Oxygen sensor issues - hesitation in closed loop

[dimin]

Having some issues which im positive is related to the o2 sensor.

Replaced the factory tired sensor 3 weeks ago with an ACA universal one. Car ran perfect prior to replacement. After replacing took it for a drive, and straight away could feel hesitation, in closed loop conditions i.e cruising. Worst is between 2500 and 3200rpm, not really noticeable outside of that.

My old V4 Sti coupe sensor failed and had the exact same symptoms, hence why I think the sensor is dodgy. So I took the sensor back to the place I bought it and got a free replacement put in. Not to long after driving away I could feel it surging yet again

Have changed fuel pump, filter, spark plugs, brand new AFM before the o2 sensor as well. Engine fully reconditioned 2000ks ago. Also have checked TPS calibration and its spot on.

So anyone had issues like this before ? Especially with trying 2 new sensors ? Are these ACA branded ones just cheap s*** ? Also, unplugging the sensor, the car runs smooth as. Also smooth in open loop, WOT conditions, etc. For first few mins of driving from cold its smooth as well. As soon as it goes closed loop though theres hesitation straight away.

Any hints, maybe Ive overlooked something ? Might try a Bosch sensor perhaps....cheers

[EJ20T]

For this problem you really need to plug it in to a scanner that reads data, someone with even mediocore skill should be able to pick whats going on from the data, the ecu can adjust the short term fuel trim by up to +/- 25%, if u have a bad o2 its likely the output will swing from close to 1v then suddenly drop to 0 making the ecu make its max adjustment and give you jerkyness in closed loop when it pulls 25% out and makes the mixture very lean.

[Timmah]

To test your Oxy sensor is working: http://www.clubsub.org.nz/forum/index.php/topic,7041.0.html

I haven't heard of ACA o2 sensor's either sorry. Are the 3/4 or 5 pin sensor's?

[gazzy2000]

havent heard of ACA, i have always used bosch universal O2's and they have always worked great for me.

Does the car run ok when the o2 is totally unplugged?

[dimin]

 gazzy2000 said:

havent heard of ACA, i have always used bosch universal O2's and they have always worked great for me.

Does the car run ok when the o2 is totally unplugged?

As said above, runs smooth as unplugged. ACA are some aussie brand that does alot of automotive electrical products.

Have just bought a new Innovate MTX-L wideband so will fit that, and use the narrowband simulated output and see how it goes. Hopefully it will be all good, and il send the shitty ACA one back

[dimin]

Oh also, have tested the output of it. It does swing between 0.2-0.8v , but quite quickly, easy twice every ten seconds. Will update after I get the Innovate one in there. Least then I can see what the AFRs are doing properly. Definitely feels like it has bad lean spots when it hesitates.

[EJ20T]

haha, if it only cycles twice in 10 seconds you have a problem

[dimin]

Update on the issue.

Got a Innovate wideband and installed it. When it goes into closed loop the AFRs swing quite badly, from 13.8-15.5, as I thought, giving the stumble/hesitation.

So i wired the Innovates narrowband simulated output to the ECU, and strangely enough it has the same behaviour?! Really baffled by this now, 3 different oxygen sensors incl the wideband, and the problem has stayed. Forcing open loop it cruises perfectly at 14-14.4ish AFRs very smooth. But as soon as it goes closed loop it stumbles with the AFRs all over the show. For now im just running without O2 input and its fine, if a little rich. Is there anything else that can affect fuelling in closed loop like this apart from the lambda sensor ? It almost like the ECU is way over adjusting the closed loop fuel trims or something.

