Tell me what your rear O2 sensor says?

[scottmcphee]

Could I ask someone to please do a quick experiment for me using a ECM reader tool that sees dynamic values of the ECM (I use the BtSsm tool being developed by a forum member). I need to compare my observation with a typical car not having the same issue I am.

 

Backstory: I'm getting a P0139 rear O2 slow to respond code. ECM shows rear O2 voltage shows stuck always at 0.6v. When testing the sensor itself, it has a nice operating range of voltage. I was able to get ECM reporting nice swings and normal voltage after cleaning grounding points on engine. But it's reverted to constant 0.6v and throws P0139 again. I want to know if it's faulty connection, faulty sensor, or some other fault that is causing ECM to be fooled into reporting 0.6v.

My basic ask is, if you have a healthy rear O2 and no complaints from it...

 

Tell me your observations of its voltage response as the ECM sees it. Go for a drive cycle taking the car from a cold engine start then up to normal operating temps on a short drive around the block, then cruise on a warm engine on flat ground with no heavy throttle inputs for a stretch... so traveling in a good fuel economy kind of way.

 

What does your rear O2 voltage do on this journey? This is the O2 downstream of the cat. conv.

 

THEN (the big ask) and what I really really need to know, when parked at your garage again after the drive, fully warmed engine, rear O2 is heated, and car still running... pull the rear O2 sensor connector (my O2 plug is easy to reach from the front of the car standing at an open hood.) Watch for fans and fingers... of course.

 

WHEN YOU unplug rear O2, while watching your ECM values what gets reported for rear O2 voltage?

Did value jump to some state... maybe in the middle of the range around 0.5 or so?

Did the value stay stuck at its last know state?

Did it drop like a rock to very low value?

Or get pegged very high on the scale?

 

For extra points rev the engine it a bit... maybe drive it around the block with O2 still unplugged.

 

Does the value stay stuck at what you see when it was unplugged?

Did it change to something else?

Did the car show fluctuation in rear O2 voltage, as some kind of phantom or "virtual O2 voltage" as you drive?

 

Top score and superhero status for anyone who does a couple more easy experiments with more sensors.. While engine is hot and running, and the rear O2 is still unplugged. Pull the EGR hose off the intake so the car sucks winds (goes lean)... report what ECM says about rear O2 voltage. Did it change? Put that hose back on. Now pull MAF sensor (on your intake) the engine should rev up... what does the rear O2 voltage say, any change?

 

Replug everything, your car will have belched several codes by now. Reset.

 

 

THIS IS A HUGE THANKS in advance if you'd be so kind to post your result.

[apexi]

I couldn't do exactly what you asked for tonight, but I did try some things.

 

When I went out to my car tonight, it had been sitting in the garage for about 1hr 30min, so it was still a little warm and not completely cold. Coolant temperature when I got in the car was 111F.

 

When I started the car and it was idling in my garage, my readings were kinda all over the place, as seen in the first attached image.

 

When driving, voltages varied depending on how much the gas pedal was pushed down. The second attached image shows what it was like during normal driving, with what I'm pretty sure was fairly consistent gas pedal pressure.

 

The third attached image shows the voltages when I was sitting in my garage, idling after I got home.

 

I disconnected the rear oxygen sensor while idling in the garage, and the voltage just dropped to 0. Revved the engine with the rear sensor unplugged, still nothing, just 0.

[08SpecB_DE]

Lazy sensors are not uncommon. If it's getting stuck you need to replace it. Much like Apexi's findings, it should be constantly fluctuating.

 

Ohm testing the harness will let you know if you have any resistance to that sensor. The sensor working after cleaning the grounds could have been a coincidence. If you are getting any reading besides zero when you unplug it then something is wrong in the harness, ecu, or your scanner. You will have voltage at the plug but it should not show up as a reading on the scanner.

[apexi]

What brand is your front and rear o2 sensor Scott? I also read somewhere that a clogged cat could possibly make the rear o2 slow to respond? But I think if that were the case, you would get a code, p0420 or something?

