Bad Throttle Position Sensor???

[MaStaMooN]

What is are the correct percentages of the TPS, when the car is off? I noticed mine wasn't near 0% nor near 100% last time I had Torque running and checking on things. My last car the percentages were like 5% and 95% with anything above or below that meaning a fault. I think mine showed like 15% and 75% or something possibly.

[scottmcphee]

Throttle sits off zero when no pedal is applied (spring holds it open preventing a complete choke). I think mine is around 10%, when pedal is 0%.

 

By the time pedal is pressed 60% the throttle has made it to WOT (100%) and all the "extra press" of the pedal beyond that shows throttle holding at WOT.

[Max Capacity]

That sensor would not fit 2006 Legacy 2.5i

 

Sorry, not much around here for a 2.5i.

[Phillip J. Fry]

Followup please?

 

I'm wrestling a p2138 pedal sensor code, that does not resurface if the MAF is unplugged.

 

I'm about to try plugging MAF back in and unplug front O2, to see if that also keeps p2138 away (dead pedal and limp home idle).

 

oh. sorry.

 

i only had codes for misfires when my engine cut out. dealership doesnt know what the problem was. smart service shoreline wasted my money and came back with possible injectors but werent certain.

 

i paid for them to put in an actual subaru 02 sensor. why they didnt in the first place i dont know. they used up all my gas and gave it back to me on e. so really all the bad gas should be gone.

 

bad gas is the only thing i can think of. ive since cleared the ecu and put in 2 bottles of berryman b12 and a costco tank of gas. the car is driving better in some ways but im getting after fires. the surging and bucking with rpm fluctuations are gone so far.

 

im still on that same tank of B12. need more driving time.

[Phillip J. Fry]

i just went to the garage and did a throttle target test. i can hear a strange noise when its opening and i see the throttle go from 50 to 70 to 98. from 50 to 98 its like 2 or 3 steps. not even close to being smooth. i can hear some noise at this location. before and after its quiet. then jumps to this...56-70-83-98...those are the steps and are not smooth.

 

the other test was throttle pos...0-50 was great then it jumps like this...64-75-84-98...its not smooth here at all. only in those crazy steps.

 

in both cases under 50% is dead on. maybe the injectors are weirdo from possible fuel that was equivalent to rain water and/or the throttle plate (i dont know what to call it) area is funked up royally.

 

the test was done on a completely cold car. engine OFF. ~64f in the garage. fuel is at 64 and so is coolant. blue light was on so this was done in Sport mode. not sport#. sport# increases throttle sensitivity or something or another. you all know what im getting at.

Edited April 23, 2016 by Phillip J. Fry

[Phillip J. Fry]

Throttle sits off zero when no pedal is applied (spring holds it open preventing a complete choke). I think mine is around 10%, when pedal is 0%.

 

By the time pedal is pressed 60% the throttle has made it to WOT (100%) and all the "extra press" of the pedal beyond that shows throttle holding at WOT.

 

yes. mine was 10-11%

 

yeah. mine went to 98 then near the floor was 100%.

[scottmcphee]

I think this is normal throttle opening program to go in big steps for second half of pedal press.

 

If you think about it, as the throttle plate pivots up more than half way, it takes much larger swings to create meaningful difference in volume of air flowing. There is probably no practical or measurable performance output difference between a 85% and 100% open throttle.

Edited April 23, 2016 by scottmcphee

[Phillip J. Fry]

I think this is normal throttle opening program to go in big steps for second half of pedal press.

 

If you think about it, as the throttle plate pivots up more than half way, it takes much larger swings to create meaningful difference in volume of air flowing. There is probably no practical or measurable performance output difference between a 85% and 100% open throttle.

 

what???

 

are you crazy?

 

if i applied your theory then why not go to 10. 20. 30? have you checked your throttle? i said i can hear a noise. if someone is using fine throttle inputs then, yes, that increment is huge. like at steady state cruising. the system will be fluctuating all over the place.

 

i can hear it jerking around like its confused. if not feathered right it goes straight to 99 from 55. a week before my incident i went part throttle to merge when the car suddenly lost a lot of pickup then i mashed the throttle more and it went forward.

 

if im right ita this throttle numbers ive found. they dont make sense. it may have dropped back even though i was with steady application.

 

also, the voltages dont have jumps.

 

if i want 60% throttle to merge i should have it. not 55 or 100.

 

you should go check yours instead of discounting it without testing.

[scottmcphee]

I just checked my throttle % and pedal % using BtSsm tool.

