Proper O2 Sensor waveform

Status
Not open for further replies.
Joined
Sep 20, 2016
Messages
493
Location
MN
Does anyone have any experience spotting the characteristics of a lazy oxygen sensor? I have a feeling the oxygen sensor in my beater is not acting as it should and is throwing off my long term fuel trim. I'll see if I can post a couple examples of what it looks like while warm idling, under heavy load at 60mph going up a hill, and cruising at 60mph with an average load on a flat surface.

You can see in the first image while idling that the waveform appears to peak at 0.8 volts, with a low of 0.1 volt. It seems to be more erratic than it should be just for idling...


In this shot this how the waveform typically looks like under heavy load at about 1700 RPM, you can see that it really doesn't peak higher than 0.7 volts and really doesn't go lower than 0.2 volts. If I understand a proper function O2 sensor I think the waveform should be constantly going between 0.9 and 0.1 volts


Here is a shot of cruising at 60mph, about 1700 RPM on flat roads with little load. Pretty similar to the above


The car is a 98 Olds Intrigue, series 2 3800 I just picked up, has 148,000 miles, dont know much history other than the sensor is likely original and I just replaced the LIM gaskets and replaced the upper plenum.

My long term fuel trims around about +12-15 at idle and +7-9 cruising. This says its probably a vacuum leak, but I have checked, and I have great idle vacuum too.

Anyone have any opinions?
 
Under heavy load the 02 input is not used, instead lookup tables are referenced in open loop enrichment. Is everything else healthy? Tight intake tubing, all vacuum lines good, PCV working?

Any MIL codes stored?

Stock Airbox?

TB cleaned recently?

Power brake booster vac line/bladder integral?

Is tailpipe carboned up w/ fluffy black carbon?

Have you tried "cleaning" the sensor with sustained WOT operation?

Does not this 3800 utilise two sensors (bank 1, bank 2) as it OBDii


I just cleaned the road salt crust water off my 02 with a bit of melted snowball "water" and my engine performance on my NEW(ish) Rogue - that had recently been lazy and slug like - perked WAY up. This bothers me as in my vast experience Lamda controlled engines have typically detuned themselves more than they've tuned themselves - so there appears a basic faulty paradigm - as in audio circles: CD music replay. I point to the oft contaminated and thus faulty Lambda sensor. Almost the last word in closed loop tune.
It truly needs a foil and a failsafe
 
The front sensor (pre cat converter) should fluctuate up and down between .2 and .8 volts under idle. The rear 02 should show a straight line at .5v
 
Last edited:
Originally Posted By: ARCOgraphite
Under heavy load the 02 input is not used, instead lookup tables are referenced in open loop enrichment. Is everything else healthy? Tight intake tubing, all vacuum lines good, PCV working? Yes, all recently gone through when doing the upper and lower intake manifold gaskets and plenum. I realize the sensor data is not used at high load.

Any MIL codes stored? Nope none

Stock Airbox? yes

TB cleaned recently? Yes, recently cleaned

Power brake booster vac line/bladder integral? checked, good

Is tailpipe carboned up w/ fluffy black carbon? nope, burning clean, seems to only use about a quart of oil every 4000 miles, which is good for this engine

Have you tried "cleaning" the sensor with sustained WOT operation? I'm a firm believe in the italian tune-up and it gets exercised every day

Does not this 3800 utilise two sensors (bank 1, bank 2) as it OBDii Now that you mention it, this engine is a little odd being OBD II and it only has one pre-cat O2 sensor, its located on the rear bank where the rear bank exhaust manifold and crossover pipe from the front bank come together




The mass airflow sensor has also been cleaned recently with electronic parts cleaner and it seems to be reading accurately. when fully warm at idle in park it will read about 3.8-4.0 g/sec which is right where it should be, which is more data pointing towards not having a vacuum leak.
 
I don't think that voltage graph is going to tell you anything more than the sensor is responding. If you google "O2 sensor response graph" you'll see why. The sensor is massively non linear so the ECU must constantly average/filter the response peaks and valleys to approximate where stoich is. If you got a Wide-Band sensor/s with signal conditioning you would get more stable output.
 
Originally Posted By: Hemispheres
The front sensor (pre cat converter) should fluctuate up and down between .2 and .8 volts under idle. The rear 02 should show a straight line at .5v
This ^. Rears, however, usually last a long time. The steady .5 volts means the converter is working OK.
 
Last edited:
They look more or less typical. You can check the O2 sensor operation by forcing it rich with propane or carb cleaner and lean by pulling the brake booster hose, plugging it with your thumb, then partially removing it while watching trims and O2 voltages.

Your elevated trims could be a few things... Could be plugged injectors, a vacuum leak of some sort (you may not have a leak directly at the upper intake- it could be internal at the lower or leaking between the upper and lower), a weak fuel pump and so on.
 
I think you will get a check engine light if they get lazy. every time I've changed a primary 02 no change was ever noticed.
 
I don't remember now but you can tell by voltage. I think the voltage will start to move slower. I have seen in happen on a vehicle I used to have and the service manual told me how to diagnose it but I don't remember the exact way to do it.
 