[scottmcphee]

Lazy sensors are not uncommon. If it's getting stuck you need to replace it. Much like Apexi's findings, it should be constantly fluctuating.

 

Ohm testing the harness will let you know if you have any resistance to that sensor. The sensor working after cleaning the grounds could have been a coincidence. If you are getting any reading besides zero when you unplug it then something is wrong in the harness, ecu, or your scanner. You will have voltage at the plug but it should not show up as a reading on the scanner.

 

My sensor appears stuck when viewing ECM parameters it is reporting 0.6v. Tool I'm using is BtSsm (Bluetooth to Android) shows real-time data. Lots of other dancing engine parameters, pedal and throttle % positions are showing my plunges on the pedal... as I'm watching O2 so I know the tool is collecting and reporting lots of data.

 

When the sensor is doing this stuck thing... I can disconnect the rear O2 at the its plug, so ECM is now blind to it and I do an actual digital volt meter test of the sensor pins and it reads a good wide range of voltage fluctuations 0.1 to 0.8 or so depending on what I'm doing with engine (pulling EGR hose to induce a lean mixture with a vacuum leak sensor drops to near 0 instantly, pulling MAF induces rich mixture sensor goes to 0.8).

 

I have to do a double check again... as to what the ECM is reporting seeing when the sensor is disconnected in this state. I THINK it's reporting 0.6v.

Which would be really weird with the sensor unplugged, don't you think?

 

And is why I begged for this experiment to be done by someone else.

[scottmcphee]

What brand is your front and rear o2 sensor Scott? I also read somewhere that a clogged cat could possibly make the rear o2 slow to respond? But I think if that were the case, you would get a code, p0420 or something?

 

First of all, THANK YOU Apexi for running this experiment!!

 

Brands were from Rock Auto... I believe Denso was one, and different brand for other.. (I've done both front AFR and rear O2). I have heard Subaru are fussy about brand and like Subaru the best. What's in there now is not Subaru. I DID KEEP THE OLD originals. Might be fun to insert those.

 

I have seen a p0420 in the past, but that went away after sensor replacements (months ago) which I think is good because I don't want to buy a new cat.

 

But that wouldn't explain what I'm seeing as a stuck rear O2 voltage from ECM perspective, when in fact the sensor is generating good range of voltages when tested out of circuit but still in its exhaust hole.

 

And I DID SEE for a short while last week after cleaning grounding points of the ECM to block lugs... the ECM was getting and reporting the O2 voltage swings that my VOM showed when testing sensor directly. So I don't think there's a complete brand incompatibility. It was possible and saw with my own eyes for a while the ECM was respecting what it was getting from the sensor. AND the car was not coding while that was going on... life was good for a couple days.

 

It's great to know the expected behaviour of disconnecting a rear O2 is to see a drop to zero at the ECM reported voltage level.

 

I will double-check of what an unplugging of my "stuck 0.6" sensor does at the ECM. If it stays stuck at 0.6... it's wiring issue or ECM problem.

 

 

Apexi: what tools do you use for monitoring ECM? And are those "seconds" across the bottom of the O2 graph?

[scottmcphee]

FYI... the real cause of my concern is not so much the rear O2 issue. It's the P2138 drive-by-wire throttle disconnect that came up a couple weeks ago.. And randomly with semi-regularity since. This is a dangerous / driveability issue (limp home). Trying to solve that one has me tearing apart pedal sensors and throttle body and my hair... and can't isolate the issue to those devices.. (could be wiring issue... may be seeing a common thread here)

 

I would also be happy to find and fix the cause of P0139. If it is a wiring issue or ECM... it could be the same thing creeping over into the throttle control area of the car. P2138 code means the pedal position A/B redundant sensors are in disagreement. (Several other threads I'm playing out that issue).

 

Here, in this thread, I truly would like to solve the P0139 O2 issue.

[apexi]

Right now I'm using an ELM327 and Torque Lite. I'm not sure if those are seconds at the bottom of the graph, that's a good question, I was kind of wondering that too. Later on I'll go check to see if it seems like they are seconds.