 

When the pedal gets over about 55%, then each decimal point 55.3 to 55.5 change starts resulting in multiple percent change of throttle .. 70 to 75.. it's very hard to try and get a specific throttle opening beyond about 55% because it is so hard to hold the pedal still enough (move it smoothly enough).

 

My throttle body buzzes and clicks in this wide margin area... it also buzzes a bit as soon as you turn the key on. I assume this is normal.

 

But I am battling a p2138 code which is a drive-by-wire error of voltages not in agreement in sensors.

 

Maybe my TB mechanism / sensor combo is hooped... dunno.

Edited April 25, 2016 by scottmcphee

[Phillip J. Fry]

I just checked my throttle % and pedal % using BtSsm tool.

 

When the pedal gets over about 55%, then each decimal point 55.3 to 55.5 change starts resulting in multiple percent change of throttle .. 70 to 75.. it's very hard to try and get a specific throttle opening beyond about 55% because it is so hard to hold the pedal still enough (move it smoothly enough).

 

My throttle body buzzes and clicks in this wide margin area... it also buzzes a bit as soon as you turn the key on. I assume this is normal.

 

But I am battling a p2138 code which is a drive-by-wire error of voltages not in agreement in sensors.

 

Maybe my TB mechanism / sensor combo is hooped... dunno.

 

ahhh. i can move it smoothly but i get ~20% jumps not 5.

 

your gas pedal sensor is not agreeing with the tps? ive read on this forum and some have removed the gas pedal then cleaned the contacts. it really is about 2 bolts and unclick the sensor. i just removed the connector and mine looked new without any dust so i plugged it back.

 

the problems ive had a few times now is going about half throttle and then trying to roll into speed but then it feels like the plate closes then my 5eat shifts to 3rd. it did this to me last week and i thought i didnt mash it down far enough.

 

my gf 2013 forester xt is smooth as can be. ive thought about plugging the access port into her car see if it can even read it then try to check her throttle numbers.

 

at this point in very highly convinced my throttle is jamming up in some way. i dont think hers makes this amount of noise and i know it wasnt like this when i bought it new 9.5 years ago.

 

my plugs are new. 02 sensor is new. i have 0 codes. just weird throttle modulation. the 2 cans of berrymen are working decently now and the car is at half tank at the moment.

 

my plan now is to get the throttle body removed, get it thoroughly cleaned and then check the tps.

 

ive since reread my reply post about "are you crazy". haha. didnt mean to come across as a donkey hole.

[Subi3snacks]

I'm having jerky throttle and loss of power. It comes and goes. One day it runs like a brand new car, other days throws a check engine light stumbles about stalls catches its self then light goes off after 20 or so miles with a code for the IACV and related problems... checked mine cleaned it, tested function ability. No vac leaks. Cleaned throttle body. Glad I checked this thread, I'm off tomorrow so I'll check the TPS and o2s and post my results.

[Phillip J. Fry]

next friday i have an appt at PIA to get the throttle body cleaned. perhaps look at the throttle numbers im seing as well.

 

im very sure its the throttle body and or tps.

[scottmcphee]

Here's my factory pedal to throttle opening mapping.

 

I measured this mapping from a scanner tool, after having cleaned the pedal sensor swept carbon traces VERY carefully with 600 grit and alcohol wipes on the traces... I'm in the process of trying to eliminate a p2138 code. I realize doing this kind of cleaning can be a quick way to buying a new pedal if the carbon gets gouged, so I don't recommend the process as a common practice. (I used strong jeweler loupe magnification and bright light with a steady hand making ultra light circular sweeping motions following each individual trace arc to remove obvious "finger" pattern polishing where the wiper had made its path over 200,000km of driving. It's pretty obvious when you look at a high mile sensor what I'm talking about.)

 

I guess I was careful enough, because the car is drivable and not immediately throwing any of the pedal codes on my first full drive cycle with lots of throttle variations including WOT.

 

Pretty linear up to about 50% pedal. Then very steep up to about 60% pedal where it hits max opening.

Everything from 60% pedal to full push is wasted. Car is already WOT.

 

This sucks, I am going to remap this with Tactrix and rom raider.

Edited May 3, 2016 by scottmcphee

[scottmcphee]

Other tid-bit...

 

An Autel scanner tool shows throttle opening % up to a maximum of about 78.8%, with pedal push of 60% onward to the floor.

 

The BtSsm tool, developed by member on this forum, shows that same pedal push opening throttle to 100%. This is the tool I used to capture the graph above.

 

So, it seems there's a difference in what the tools report as the opening, or how the tool is collecting and (doing math?) on the TB values.

