I've noticed that operation in bitterly cold climates, especially short trip operation, puts a lot of trash into the intake. There is a lot of oily vapor introduced into the intake due to the PCV system

One thing I've noticed about Toyota VVT motors, they don't appear to use a separate EGR valve system. The software commands the VVT to work similar to a separate EGR system

In other vehicles, after only a year or two, I've had to clean TB blades and IAC passages, due to heavy deposits as a result of the PCV and EGR.

Forum member Hobbit dissected a Gen I TB that appeared sticky to him

http://www.techno-fandom.org/~hobbit/cars/tb/

I personally would like to vote Hobbit as MVP on the PriusChat team. He really goes the extra mile to dig into things. Plus, he isn't Evil like I am

Seriously, due to how the VVT works, some forum members assertion of the TB being "washed" by gasoline residue is busted. Now that I really think about it, that "puddle" we see in the intake manifold is the result of residue from valve-overlap "internal" EGR function, and PCV vapor

At the very least, that TB in question should be examined for deposits. Good catch

 

Bob

You get me a bottle of gin, I'll absolutely give this my "A" game effort

jay

ps none of that cheap stuff, I prefer Bombay or Tanqueray

 

Those reports conflict somewhat. The same plane on the same day has the same problem (according to the day and tail number). Yet one report states that in the first case it probably "pilot unfamiliarity" and the second finds the potentiometer problem. What's your take Jayman?

 

The problem with checking signal integrity/bounds is that the control system can't read minds. Eg: red light, foot on brake. Wheel speed sensors and shaft speed sensors register 0 mph, which is what you'd expect.

Suddenly, the accelerator pedal sensor indicates full travel, eg floored gas pedal. Is that a genuine operator input, or is it a "bad" signal? Would a typical driver have a foot on the brake pedal, *and* a foot on the gas pedal, at the same time?

Well, if they intend to do a brake stand burnout, probably. But how would the computer know that?

With critical process safety systems, where failure or misinterpreted control input (Eg, momentary loss of signal input, EMI, etc) can reasonably be expected to cause a serious situation (Explosion, loss of life, city consummed by fireball, etc), there is "voting." Typically, 3 separate inputs are used, and the software "votes" the inputs.

So if 1 out of 3 go bad, the bad one is voted out of the loop. Where it gets tricky is if 2 out of 3 go bad, sometimes you get very unpredictible results from the software

I agree that only a complete "dump" can uncover such situations. The hard part in that scenario is trying to convince Toyota that when the dump claims you floored the gas pedal, you actually did *not* have your foot on the gas pedal

The cruise control is easily outvoted by manual pedal input, is cancelled by brake pedal input (Assuming the brake switches are correctly adjusted and working!), and will only slowly accelerate with a Resume command

The accelerator pedal assembly could have an internal problem, as Mr Wilson documented on that Gen I pedal assembly. There could be a wiring problem between the pedal and the ECU (Chaffed wires, momentary short or open, etc)

I'm convinced that as the fleet ages, we will see more of these issues pop up. Entire fleets of aircraft were retired due to bad wiring, with the cost of rewiring exceeding the value of the aircraft, and the safety related issue of wiring demanding they be grounded

As wiring ages, insulation can break down. Corrosion can start in wiring bundles and especially in connectors. Ground points can corrode and introduce intermittent floating ground problems. An intermittant floating ground due to a loose battery shunt almost crashed a Boeing 767 over the Atlantic, and did introduce significant system failures on landing at Logan

NYC96IA116

The real question is: are the manufactures and *especially* dealerships understanding this issue, and ready to deal with it? Intermittant problems are the most difficult to resolve