[scottmcphee]

Ok, updates.. I unplug rear O2 sensor and ECM tool is reporting 0.6v stuck there like a pig in the mud.

 

The car also throws another code for rear O2 heater too low when I unplug. Ok fine, that's understandable.

 

I check the sensor directly and its showing good signs of life.

 

So I'm wondering is my ECM bad, or is my tool bad? I should be seeing the value drop to zero when unplugging, right?

 

I decide to play around a bit with the rear O2 socket that the sensor would normally plug into, to see if I can induce a reading on the ECM tool.

 

With sensor unplugged, I test for voltage present on the sensor lines going to the ECM. Nothing appreciable across the two sensor lines.

 

Next I want to see a zero volts or some change at the ECM. So I jumper the two sensor lines together to make a short circuit of where the sensor would have appeared in circuit. Still stuck at 0.6v. Then I ground both sensor lines to engine block grounding point. ECM still showing stuck at 0.6v

 

But I notice engine rev difference when I ground out the sensor, and a rev change again when I unground it. Subtle but very repeatable. The engine is reacting to O2 sensor input changes but the tool is showing 0.6.

 

I have the feeling my tool is not reading and or computing and or showing rear O2 level correctly.

 

For giggles I tried a similar experiment on the front AFR O2 sensor. When I unplug it, my tool show the AFR at 14.7 and stays stuck at that number with engine revving. It actually took a restart of the App and a reconnect to the OBD dongle and maybe an engine restart to shake the 14.7 and get back to reading the AFR again and showing it correctly once it was plugged back in... it didn't just start showing the correct numbers automatically upon replugging AFR.

 

So the tool has some interesting "default" value display behaviour.

 

I am going to get that figured out first, or switch tools.

[scottmcphee]

Right now I'm using an ELM327 and Torque Lite. I'm not sure if those are seconds at the bottom of the graph, that's a good question, I was kind of wondering that too. Later on I'll go check to see if it seems like they are seconds.

 

I tried a little blue ELM 327 bluetooth plug and torque on my car, and it would pair with bluetooth with my Android, but torque could not read or find the protocol of the ELM and never got as far as connecting to the ECM.

 

I actually buy things like that in pairs, they're so cheap... but both of them had the no connect issue. Gave one away and it worked fine on my buddy's GM truck with torque.

 

I got the feeling that some of these things are very fussy to my model year Subaru, that's when I decided to try BtSsm... and got that working pretty quick.

 

What brand or type of ELM are you having success with?

[08SpecB_DE]

Definitely sounds like your scanner is not working properly. I don't have a wiring diagram in front of me but is it a 3 wire harness? I'm going to assume one of those wires is a signal wire. Hopefully you didn't back probe that wire.

[scottmcphee]

My rear O2 is 4 wire. 2 heater. 2 sensor.

 

Why do you hope I didn't back probe a wire?

[08SpecB_DE]

You can fry the pcm by accidentally sending it power.

[scottmcphee]

My tools were actually fine. Reporting what they read from ECM. I verified with a second scan tool of different type, and hung a volt meter off the suspect pin.

 

The physical voltage present inside the ECM at the O2 input pin was stuck at 0.6v. So I had 3 things telling the same story.

 

When the rear O2 harness checked good, end to end.. I suspected the ECM and its connection to the outside world.

 

Looking closely at the circuitry on the ECM, I see a weak pullup resistor on the board for the rear O2 incoming line. This explains why the line floats up inside the ECM when there is no cable (harness) connected to that pin. It is such a high impedance pullup, that just hanging a wire on pin brings it down. But when no good connection is made at the pin for that wire, it continues to float high inside the ECM.

 

The problem seems to have been poor pin connection at the ECM side. I cleaned pins there very carefully and got the ECM to start reporting the normal good voltage swings the O2 sensor was trying hard to tell it!

[08SpecB_DE]

Excellent! Glad to see you got it fixed.