410 replies to this topic

[Robbks]

When you look at expected sensor voltages and resistance readings in the FSM across the years they are all a little different at 90'c

only use Subaru Genuine components for engine sensors and thermostats
Tridon are Supercheap rubbish.

use
Subarugenuineparts.com
SubaruOnlineParts.com
Patrsouq.com
And enter your VIN for the right part
 

[LukeFranky]

Yep, learning the hard way not to cheap out.

 

Anyway, correct me if I'm wrong but it looks like my rear O2 sensor is consistently reading very rich on throttle? If this reading is correct, then what does that mean if my front A/F sensor is reading lean?

 

Also, everywhere I've seen the rear sensor is post cat. Mine is before the cat. Is this normal?

[duncanm]

no - FSM says rear O2 is before the rear cat (don't forget the front primary downpipe cat).

 

If your front sensor is lean, and the rear rich - then one sounds kaput.

 

May explain a lot - ECU is confused by conflicting AFR inputs.

[LukeFranky]

(don't forget the front primary downpipe cat).

- Ahh, that didn't even occur to me.

 

If your front sensor is lean, and the rear rich - then one sounds kaput.

- Both are new. They were replaced when the heater element in the front one died. Whoever I ordered it through assumed I meant the rear O2 and sent that by mistake so I ended up with both. That means I have a spare for each but the front one only works once the car is hot.

[Robbks]

Rear sensor (o2) is pretty only used for closed loop running.
Front sensor (AF) looks at mixtures all the time and is responsible for AF corrections in both CL and OL

but MAF scaling also plays a huge part in AF readings and needs to be correctly scaled for the airflow past it so that the o2 sensors readings make sense

[LukeFranky]

Found an interesting article...

 

"Real-world air-fuel sensor diagnosis: 2002 Subaru Forester P0130 and P0171

One of our best customers brought their vehicle in because it had the check engine light on. The vehicle otherwise ran fine. So, she brought the vehicle and the light happened to be off at that moment. So, we did the oil change and sent the vehicle on its way. A few minutes after she left the check engine light popped back on. This is when the fun began.

The first thing we did was scan the codes.

After this we checked for TSBs and finding none, we looked for hits on Identifix. Apparently a lot of air-fuel sensors go bad, but the test recommended by Identifix perplexed us. It said to replace the sensor if the post-cat oxygen sensor was rich while short term fuel trim was lean.

Graphing the data showed some interesting results.

Obviously the STFT was totally off and was indicative of what would have been an oxygen sensor shifted lean or a major vacuum leak. The method that Identifix recommended was to look at the rear oxygen sensor data to see if it was “rich,” which would obviously indicate that the air-fuel sensor was stuck lean and thereby commanding fuel until the system was in reality rich though theoretically running lean. This appeared to be what was happening.

The rear oxygen sensor was at 800 mV, which is on the high side... I guess. However, that’s just not enough of a smoking gun for us.

So, we needed to find out if the air-fuel sensor was meeting specification. The Autoland Scientech Vedis II had a PID that gave the air-fuel ratio sensor as Lambda. Pardon the bad picture, but these screen captures are taken during real shop conditions. As you can see, the Lambda was elevated into lean territory, here captured at 1.21.

We added propane and the sensor would not budge. It was pinned lean.

A few minutes after we were done with the test the sensor started working well again and the Lambda fell to 1.00. STFT was normal. As far as we were concerned, we caught an intermittently bad air-fuel sensor in the act. However, we wanted to get the voltage specification for this vehicle when it was good, because manufacturers tend to use the same voltage for all the vehicles they have.

Testing this sensor required no fancy back probing or looking up anything on a wiring diagram. The sensor had a cover over an area with a positive and negative sign, designed to place to meter leads (Figure 1). On our meter we read 2.44 V. We simply replaced the sensor, checked Lambda, and were happy with what we found. The car was sent on its way and has not been back since."

[natho]

OK, I've had another look at the logs. Looks like the places where the mixture is the leanest is where there is no AF correction. I suppose that makes sense - but what I can't understand is why the rear O2 sensor shows a rich signal when the front shows a lean mixture. I've never checked these logging parameters on my car either.

 

what were the part numbers you used to replace your oxygen sensors?

[LukeFranky]

I don't know the part numbers to be honest. I got them through my father's workshop. They are OEM and looked exactly the same as what was on there already so never thought they were incorrect.

I got it to his workshop again today. We're testing thr fuel pump, data logging it with his scan tool and then seeing where that takes us.

[natho]

i did some data logging on my car late last night. it also has the silicon inlet you fitted (funny story - it was the factory pipe on the b4 i had a few years back that was sent to china to get this made!), along with (as of yesterday) a cold air intake. primary turbo is also a modified one - vf32 core, ported p14 housing and vf31 comp cover. you'd expect all this to mean the tune is way off, but the AFR corrections are quite minor compared to yours.

