Lean combustion?

[Chas54]

Thanks guys

I have a few thinks to consider now. KenS has a trace that looks somewhat similar to my faults. Disconnecting an injector interferes with the O2 trace . Now i'm wondering if i have some grotty wiring at the injector supply? Ill have to have a closer look at KenS trace this evening.

Im going to order a new O2 sensor and looking at the Walker sensor on ChrisM trace , i think ill try rock auto for one.

My cleaning technique involves a baked bean tin full of E5 petrol placed in my water filled sonic bath and the sensor placed in it , but without the cable section submerged. Moderately heated 30c for 30 minutes . I haven't thought to try white spirit as i felt it was to oily but am reconsidering this now. I do have some acetone left over from fibre glassing clean ups , i may try this .

If i understand right then, when i see a 14.7 to 1 fuel air ratio , it is most likely a set target as to a live figure when in closed loop.

One thing i'm doing today is checking the feedback cable from the O2 sensor to the ECU plug. As i know the 12v heater connection is OK and when i get my fault there is no (or very little) voltage produced.

I have been wondering what determines the O2 sensor sampling rate frequency ? It surely must be something within the ECU . Its not a steady trace and seeing differences in peak to peak time values has me wondering if whatever parameters are used in the sampling rate calculation , maybe one of those is stopping it working.

Yesterday evening after replacing the wheel arch i had a short run of about 5 miles. I let the car warm up on the drive and all seemed well , running smooth and no ECU light on. I hadn't cleared the faults recorded though . It ran really well until i had to stop at some road works for a few minutes and round about that time the ECU light appeared . The road was not rough at this point though.

Ive attached my elanscan trace for this run .
When i look at it i see the O2 sensor started working around 25secs , then stopped at 82 secs. It didnt start working again until 500secs although at around 300secs the voltage started picking up . It also stopped again a couple of other times but i didnt spot these as the engine was running smoothly.

after wheel arch inner.zip

Thanks for taking the time to investigate my problems gentlemen , its much appreciated .

regards

Chas54

[Chas54]

ChrisM

Do you have a part number for the Walker O2 sensor


Regards

Chas54

[chrism]

Bosch 15701 is popular and lead length is good (some can be short, I gather...)

Walker 350-33022 is above, Denso is below. Different designs...

Erm, when I was testing sensor leads ages ago I saw a really small almost 'background noise' trace on my signal -> ecu lead - so that might be 'normal' - and I'm not sure there would be a lot to measure...

Edit: A bit confused, were you talking about heater voltage down or signal voltage down when the O2 signal dropped out?

ps need to try my spare Walker first?

[KenS]

Chas54 - I need to stress that I am learning as we get more plots and trying to make sense of what we see. I cannot be 100% sure this is leading to the O2 sensor but there are a few things now pointing at it. It is always possible that it's a combination of issues, say a slow to respond O2 Sensor together with a sticky fuel injector which together gets seen when the power demand changes, giving an excessively long period of lean so the ECU reverts back to Open loop. My plots show that those small periods of lean (and probably rich) do occur when demand changes and are longer with a failed injector, but they are short enough not to be timed out by the ECU. But can a slow O2 sensor alone cause the time out or does it need a fueling issue as well? Hard to be sure.
The slow warm-up of the O2 sensor on cold start also points to the O2 sensor, which could be the heater failing (but now crossed off the list). So what could cause the slow O2 voltage start, either truely weak mixture, low exhaust gas temperature or slow O2 sensor? The O2 sensor seems to be the common factor. Is there any history on your other O2 sensor?

On the sampling of the O2 sensor: As far as I can find out (happy if anyone can confirm or correct this) the O2 sensor, when fully warmed, gives a voltage crossing 0.1 to 0.9V when the O2 in the exhaust gas equates to the magic 14.7 Air/Fuel ratio, so as each pulse of exhaust gas passes it swings across that voltage range. If there is (lower?) content of O2 it doesn't get up to that point and stays below 0.1V, if there is high content of O2 the voltage stays above 0.9V (might have the high /low contents round the wrong way). The O2 has to make its way to the surface of the sensor, if the sensor surface has been contaminated it reduces the sensor surface area and filters the absorbtion of the O2 so its response can be slowed.
So I believe if an oscilloscope was used to monitor the O2 sensor voltage it would show a 0.1V-0.9V oscillation at a frequency = the engine speed x 2 (4 cylinders, 4 stroke). So at idle I think we have O2 sensor voltage at about 30 Hz. ElanScan will be sampling this data at a lower frequency, so we see an aliasing effect where the Elanscan depiction is significantly slower than actual and at points it may miss pulses.

aliasing.jpg

So in summary, worth doing all the 'free' checks first to rule out as much as possible. Trying another O2 Sensor next highest on the list but no guarantee, especially if you have confidence in your 2nd O2 sensor.

