[Brek]

So on my most recent trip and with a little cooler conditions (55f-20f) I started to experience a really fun issue. It usually happens in the morning when the car is cold but basically, I will start driving and somewhere down the road (within the first 20 minutes of driving) I will shift gears (any gear, up or down) and my gas pedal will quite working and my car will try to idle at about 2k Rpm. I then have to coast or if I am stopped clutch bump my car to the side of the road. I have to turn off the car and let it sit for about a minute, I will try to start it once and it won't go, give it a second shot and it works like a charm as if nothing happened...

I would like to try and figure this one out as it can be a little sketchy as I am sure you all can imagine...

Ive done a little research and it sounds like it could be the throttle control harness, its just odd that I have only experienced this problem in slightly colder conditions...

Any help or ideas with this would be great!

 

[kybishop]

Do you get a CEL (check engine light) or anything?

Had issues on our 2006 Outback. Checked connections and all looked ok. Ours ended up being the throttle position sensor that is built in to the gas pedal assembly.

This thread should help you trouble shoot and figure it out; http://www.subaruoutback.org/forums...4906-code-p2138-throttle-position-sensor.html

 

[gbhrps]

Brek,

Research shows that the "likely" culprit is the Throttle Position Sensor on your accelerator pedal. Up under the dash of the pedal you'll find a wide 6 wire connector. Internally it has an orange rubber seal that may have let in moisture to corrode some of the pins. Another occurrence shows that the connector may not be seated properly to make a good contact and needs to be removed and then reseated properly, or a loose wire to one of the pins.

Likewise, the other end of that wiring harness goes to the throttle body up on the engine, and its possible the same things may be going on there.

If neither connectors are the solution, you may need a new Throttle Position Sensor, or the engine computer may be at fault.

If it were my car, I'd deal with those connectors (loose or corroded) and if that doesn't solve the issue, I'd get a good independent mechanic to check the system out before I dropped any money into parts.

Cold shrinks things, heat expands them. So the temperature differences are making/breaking some internal connections somewhere in that system. Good Luck!

 

[CNY_Dave]

My gas pedal on my '05 forester does what's described.
Did it for several weeks before throwing the code and putting it into limp mode (a managed 2500 rpm or so, just enough to creep along).
A shut-off and re-start allowed it to be driven.

I use the cruise control a lot and it has only thrown the code while using cruise.

I think the contacts at the idle position on the pedal have fretting wear from miles of vibration with the pedal in the idle position.

 

[plain OM]

@CNY_Dave

Is the 05 Forester drive-by-wire? (I suspect it is but want to confirm.) If so, then the pedal main and sub sensor voltages can be read/logged e.g. using Romraider. It's one way to verify if the pedal is sending bad data to the ECM, triggering the CEL. There's a recent thread where the cruise self-cancels after about 17 minutes. The pedal is suspect but waiting for follow-up. http://www.subaruoutback.org/forums...51969-cruise-control-cut-out.html#post5154826

 

[CNY_Dave]

Yes, drive by mother-may-I wire.

 

[Schaeff58]

To all:

This just happened to me yesterday. I have a 2006 outback, 3.0r L L Bean edition, with about 212k miles. Ive had it about a month. This is the first major issue. Got about 2 miles away from the house and I had no throttle control. I pulled over and turned around and headed home at about 2k rpm crawl. Was glad I made it up the small hill. It threw a code, as CEL was on, have not cked it yet, cruise light was flashing too. Got back to the house and parked. Moved my old faithful '97 outback and then then tried to put the 2006 in the other parking spot the '97 had. The 2006 did not start the first time, but did the second time and I had throttle control back. I drove the '97 to work. Have not looked at 2006 yet. I have read many, many posts here and feel certain it is accelerator pedal. I am also going to ck the throttle body to see if its dirty, throttle body sensor and if I can see it the wiring harness under the intake manifold and pull the code. It seems that most cases it has been the accelerator pedal. I will report back.

 

[Schaeff58]

Ok, reporting back!
Read the code, it was the P2138. I first cked the plug and it was in solid. Looked over the wires to the fitting and it all looked ok. I pulled the pedal and took cover off sensor. You can see the wear that it has... I used some electronic cleaner and sprayed all the parts, plug, fingers etc... Put it all back together and drove it... it worked but I ordered a new pedal just the same. I wont drive it on the road until I have it replaced. I ordered part # 36010AG021 Off ebay. https://www.ebay.com/itm/2005-2006-...281436&hash=item1a2d8795e0:g:-e0AAOSwZtlZ5bI8 112 bucks! Says oem... cheapest I found it.
I also looked at the throttle body, it was ok, not eat off it clean, but not bad. Looked wiring harness over and felt it, no damage I could tell, so I really think its the pedal

Thanks to all for posting their experiences and fix.

This is a great place for help!

 

[plain OM]

. . .You can see the wear that it has... I used some electronic cleaner and sprayed all the parts, plug, fingers etc...