Will update if I sort the issue, but if anyones had similar problems and fixed it feel free to post up ! Cheers

[Guest]

i had similar problems same as you.but i didnt find it in the 02 sensor i found the afm was faulty.i pull all the intake off cleand it made sure everything was tight,clean the 02 sensor and cleand my afm will some afm cleaner,unpluged battery and left over night.

after that its been sweet .no problmes since.i duno if it was the 02 or afm but it was one of them

[dimin]

 sti8u said:

i had similar problems same as you.but i didnt find it in the 02 sensor i found the afm was faulty.i pull all the intake off cleand it made sure everything was tight,clean the 02 sensor and cleand my afm will some afm cleaner,unpluged battery and left over night.

after that its been sweet .no problmes since.i duno if it was the 02 or afm but it was one of them

Hmm the AFM is brand new, have cleaned with AFM cleaner, no change. Have now unplugged the o2 completely, so running open loop all the time. Runs very smooth with AFRs around 14.3-14.5:1 at cruise. With the o2 plugged in it cycles at least 1 AFR point either side of what I target(14.7 is the stock closed loop target). Just gonna leave it unplugged for now, its running very close to 14.7:1 anyway and runs so much nicer so will see how it goes.

[gazzy2000]

you sound like you know what you are doing, how about checking the TPS adjustment/voltage? easy enough to do and it is a sensor that is functioning during the conditions you describe, , just a thought... as if it were me i would be eleminating those sorts of things

[Guest]

 dimin']

[quote name='sti8u said:

i had similar problems same as you.but i didnt find it in the 02 sensor i found the afm was faulty.i pull all the intake off cleand it made sure everything was tight,clean the 02 sensor and cleand my afm will some afm cleaner,unpluged battery and left over night.

after that its been sweet .no problmes since.i duno if it was the 02 or afm but it was one of them

/quote]

Hmm the AFM is brand new, have cleaned with AFM cleaner, no change. Have now unplugged the o2 completely, so running open loop all the time. Runs very smooth with AFRs around 14.3-14.5:1 at cruise. With the o2 plugged in it cycles at least 1 AFR point either side of what I target(14.7 is the stock closed loop target). Just gonna leave it unplugged for now, its running very close to 14.7:1 anyway and runs so much nicer so will see how it goes.

hmm i wonder why its like that,kinda strane dont you think.i havnt tried my car with the 02 unplugged before,mine only hesitates in say 3-4th gear going up hills and stuff.

i gues its better to run a lil rich rather than lean.

[QRAF]

have you upgraded the turbo from the factory one? as mines been doing the same thing since a put my vf22 in i've replaced about 4 o2 sensors the fuel pump and the afm and its made no change i was thinking it was my injectors as after i put some injector cleaning in it goes away but come back eventually.

mine feels like i'm in a boat sort of a hesitation feeling.

be keen to know if you solve the issue as my car just runs rich with no sensor plugged in.

[Guest]

 QRAF said:

have you upgraded the turbo from the factory one? as mines been doing the same thing since a put my vf22 in i've replaced about 4 o2 sensors the fuel pump and the afm and its made no change i was thinking it was my injectors as after i put some injector cleaning in it goes away but come back eventually.

mine feels like i'm in a boat sort of a hesitation feeling.

be keen to know if you solve the issue as my car just runs rich with no sensor plugged in.

he prty much how you described it it what its like!mines not bad thoe.only happend unload on hills ect.iv replaced all except for 02 and afm.but they both seem to be ok now after some afm cleaner..

[gazzy2000]

also as part of the elimination process, i would also try either knock sensor reloction, replace the knock sensor or check the torque and angle settings..

If the motor has been out etc the knock snensor should have been torqued to 24nm, this apparently is critical as is the 60 degree angle from the back of the engine (where you found it)

(obviously a clean mounting surface, sensor and bolt is a must for a solid connection)

An il adjusted or faulty knock sensor can make the timing go al over the show, couple that with sensitive/touchy/intermitant o2 issues and you will get surging and hestitation etc.

(NOTE: excessive noise and/or interferance on the knock sensor wire can create false signals to the ECU too creating the same effect, but this a more of a last resort)

Faulty or il adjusted knock sensors dont neccessarily throw CELs either

Plus, to state the obvious:

- change the [glow=red,2,300]fuel filter [/glow] too

- make sure you have the same profile tyres on (mismatch tyres can cause the C/diff to bind up creating a slight hestiation feeling, making you think its engine/fuel/ign when its actually the drive train)

- earthing kit - clean all earths, batt terminals or fit an earthing kit, degration in the factory earthing system in ageing subarus is well known (possum link publications insist you run new earth wires when installing their plugin ECUs in GC & GFs )

am keen to hear what you find.