[KenS]

With reference to the 'after wheel arch inner plots' That section between 82sec and 500sec is odd, the O2 votage has stopped oscillating for a long period of time but it has not dropped out of Closed loop. Further on when the oscillations stop for a number of seconds, it does drop out of Closed loop. How can it stay in closed loop for over 400 seconds when its not receiving O2 sensor pulses? And it doesn't look like the ECU has tried to increase the injector pulse width/duty cycle to recover from the apparent Lean mixture. I'll see if I can learn anything more from the Closed loop service manual.

after wheel arch inner.jpg

[Chas54]

I found a Bosch O2 sensor on Rock auto but they cant send it to the UK as it is uncertified for Russian steel content .. Cant find another Bosch one at the moment.
Ive actually found a Walker sensor now at £55 in the UK on ebay so ive just bought it.

I can see what you mean about exhaust pulses at 32htz at idle and its the elanscan capture rate being the deciding factor in the pulse rate I'm seeing.

The" after inner wheel arch " run I think the O2 sensor stopped reporting back . I am pretty confident the 12v heater supply is now there all the time. The interruption is in the feedback to the ECU. All the time the 12v heater is in service and warm enough i should have this return to the ECU at some voltage . According to fault 44 code write up the pin D7 to D6 voltage at the ECU would be around 0.45 volts and vary by 1 volt higher if rich , down to 0.1 volt lean.

So even if the ECU decides that the throttle is wide open and so it changes to open loop to run rich and suppress any knock, running in open loop and using remembered values . Then the feedback voltage from the O2 sensor should still be evident on the elanscan trace . I'm thinking i might try running with the volt meter across the O2 sensor feed back cable at the wing plug and a convenient earth and see if i can spot the voltage dropping off when the fault occurs .
I could do this at the ECU plug across D6 & D7 but entering my wire probe into the plug block is awkward, although it may prove worthwhile . Hampered a bit by rain at the moment .

Thanks again , Its an education .

[chrism]

When you compare O2 sensor with road speed in the latest trace (I didn't overlap them for clarity but that enabled closed loop to be in the frame too) it seems that the sensor tried to fire up, couldn't but stayed in closed loop whilest idling.



As soon as any throttle was applied for movement, the sensor sparked up a little better - perhaps higher pressure exhaust heat - and stayed closed loop before falling out and trying to re-establish closed loop at smaller throttle openings. The last bit of closed loop on the speed trace might be explained by the coolant starting to heat up after the car stopped.



I'm still thinking 'naffed' O2 sensor and the need to try one which we know works?

edit - just seen that you've a new bought one

[dapinky]

I found a Bosch O2 sensor on Rock auto but they cant send it to the UK as it is uncertified for Russian steel content .. Cant find another Bosch one at the moment.
Ive actually found a Walker sensor now at £55 in the UK on ebay so ive just bought it.

Still listed on Amazon UK for under £40....... https://www.amazon.co.uk/Bosch-15701-Ox ... ZI4IO?th=1

Every so often Rock Auto chuck them on a 'closeout' deal (last time they were listed at about £17)...... but as they can't confirm that Vladmir isn't taking a cut, wont send them to UK...... however, if/when they go on a good price offer, I intend to buy a few via a friend in USA for future LEC requirements. I know this doesn't directly help now, and you have bought a Walker one, but may help someone looking in the future who searches and finds this thread.

[KenS]

I have put an oscilloscope on the O2 sensor and measured its output during warm idle and got a surprise. My belief that the sensor measured each exhaust pulse giving a frequency of around 33Hz and then sampled at a slower rate by Elanscan is completely wrong. The oscilloscope trace is:

O2Voltage-sml.jpg

The sensor output at idle has a period of around 1.4 seconds (0.7Hz) which is low enough for elanscan to track and display:

O2sensor-voltage.jpg

So at idle the ElanScan traces are fully representative with no possible upset or significant distortion caused by aliasing. I'll try and get an oscilloscope plot at around 3000rpm to see if the pulse period is constant or related to rpm. Seems like the pulse duration must be current driven by the ECU?
Sorry for adding possible confusion in interpreting the plots, a bit of over thinking!

[chrism]

So does this mean that Chas54's 'scratchy' black sensor trace shows that his O2 sensor is less healthy than my blue traced Denso sensor, since my trace seems smoother, stronger, more frequent and higher?

ie Could height and smoothness be a measure of sensor health? Could Chas have somehow degraded his sensor by cleaning it (even though I totally abused mine, believing it was knackered...)?

[KenS]

Chis, I think you are right, the trace could still be degraded by noise, poor connection or bad earth but now Chas has checked these out it does suggest the poor quality trace is due to the sensor.

I ran a 2nd scope plot at around 2500rpm and the O2 sensor voltage plot looks very similar, so the O2 sensor pulse frequency is fairly well fixed, it is not proportional to rpm:

2500rpm.jpg

Probably over-thinking this again, if the pulse frequency is not related to RPM it must be set by the ECU, so there is a very slight chance the ECU has a part to play in this. A quick Google gives:

ai-text.jpg

I guess we are Wideband.