 

in case you're wondering, the car is getting dump pipes and a front mount in the next week or so. wideband is going on when the dump pipes do, so i'll re-scale the MAF after that. once all that's done, i have a ported ECV and exhaust housing for the secondary turbo.

[LukeFranky]

So we just finished playing for now. My old man focused on the cam shaft sensor and thinks he found some weird voltages. He's guessing that maybe the other sensors could be being affected by weird voltages too.

 

We pulled the two front plugs, both show signs of consistently running rich while the exhaust showed signs of running lean.

[LukeFranky]

Everytime we reset the ECU the car runs better but pregressively gets worse. This leads me to think the car thinks it is lean but actually isn't. It is also pulsing the boost again.

[natho]

it might be worth double checking those oxygen sensors are correct. perhaps start by swapping the original front one back in, reset the ecu, let the car warm up and drive while data logging?

[LukeFranky]

Car is definitely running rich. Drove behind it today, saw it blow smoke. We're looking at the O2.

 

Anyone know where the ECU earths to?

[LukeFranky]

So I'd thought it might help if I summarised my investigation so far.

• ECU thinks the car is running lean.
• Rear O2 says the car is rich.
• Shows signs of running rich, spark plugs, black smoke, etc.
• Surges at low boost. e.g. car might pulse between 3 and 6 psi with a fixed throttle on the highway.
• Misfires at idle/low revs with a cam shaft sensor code.
• Misfire at idle/low revs gets worse when voltage drops (i.e. lights on)
• Resetting the ECU makes the car run better for a while.

 

Notes:

• Both O2 sensors aren't old. Believed to be correct OEM parts, part number has rubbed off now.
• Cam shaft sensor appears to be a bad earth or cabling. Possibly alternator.

 

Questions:

• Where does the ECU ground to? I'd like to clean up this grounding point in case a bad earth is affecting the cam shaft and O2 sensors.
• What possibilities could lead to a false positive lean code?
• Anyone know what voltages both O2 sensors should be at, idle or otherwise?

Cheers,

[natho]

I'd suggest trying your original front o2 sensor and seeing if this makes a difference. It's not that hard or time consuming to swap it.

 

Make sure all the extra engine earths are properly attached - i had some CELs at one point that cleared up just from re-connecting a couple of these i'd forgotten.

 

Will PM you about the factory manual which is probably your best bet for working out the earth points etc.

[Robbks]

Another option with the rear o2 is to disconnect it completely.
the ECU will throw a code but should not throw limp-mode.

It "should" force the ECU to run in open-loop and you only have to deal with troubleshooting one part of the AF corrections at a time.

 

[LukeFranky]

Another option with the rear o2 is to disconnect it completely.

- Good idea. Will try it.

 

I'd suggest trying your original front o2 sensor and seeing if this makes a difference.

- I'll try, but the cable on the original front O2 sensor got cut somewhere along the line. Going look at my mates WRX to see if it has the same one, might be able to try it.

[RX25SE]

I thought the rear (post cat) sensors were for monitoring the cat only?

 

Post cat O2 sensors do not usually report the burn to the ECU.

 

Have you tried a sniff of propane in the intake (upstream of MAF) and see if it drives the front O2 sensor rich?

[duncanm]

• Anyone know what voltages both O2 sensors should be at, idle or otherwise?

 

FSM says:

 

• Front and Rear heater 0 - 1.0V ON and Idle
• Front (A/F) Signal (+): 2.8 - 3.2V (ON and Idle)
• Front (A/F) Signal (-): 2.4 - 2.7V (ON and Idle)
• Rear Signal: 0V (ON), 0-0.9 (Idle)

 sensor_voltages.png   97.73KB   4 downloads

 sensor_voltages_2.png   106.46KB   4 downloads

[LukeFranky]

Yeah. I did actually spray a bit of a can of flammable liquid. Front O2 didn't register much of a change if any.

If you asked me yesterday I would have told you my front O2 is genuine, but knowing my dodgy memory I looked into where it came from. Here comes my guilty confession: It looks like it's actually a knock off. My reciept found me out. There's no way I paid $150 for a genuine part.

I remember now. There was a mix up and I was sent a genuine rear O2 sensor instead of the front. When I was quoted for the front I must have gone looking elsewhere. Idoit.

This current O2 must have worked while I was having no primary boost because it never registered a problem until I got boost back.

Tried rewiring the old O2 but once those cables are cut it is difficult to rejoin them, solder doesn't work. Ecu complained about it right away.

Bit the bullet, ordered a new genuine Denso front O2.