[Guest]

iwill check the tires and earthing kit.my earthing kit goes everywere throughout my engine bay.most of them look prty well earth.as for the tires have yet to notice anythgin diffent but iwill check that too

cheers gaz

[gazzy2000]

frig! almost forgot when changing sensors like the o2 and knock etc you must reset the ECU to let it it relearn the new and hopefully correct settings.

i fitted a new o2 into my STI (97) one time and it didnt do squat, reset the ECU and it was a different car.

Reset using the green and black plugs method & reset everysingle time you make a change or adjustment otherwise your wasting your time

[dimin]

Cool cheers for the suggestions gazzy.

Have checked and adjusted TPS, it seems fine. Knock sensor was the next thing to check, so will do that when I get a chance, engine has been out so it may be a high possibility that its not torqued or positioned correctly. Also, QRAF, no turbo upgrade, its just the stock VF28. But it does have an aftermarket downpipe, though the sensor is in the same position as stock. Have done fuel pump and filter, and spark plugs too. Leads looked fine under close inspection but perhaps I may replace those too eventually.

Will check knock sensor and update if it has an effect.

[QRAF]

i've also changed the fuel filter and spark plugs

[gazzy2000]

here is the article on knock sensor relocation http://www.imprezasport.net/forums/index.php?showtopic=785&st=0

it is based around the N/A imprezas sports that the poms are right into, but look at the symtoms when they are crook... al sorts of surging and hestiation.. bad fuel economy etc etc

when i had hestiation probs i spoke to PBMS and they are familiar with the K sensor relocation fixing those sorts issues... so the relocation is certainly not to be ignored & should be eleminated as a possible cause

[QRAF]

it says its not recommended to do to turbo cars ???

Still might give it a go tho

[gazzy2000]

 QRAF said:

it says its not recommended to do to turbo cars ???

Still might give it a go tho

when i trouble shooting det issues i relocated the sensor to that location they recommend, then right up on the manifold (for sh1ts and giggles) and it still worked.. i.e. it still detected the det and the ECU was hard out trying to pull the timing. So it works fine in my experiances at least.

[EJ20T]

You could have a small air leak somewhere, are you sure its not a small missfire when the mixture leans out thats throwing the mixture out at the o2 sensor then feeding the ecu rubbish? as a test get the o2 hot, then hold the revs at around 2.5 steady, then check the o2 output in this situation, u should see a fairly rapid cycling between .2-.8 v if it goes out higher or lower u need to determain if the ecu is causing it or a slight missfire, and this is where a scanner can solve a problem in 10min vs hours of guessing

[dimin]

Hmm a scanner may be the go. Im at my wits end now with the issue. Tried another different branded o2 sensor, brand new, same problem. The closed loop mixture cycles too much, like 14 to 15 afr's, thats far too much of a swing IMO, you can feel it going rich-lean which makes the car surge forward then pull back, surge forward pull back ,etc. I would of thought it should hold close to 14.7, maybe a .2 of AFR point variation but nearly a whole point is of course gonna cause driveability issues.

It almost seems like the ECU isnt controlling the short and long term fuel trims properly. Im thinking best way out is im planning to take it for an ECUtek tune, might see if they can tune it open loop only and disable the o2/closed loop fuelling. Did that on my V8 sti and it ran the best it ever did. So one thing to look into. Only other option is try a proper genuine subaru bosch one, maybe the aftermarket ones arent up to the task, even though I've never had an issue with previous cars and aftermarket sensors.

[gazzy2000]

i take it you tested for air leaks? (can of ether while car is running around inlet etc n watch for A/F ratio going rich)

failing that. why not borrow an ECU?