Hopefully an O2 Sensor swap is the fix, and at that point everything sort of makes sense!

[dapinky]

Most of this thread has gone well over my head........ but....... I have something which is probably completely unconnected to throw in!!!!!!!

On my car, the O2 sensor 'drops out' when the headlamps have been on for 5 minutes or so. I know what causes it, and have a 'work around'....... however, it is specific to the fact that I changed my headlamp units to LED units.

The LED units required a resistor to be wired in to get the pods to work properly, and when this resister got hot it managed to cause the O2 sensor to give duff readings, engine management light on, but car still runs properly. (on open loop, presumably).

The fix was to move the location of the resistor to the NS headlamp, away from the O2 sensor.

Now, if you still have the OE lights, this isn't going to be your cause, BUT the fact that something nearby can cause interference may be an avenue of exploration!!!!

[Chas54]

Dapinky

Was that resistor one of those great big wire wound on ceramic things/

I can imagine one close to to the O2 sensor cable having an inductive effect possibly. But it still sounds strange. I haven’t got anything that close on the wing but maybe there could be something laying close on the feedback cables route to the ECU. Offhand though, i cant think of anything .l


Ive had an ebay message from the Walkers O2 sensor supplier pointing out that the ordered sensor is not compatible with my vehicle . But as it is compatible with a similar years Isuza ive gone ahead.

I’m still trying to figure out how the varying voltage from the O2 sensor interacts with the normal voltage at the D7 pin (0.32v) . If the O2 sensor causes this voltage to vary between 1v and 0.1v why do i see a higher voltage on elanscan for instance 3.5v ?

Cant really do much else until the new O2 sensor arrives but i will try to check across the D6 D7 pins on the ECU with the engine running .


Chas 54

[Chas54]

KenS

Based on nothing I know but reading up on narrow and wide band O2 sensors, I think we have narrow band..

There has to be more cables for a wide band .

Chas54

[dapinky]

Dapinky

Was that resistor one of those great big wire wound on ceramic things/
Chas 54

It's one of those big (gold coloured) aluminium things with the built in heat sinks..... I have no idea if it causes induction, or even if it is something inside the LED units, but shifting it about seems to have sorted the issue I was having...

... have you got any LED indicator bulbs? If so, have you got a ballast resistor fitted somewhere for those to flash at the right rate???

[Chas54]

No, still got the old filament bulbs all round

[RonR]

I’m still trying to figure out how the varying voltage from the O2 sensor interacts with the normal voltage at the D7 pin (0.32v) . If the O2 sensor causes this voltage to vary between 1v and 0.1v why do i see a higher voltage on elanscan for instance 3.5v ?

The original 'elanscan.ini' had an error in the multiplier used to calculate the voltage, my updated version has a correction for that, plus lots of other tweaks to give a bit more information.
Available here: viewtopic.php?f=34&t=30234#p382043

[RonR]

Is the 0.32V figure you quote for pin D7 a voltage reading taken at the ECU or at the O2 sensor?
Your scan output from cold shows 0.45V at the ECU, and I've always seen 0.45V on pin D7 with engine cold.

O2 at Startup.jpg

As I understand it, with the sensor plugged in the voltage varies around this centre voltage, kind of like a hysteresis loop as the ECU adjusts the fuel delivery and the mixture goes above or below 14.7:1.

[chrism]

<- My Denso initiates at from cold at 0.45V

<- Chas54's sensor also initiates from cold at 0.45V

Yep I agree with Ron's figure of 0.45V at ecu (pin D7) for both mine and Chas54's data traces - and my upper trace seems to show Ron's variation around this figure as my sensor does what's expected.

I guess the new ecu's trace might be able to spotlight this issue...which has encouraged some fascinatingly 'deep diving'...

Plus wherever the 0.32V signal to D7 is measured how come EScan 'sees' 450mV on the most recent scan (and seemingly, on all Chas' scans) Again, just asking...

[KenS]

Have I finally got an understanding of the normal O2 Sensor Voltage characteristics right?
The O2 sensor simply gives out a voltage dependant on the exhausts O2 content, for a weak mixture that voltage is around 0.1V for a rich mixture the voltage is around 0.9V. For an ideal 14.7 air/fuel mixture the voltage would be around 0.4V but the high gain/sensitivity of the O2 sensor combined with the relatively slow closed loop cycle time of the O2 sensor/ECU/fuel injectors makes achieving a steady 0.4V near impossible. So in normal closed loop operation the ECU is continuously adjusting the fuel volume causing the O2 sensor voltage to oscillate between the lean and rich values at a rate determined by the systems closed loop cycle time (around 0.7 seconds).
Then, according to the Closed Loop Service Manual, if the O2 Sensor voltage stops oscillating for more than 5 seconds the ECU drops out of Closed loop. On re-establishing an oscillating sensor voltage the ECU will re-enter closed loop mode.

Confirmation or corrections appreciated.