The close-up of the resistance traces is interesting. The two middle ones look unusually rough, whereas the two outer ones appear to be smoother overall (except at the ends, where the spring contacts probably don't have to reach). I'm fairly sure those surfaces should be smooth to the eye (and camera). I wonder what the matingspring contact surfaces of the sliding elements were like. (Perhaps notably worn?)

 

[johnre]

The close-up of the resistance traces is interesting. The two middle ones look unusually rough, whereas the two outer ones appear to be smoother overall (except at the ends, where the spring contacts probably don't have to reach). I'm fairly sure those surfaces should be smooth to the eye (and camera). I wonder what the mating spring contact surfaces of the sliding elements were like. (Perhaps notably worn?)

Interesting indeed. Probably a thick film process, and there is clearly some redundancy designed into this circuit. And the two traces that say "rough" on your thumbnail illustration have pad contacts on one end of the film, whereas the two that say "smooth" are on both ends of the film. All the pads have marks, indicating a pressure contact.

I agree, it would be interesting to see the arrangement of the sliding contacts - to see how the entire circuit is wired.

Was the compartment this circuit was inside sealed or vented? That would give some clue about failure mode - moisture from internal condensation will be an issue if it's vented.

 

[Schaeff58]

The close-up of the resistance traces is interesting. The two middle ones look unusually rough, whereas the two outer ones appear to be smoother overall (except at the ends, where the spring contacts probably don't have to reach). I'm fairly sure those surfaces should be smooth to the eye (and camera). I wonder what the matingspring contact surfaces of the sliding elements were like. (Perhaps notably worn?)

I didnt think to snap a pic of the part that makes contact on the worn surface. I will do that when I take it back out after I replace the part. I did look at them, they were not the copper color they were silver, but I did not look at the wear.

 

[Schaeff58]

Interesting indeed. Probably a thick film process, and there is clearly some redundancy designed into this circuit. And the two traces that say "rough" on your thumbnail illustration have pad contacts on one end of the film, whereas the two that say "smooth" are on both ends of the film. All the pads have marks, indicating a pressure contact.

I agree, it would be interesting to see the arrangement of the sliding contacts - to see how the entire circuit is wired.

Was the compartment this circuit was inside sealed or vented? That would give some clue about failure mode - moisture from internal condensation will be an issue if it's vented.

To answer your question Johnre I dont know if it vented or not. The sensor seemed to be sealed... it had a nice intact o-ring type seal along the outside edge of the unit. Nothing in the way of condensation or dirt in there... There was a black coloring of the shaft in the center that protrudes about 1/4" that the slider pivots on...I wiped it off I thought it was the layer of material worn off the contacts. (but have no clue).

 

[johnre]

To answer your question Johnre I dont know if it vented or not. The sensor seemed to be sealed... it had a nice intact o-ring type seal along the outside edge of the unit. Nothing in the way of condensation or dirt in there... There was a black coloring of the shaft in the center that protrudes about 1/4" that the slider pivots on...I wiped it off I thought it was the layer of material worn off the contacts. (but have no clue).

You've essentially answered my question when you noted that it was O-ringed. That pretty much eliminates moisture or chemical contamination as the failure mechanism, and instead makes my suspicion fall on the one remaining possibility - plain old wear and tear on the wiper - as everyone else here already suspected.

The stuff on the shaft might well be exactly what you think it is. Very small particulates that are scrubbed or worn off don't just randomly fly off every direction and deposit wherever; they tend to bounce around a while and eventually find something to bind to, either chemically of through electrical (ionic) attraction. For whatever reason, they must like the rotor shaft.

Subaru probably knows about this wear out mechanism, as they put redundancy into their design. It was probably tested on a cycling machine for the equivalent of several hundred thousand miles in what's called accelerated testing (honest - that's what it's called; no pun intended!). However, this can't always mimic the reality of conditions we put our cars through, and it does appear from these photos, your other observations, and owner anecdotes that the design is marginal here.

Thanks for the disassembly and photos - nice of you to do this for us.

 

[plain OM]

Interesting indeed. Probably a thick film process, and there is clearly some redundancy designed into this circuit. And the two traces that say "rough" on your thumbnail illustration have pad contacts on one end of the film, whereas the two that say "smooth" are on both ends of the film. All the pads have marks, indicating a pressure contact.

I agree, it would be interesting to see the arrangement of the sliding contacts - to see how the entire circuit is wired.

Was the compartment this circuit was inside sealed or vented? That would give some clue about failure mode - moisture from internal condensation will be an issue if it's vented.

There's some more photos in http://www.subaruoutback.org/forums...8-throttle-position-sensor-4.html#post4879322, and perhaps elsewhere in that thread and others relating to the code.

 

[johnre]

There's some more photos in http://www.subaruoutback.org/forums...8-throttle-position-sensor-4.html#post4879322, and perhaps elsewhere in that thread and others relating to the code.

Thanks for the link; one post in that thread here:

http://www.subaruoutback.org/forums...138-throttle-position-sensor.html#post4537497

has very clear pictures of the elusive wiper construction.

