Lean combustion?

[andrewoitS2]

Hi All,
Just a warning, I couldn't import from ROK Auto for a new O2 sensor and used Amazon (it comes from the Amazon USA) yes original Bosh and cheap! Only issue is Amazon look like they kick them around and they are a very sensitive bit of kit. I ordered for about £78 from SJ and it came perfectly packaged.

Issues I had - have. After winter lay up the car was running rich and under turbo load vile kick backs. Elanscan showed that the ECU temp gauge had failed, I presume causing the care to believe it was running cold and constantly over fuelling and either causing the O2 sensor to fail (or very bad timing) Changing both transformed the car back to normal.

I jumped in here as my ElanScans did and still show chunks of data missing, I need to re load Elanscan (sure its the new Beta version, and try another cable)
I have cleaned pretty much every connection I could get to with WD contact cleaner including under the driver side arch

[Chas54]

Ive checked out the cables from O2 sensor to the ECU pin D7 and also the earth pin D6 . I have mauled and waggled all the cabling without producing a loss of continuity. Ive had the plugs off the ECU for a little light WD40 spray wipe over and reseted them well.

As i am waiting for the new O2 sensor delivery ( possibly this morning) I thought id do an elanscan from cold with my N0.2 O2 sensor in place .

As i have had the battery disconnected for the ECU work, all the faults are cleared . So to start i had to keep blipping the throttle until it would run evenly. Spluttering a bit while idling until coolant reached 82c or just over . Then running smoothly.

Looking at my scan , i can see the O2 sensor worked initially then stopped and didnt work again until the coolant reached the 82c mark. I had a voltmeter across the O2 sensor feedback cable at the O2 socket and the engine as earth . the reading correspond with those of the elanscan . I tried across the O2 heater and that should around 13v with little deviation. So i think the heater was energised OK but just not hot enough to work. Although it appears to have worked immediately..

There is also a at 1460 secs a quick blip of the throttle just to see if the engine recovers to idle OK afterwards. It did but the elan scan trace showed the O2 sensor feedback voltage dropped .. From the manual i think the scan stops at wider throttle openings and maybe that's what happens here ??

Anyway its cooling off now and my next trace will be the new sensor i hope.

Just to clarify my elanscan is on an old acer notebook running windows 7 . I have a 2nd version on a desktop PC running windows 11 but used only for viewing results, not recording

I may see if i can run the latest updated elanscan on my notebook but feel sure I'll have a battle getting the port to sample correctly afterwards. So i wont try yet .

Thanks again for all the advice and insights.

Chas54

after check ECU O2 connection S2.zip

[Tuga2112]

TLDR

a suggestion.

get the multimeter to track the voltage on the o2 sensor power line (the chocke wire from the alternator) see if there is a correlation to that and the scan readings.

im wondering if the regulatorh puts enough for the warning light to not be visible. but not enogh for the o2 sensor to activelly warm up. after the first "burst" or energy

[chrism]

Yes one of my random thoughts on the first page of the thread echoes Tuga's point about O2 sensor supply 'oomph'...

Original alternators, regulators and their connections are getting long in the tooth now. It's the next thing to consider if the new sensor doesn't deliver, I reckon.

Edit: again it looks like your O2 sensor might be getting 'lit up' by exhaust heat instead:-



Edit edit: And isn't it a coincidence that the sensor seems to start oscillating again just as the thermostat opens? Or could that be significant?

nb your trace started initially at 0.45V again; could your own 0.32V measurement be a half-cycle max or min? Isn't this what you'd get with a m/meter measuring a voltage through a diode rectifier? (Thoughts of 'root-mean-square' and 'half-wave rectification' stuff is tugging at my addled memory )

[Chas54]

Sorry Chris i may have given you a wrong steer here. The 0.32v is the low reading quoted from EMJ5 Code 44 error circuit description ..

I also thought the O2 sensor burst back into life at the thermostat opening point.. This started me thinking maybe the O2 sensor heater isn’t up to heating the plate enough until the engine and hence exhaust system is up to a high enough temperature . The fault light having been experienced when dropping the engine speed to 30mph to travel through villages and the exhaust flow and engine cooling a bit, then causing a little chilling of the O2 sensor. Its a stretch of the imagination but Tuga has suggested a lack of electrical power may cause the heater to appear Ok as the ignition light is off and i can measure 12v at the terminals. If i can prove this it would be a revelation .. Any way the new O2 sensor is here .
I am going to fit it and monitor at the body socket for the O2 plug, both the O2 feedback voltage and with a second meter the heater voltage.

If i can prove a drop in heater supply voltage correlating to the loss of O2 feedback voltage i could try removing the ignition to O2 heater plug +ve supply and replace it for a test with a directly cabled supply to the battey +ve . Applied when engine is running ..

Or maybe this new sensor will just work straight out the box ..

Chas 54

[Chas54]

new Walker W350 33022 O2 sensor.zip

New Walker W350-33022 O2 Sensor.