 

Maybe the BtSsm tool is "smarter" about the min/max range of the TB sensor and scales it, knowing that full range opening value is 100% throttle and there just isn't any more to be had. So it reports TB 100%. ?? dunno just a theory.

 

I have not done a visual on the TB plate to see at what pedal % has it opened fully for real. But I'm pretty sure when the Autel says 78.8% it has fully opened, and that's the place where BtSsm says 100%.. there would be just no more to offer past that mechanical limit.

Edited May 3, 2016 by scottmcphee

[Phillip J. Fry]

this is great information man. on thursday and friday i will throw back what i find in my car and can try to reply with some theories or something. friday i have the throttle cleaning appt.

 

thank you

 

you have spent a great deal of time. its def helping those of wihc wish to learn a bit more detail about what the car is doing.

[OBieXT]

Here's my factory pedal to throttle opening mapping.

 

I measured this mapping from a scanner tool, after having cleaned the pedal sensor swept carbon traces VERY carefully with 600 grit and alcohol wipes on the traces... I'm in the process of trying to eliminate a p2138 code. I realize doing this kind of cleaning can be a quick way to buying a new pedal if the carbon gets gouged, so I don't recommend the process as a common practice. (I used strong jeweler loupe magnification and bright light with a steady hand making ultra light circular sweeping motions following each individual trace arc to remove obvious "finger" pattern polishing where the wiper had made its path over 200,000km of driving. It's pretty obvious when you look at a high mile sensor what I'm talking about.)

 

 

Now that you have carefully cleaned the sensor its time to goop it back up again. You need to smear a very, very thin layer of Vaseline across the carbon traces to seal out the atmosphere. Otherwise the oxygen will begin attacking the resistive material and over the course of 3-6 months it will become worse than it was before you cleaned it. Alternately, a shot of Tri-Flow will usually do the trick with out the need to disassemble. At least I've had good luck with it on audio equipment and servos for R/C airplanes.

[scottmcphee]

How about thin wipe of GM/Delco dielectric grease that is meant for automotive electrical contacts, instead of Vaseline?

 

I don't mind disassembling the sensor to do things right, no worries there. Besides I'm getting really fast at taking off the acceleration pedal off the car (2 nuts, 1 plug) and opening the sensor (2 screws) on my bench. NO WAY could anybody do an operation like this on a sensor while the pedal is mounted in the car.

Edited May 4, 2016 by scottmcphee

[OBieXT]

That may work also although I haven't personally tried it. It just has to be thin enough that the points on the wiper arm push through it to make contact with the resistive material. I'm not sure if it would thicken up over time and cause issues. Worse case is you have to clean it off with alcohol and try something else.

[scottmcphee]

How about thin wipe of GM/Delco dielectric grease that is meant for automotive electrical contacts, instead of Vaseline?

 

I don't mind disassembling the sensor to do things right, no worries there. Besides I'm getting really fast at taking off the acceleration pedal off the car (2 nuts, 1 plug) and opening the sensor (2 screws) on my bench. NO WAY could anybody do an operation like this on a sensor while the pedal is mounted in the car.

 

This seems like a sane source, that says it's OK to use silicone dielectric grease on pots:

 

http://www.w8ji.com/dielectric_grease_vs_conductive_grease.htm

 

In fact, the guy seems like the guru of the topic. Compares to Vaseline (low melting point).

 

 

 

 

Commonly used on:

 

Low power potentiometers or volume controls

for:

wear, electrical noise, friction

materials:

carbon and brass

Edited May 5, 2016 by scottmcphee

[scottmcphee]

Now that you have carefully cleaned the sensor its time to goop it back up again. You need to smear a very, very thin layer of Vaseline across the carbon traces to seal out the atmosphere. Otherwise the oxygen will begin attacking the resistive material and over the course of 3-6 months it will become worse than it was before you cleaned it. Alternately, a shot of Tri-Flow will usually do the trick with out the need to disassemble. At least I've had good luck with it on audio equipment and servos for R/C airplanes.

 

Hey, are you saying you actually cleaned a Subaru pedal sensor, and found you had to do it again because of no grease? ... or just based on other similar experiences?

[OBieXT]

No, I haven't cleaned a Subaru pedal sensor. This is based on other experiences. The pedal sensor is really just a potentiometer and I have cleaned enough of those to recommend you put something on it or you will likely become noisy again (electrically noisy, not audibly noisy) in short order.

[scottmcphee]

I'm just at the point of having a good pedal again that (knock on wood) hasn't thrown a p2138 code for days... A new record!