I note there are gold plated pressure contacts on the pins that contact the six pads on the top of the film circuit - that's good - but these pins never move on the pads except when it's assembled or disassembled. The wiper fingers, on the other hand - three on each beam - appear to be silver plated, but can't be sure. One thing for certain is that they aren't gold plated, which isn't good - it takes a lot more pressure to assure good wiper contact without gold, and those wiper fingers are clearly gouging out troughs in the films (the three lines on each film). That establishes a known wear-out mechanism of the film and/or wipers related not to time, but the number of times it scrapes the wipers over the film.

I promise a more thorough analysis when (not if) mine fails!

 

[Schaeff58]

You've essentially answered my question when you noted that it was O-ringed. That pretty much eliminates moisture or chemical contamination as the failure mechanism, and instead makes my suspicion fall on the one remaining possibility - plain old wear and tear on the wiper - as everyone else here already suspected.

The stuff on the shaft might well be exactly what you think it is. Very small particulates that are scrubbed or worn off don't just randomly fly off every direction and deposit wherever; they tend to bounce around a while and eventually find something to bind to, either chemically of through electrical (ionic) attraction. For whatever reason, they must like the rotor shaft.

Subaru probably knows about this wear out mechanism, as they put redundancy into their design. It was probably tested on a cycling machine for the equivalent of several hundred thousand miles in what's called accelerated testing (honest - that's what it's called; no pun intended!). However, this can't always mimic the reality of conditions we put our cars through, and it does appear from these photos, your other observations, and owner anecdotes that the design is marginal here.

Thanks for the disassembly and photos - nice of you to do this for us.

Happy to contribute!

My car was well maintained from the folder I got from the previous owner, and the accelerator pedal was not documented as being replaced. I have not put a thousand miles on this yet, so the 212K already on this car surely approaches the 'accelerated testing' literally.One question I had was answered about redundancy in this part for the wear issue. To me it looked like at parts of the very severely worn places were completely done, gone, and minimal contact was being made. And I think you mentioned it about gold vs silver contacts. Mine were a silver color, and I plan to take the sensor apart and add those photos as well of the wiper.
I am happy it is an easy, but not so cheap, fix. But if I get another 212k out of replacement pedal, the cost per mile is nothing! :smile2:

 

[johnre]

I am happy it is an easy, but not so cheap, fix. But if I get another 212k out of replacement pedal, the cost per mile is nothing!

I agree, amortized over 212k miles, it's not much. But IMHO such a "mission critical" item as this should not wear out over any reasonable projected life span of the vehicle, especially considering (1) it directly replaced something that did last that long (a mechanical linkage), and (2) there are alternative potentiometer designs available for this function that do last this long, albeit at a higher cost.

 

[CNY_Dave]

This post might be redundant- there are 2 resistance circuits, the computer constantly compares the 2, a mis-compare triggers one of the pedal codes.

 

[Schaeff58]

I agree, amortized over 212k miles, it's not much. But IMHO such a "mission critical" item as this should not wear out over any reasonable projected life span of the vehicle, especially considering (1) it directly replaced something that did last that long (a mechanical linkage), and (2) there are alternative potentiometer designs available for this function that do last this long, albeit at a higher cost.

I wholeheartedly agree with you. And you are right the cost is n ot the point. Not trying to say this failure is ok in any form, so point taken.

I got my replacement pedal yesterday and swapped it out. Took a whole 10 mins. I test drove it, and then to work this am. Man it feels way different, better response.

I said I would include a picture of the wiper. I took several, but it is hard to get a good one, this one is the best, and the rest all look about the same. Sorta like a three toed cats claw on each arm. That sure seems like it would not last as long as it did. Something has to wear when its set up like that... I would have thought it would be on some sort of an outside corner, or a flat slider and not this rake type configuration.

 

[johnre]

This post might be redundant- there are 2 resistance circuits, the computer constantly compares the 2, a mis-compare triggers one of the pedal codes.

A redundant post about redundancy.:smile2:

I'm not going to let Subaru off easy on this one. Redundancy only works well when (1) the probability of failure of each contributing part is low, and (2) there is not much statistical correlation of failures; i.e. there is no mechanism that synchronizes failures to a predictable time or wear level.

If you have neither situation, all you can do is exactly what Subaru has implemented - throw the well-known P2138 code up, which may allow you to operate it long enough to get to a repair facility. It's sort of like putting a wear indicator on brake pads - you hear the first one squeal, and that means it's past time for brake service. However, it best be done promptly, as the others are likely to be in a similar state, at least for that axle.

Looking over all of the pictures provided (thanks - you all know who you are!) leads me to believe that those of us who get a P2138 code and then see it go away probably experienced the initial wiper "claw through" of only one small spot on one circuit. The spot was so small that the wiper didn't hit it again for a while, hence it "got better". But this meant the films were almost worn through everywhere, and time was getting short. At some point, it's down to only one circuit working, and the ECU is carrying a P2138 code all the time. Then that circuit fails, and we know what happens then.

Subaru engineers understood all this, and their only response was to make the circuit redundant in the fashion I described.