Those graphs are from Torque Pro by the looks of it.

Just something to keep in mind in that case:

The OBD2 dongle you use to interface with the car and Android device has a relatively slow and limited response time and update period. From personal experience I can tell you that the front O2 voltage graph fluctuates a lot more than what Torque Pro is able to indicate - in my case I am comparing that same graph seen on Torque Pro with my Mercedes STAR computer and also my VAG interface. My Autel 802 scanner also produces way faster changing waveforms with much higher/lower peaks because it is also able to respond and track the values seemingly much more accurately.

Re your posted idle O2 graph ....................Even with that known limitation you should be seeing a way more constant cyclic sine-wave like O2 waveform at idle with certainly no periods of "inactivity" like you seem to be experiencing there.

Id say there is something not right there - could be the O2 but one would have to take a more detailed look at the short and long term fuel trims (at least!) - that is assuming everything else is in good shape and working as it should

"My long term fuel trims around about +12-15 at idle and +7-9 cruising" .................. LONG - sure? - explain why so wide a range?
 
Last edited:
Tman220 -
Try clicking the "faster communication" box, and see if the graphs differ from what you've posted:
C8ljpuJ.png
 
Yeah - even with that ticked you are not going to get near what the professional tools are displaying
 
It'd be wonderful if that instead of telling us, you'd show us the superiority of your professional tools on a J1850 VPW (10.4 kbit/s) application such as this.
 
Originally Posted By: Ramblejam
It'd be wonderful if that instead of telling us, you'd show us the superiority of your professional tools on a J1850 VPW (10.4 kbit/s) application such as this.


I agree 100%. The trick to getting pro-scanner like accuracy with apps such as Torque is to buy a top end bluetooth OBD II Unit. The quicker the refresh rate on the signal from the bluetooth to the device the more accurate the real time info displayed is.

Torque Pro actually uses the manufacturer created PID (Parameter Input Device) so it's virtually the same as a high end reader. The only thing it doesn't do is actually let you send information to the sensors as you can do with something like a Tech II or Snap On Solus.
 
Last edited:
Originally Posted By: Ramblejam
It'd be wonderful if that instead of telling us, you'd show us the superiority of your professional tools on a J1850 VPW (10.4 kbit/s) application such as this.
ISO shouting at me for giving valid info based on personal experience in the past, why don’t you consider the fact that I am not necessarily in a position to immediately post the obvious "evidence" you apparently crave. I will be in the shop later today and I can then post comparative graphs on eg a MB W168 I am currently working on ................ so don’t go getting your panty in a bunch for nothing yet
 



This is an example of what I am talking about: (same info off the same vehicle for the same period)

W168 had a brand new O2 installed and the ECU was allowed to get used to it over a few heat cycles and about 50km distance since replacement - as you can see the ST FT and LT FT is just about as perfect as one could expect from such an old control system.

The Autel 802 scanner there with its dual display can show and track very fast fluctuations in the blue section (updating about 10 times a second judging by eye), and even in the white section the values shown is obviously much closer to reality. If I were to also fire up the Mercedes Star computer, I can produce even better graphs with even faster and better curves for the O2 or any other sensor you care to mention due to its vastly superior scanning/logging/updating abilities

I like to use Torque and the dongle to very quickly demonstrate funky mixture control to a customer - that particular Torque layout of mine I posted is worth about 2000 words IMO

In general LT FT needs to be close to 0 and shouldn’t change faster than about every 5 minutes (if even that) and ST FT avg ideally too (though osscillating up to a few times a second as per ECU control abilities) ..................... that is if the motor is properly warmed up and the exhaust and cat is at normal temp too
 
Originally Posted By: LvR
Originally Posted By: Ramblejam
It'd be wonderful if that instead of telling us, you'd show us the superiority of your professional tools on a J1850 VPW (10.4 kbit/s) application such as this.
ISO shouting at me for giving valid info based on personal experience in the past, why don’t you consider the fact that I am not necessarily in a position to immediately post the obvious "evidence" you apparently crave. I will be in the shop later today and I can then post comparative graphs on eg a MB W168 I am currently working on ................ so don’t go getting your panty in a bunch for nothing yet
Let me clarify what he actually meant. (That would be first!) He was implying that the protocol itself used on the 1999 vehicle is slow and that no standard tool will be able to communicate it with any faster. For crying out loud, he even gave you the bit rate and that should have given you a clue!

Of course if you use HDS or similar manufacturer specific proprietary tools which do not use that slow protocol, in theory you can get faster waveforms.
 
Last edited:
I was using cheap OBDII bluethooth interfaces and finally got tired of them for being so slow. I bought a OBDLink MX Bluetooth Scan Tool from scantool.net and it is lightening fast. No way to follow it visually but the data recorder can capture the data in milliseconds to a excel file.
 
Which vehicle and which protocol? If it is newer vehicle and CAN bus, then a better scantool will have higher performance.
 
Status
Not open for further replies.
Back
Top