 

Scared of possibly wrecking a good thing by putting a greasy film on the carbon traces.

 

Did you know the wiper of the pedal sensors is not your typical looking thing you'd find on a regular pot? The wiper has tiny metal fingers bent and positioned to look like scrapers that sweep the trace. Just to see them you'd think they'd chew their way through that material in no time scraping back and forth... but it must be very hard carbon material. Put another way, they seem very capable of cleaning the surface they want to run on!

 

These wipers are not like the back side of a dimpled spoon shape rubbing back and forth like a saucer floating over water.

 

I do worry that putting dielectric grease film on those carbon traces might clog up the "fingernails" of these wipers, interfering with the electrical contact they need to make.

 

For now, no grease. I'll see if oxide becomes a problem.

Edited May 6, 2016 by scottmcphee

[CPUNeck]

Has anyone come up with conclusive evidence of pedal sensor failure? I'm not sure how to actually test it with conventional test equipment (multi-meter, oscilloscope, etc.) I understood those sensors were more like hall effect with no direct contact (wiper.) My car is not running right, and not throwing codes either...

 

New - fuel filter, fuel pump, Front / Rear OE O2's, spark-plugs, upper engine main harness... on and on. Smoke test=good, Exhaust leak=none.

 

Symptoms:

Sporadic fuel trims (14-16% LTFT) with STFT swinging wide - and + usually up to 25%.

 

Knock at steady load (65 mph, level road) with "light" throttle application (fine learning -1.4 to -2.4.)

 

AF Learning D / 4 = -11 to -14% (applied to open loop operation.)

 

Downshift 5EAT (2nd), turn right, ~90% throttle the car would almost nose dive and sounds like the engine is gurgling, knock, then finally it'll accelerate.

 

Voltage swings on one way commute typically 12.8-14.0v, with an average of 13.6-9v. Cold in the morning, volts really just fluctuate 1-2 tenths of a volt (14.3, 14.2, 14.0v), afternoon when hot those same fluctuations swing 3-6 tenths of a volt (14.4, 14.0, 13.8, 13.2, 13.5, 14.1v.)

 

Mechanical fuel pressure gauge exhibits very rapid flutter under certain circumstances. 1) Under light load, like drive, throttle at most +3% (right off idle), the needle "flutters" over an 8psi span. 2) Under high load, like passing at 80, throttle at ~80%, same flutter just higher pressure.

 

Key-On,engine off - throttle plate unable to follow accelerator (at 1/3 throttle the plate seems to jump to 1/2,) also it sounded very loud and I clearly heard what sounded like the plate binding just when it would "jump."

 

I Suspect:

1) The accelerator sensor could be feeding bad signals to the ECM sporadically.

2) The throttle body assembly could be damaged or failing.

2) The alternator might be losing a diode or the rectifier, leaking A/C voltage into the system causing the issue.

 

Pedal would be the simplest so I hope y'all had good luck.

 

Thanks

[Timmytech]

I'm working on an 05 legacy gt turbo. I recently replaced the clutch but after starting it up and attempting to relearn the tps I was unable to. I tried both the manual way by disconnecting the battery then letting it idle yadayada. I also tried via scan tool. Neither would take. It's throwing a p0123 code. It was throwing an o2 sensor code also, but it has aftermarket exhaust system that deleted the o2 sensors and the resistor tricking the ecm had fallen out, I replaced it and the o2 sensor code went away. Reading the position sensor, it reads that the tp is at 100% when idle and no foot on the accelerator and it chugs, push the accelerator to the floor and it idles normally. The accelerator pedal position senators are reading seamingly normal and changes normally when the accelerator pedal is pushed to different spots, voltages are within normal parameters. I also pulled the throttle control motor relay out from behind the dash and it is still good. I swapped a throttle body out of known well running legacy, same problem. The tps angle percentage is reading 100% no matter what. Sub throttle sensor voltage is steady at 1.7 v and main throttle sensor steady at 5v. The other issue I see is the throttle motor duty % is at 0 and does not change and throttle motor voltage is showing 0 and these don't change at all whether the accelerator is being pushed or not. I can't find any pinched wires or damaged connections and I'm losing my mind over this thing it's been here for months. Could the o2 sensor be causing this somehow?? Any insight or help would be much appreciated. Edited January 2, 2017 by Timmytech

[Max Capacity]

http://legacygt.com/forums/showthread.php/idea-web-links-saved-various-parts-219238.html

 

Check in there for a O2 sensor that we know work.

 

http://people.csail.mit.edu/ilh/vacation/

 

That may help too.