First glance suggests it has far more slots for the gases to get at the hotplate.

I've attached my latest scan with the new O2 sensor fitted . I also had two volt meters fitted across the heater plate and the O2 feedback to ECU.
Both appeared working with 13.5 volt and varying voltages from the feedback.. My initial thoughts are it doesn’t look that much different from my earlier one today. The sensor produced up to 3v until around 60secs then dropped back to 0.4 ish . Corresponding to my voltmeter . Roundabout 350 secs and 76c the sensor burst into life with a good steady swing close to peaks around 3v.. Revving the engine around 700secs shows the voltage drops off indicating lean mixture.

After switching off and stopping the elanscan, i noticed the sensor feedback voltage rising to about 0.12v . I guess the plate is hot enough to cause the reaction even then.

Next test is my usual route for an in action scan. I will have to unmount my voltmeters though, they are perched a bit precariously ..

Chas54

[Chas54]

after check ECU O2 connection S2.zip

earlier today elanscan prior to O2 sensor swap

[chrism]

Yes I can see that revving at 700 upset the O2 voltage swing briefly but your a:f ratio and closed loop stayed 'normal' according to your data. I can't see that the mixture weaked or you lost closed loop from your scan:-



The newest sensor started up quicker but struggled after about 40s this time coming back to action after about 350s, previous sensor dropped right out after 70s and didn't start up again until 750s...

Well the new sensor is responding more eagerly, it seems but I can't think that you'd notice any drop in performance due to fuel weakness?

It'd be interesting to see your subsequent 'action scan' data and to see how the check engine light, a:f ratio and closed loop respond.

[chrism]

...just in case the black lines cause confusion in the last scan...

And I wonder how the resistance of the two heater elements compare? I'd have expected the Walker to be a lower resistance for higher current to activate the sensor? My Walker was 3ohms - did you have a reading for your Walker and any of your old sensors? That might support Tuga's lack of regulator supply?

[Chas54]

Chris

Just been out and tested all 3 O2 sensors. Ive discovered a faulty lead on my Fluke meter as well which may explain my previous strange resistances . I do have a second cheap Chinese meter so tested with both to confirm these

The new Walker O2 sensor measured on the plug terminals 5.5ohm
The No1 O2 sensor 14.3 ohms
The No2 O2 sensor 4.4ohms

either No1 or No2 is a Bosch as i have the box but they are not marked themselves..

No2 was the elanscan earlier this morning .

Chas54

[chrism]

Chas 54 sounds like there's nothing wrong with those heater circuit resistances if you measured those ohms using both heater wires (white I think...) within the black m/plug terminals at the end of the sensor? If they were 20ohms+ then I'd suspect.

No.2 was probably the Bosch since mine was 4ohms...esp if it has the same fluted end as the Walker. The denso has the non-fluted holed-cylinder-end design.

So I suppose this next...i could try removing the ignition to O2 heater plug +ve supply and replace it for a test with a directly cabled supply to the battey +ve . Applied when engine is running ..

I'm not sure I'd bother to measure the O2 feedback voltage in the first instance since an EScan's 'Oxygen sensor Volts' graph will tell all - just wack your new Walker back in with the revised heater supply and do a 5min (say) log to see if the problem's still there by looking at the resulting graph...

Your Walker looks good and your Bosch looks okay to me, The pink trace must be no.1(?) cos it looks a bit sickly. It looks to me like its higher resistance might be causing it to struggle, perhaps.

[RonR]

Following on from a previous observation:
Even with the new sensor, the voltage drops to 0 and hovers around there for 10 seconds before jumping back into life.

Walker last 150 seconds.jpg

Could we be looking at a stray strand of wire on a pin / cable occasionally earthing itself? Maybe even inside the ECU?

[chrism]

Hmm...if that's so it was the scan taken earlier that day in 'after check ECU connection S2.ecu' with whatever sensor that was, presumably the Bosch at ~1500s, and before?.

And superimposing the blue later-in-the-day Walker scan the same thing might have happened at 700s(?) after that initial struggle to light up:-



Interestingly both these sensors started up at 390mV rather than 450mV today...

Let's see how the heater +ve substitution goes...fingers crossed...

[chrism]

Something else I've noticed: If you stretch out the Walker sensor trace (below) and hold it close to the Bosch trace (above)...



...the 'downtimes' seem to coincide and the times these occur for the newer Walker are as-near-as-dammit halved. The overall shapes seem unusually close too.

That would really have to be a coincidence wouldn't it? Could there be an explanation?

[Tuga2112]

Could we be looking at a stray strand of wire on a pin / cable occasionally earthing itself? Maybe even inside the ECU?

my understadnign is that the flaw has a rather predictable timing. something like that would be re-ocurring every now and then instead of being a single failure

[Chas54]

Ive tried a direct feed to the +ve side of the O2 sensor plate heater but as the engine was idling well already and voltage was there it didn’t really prove much. But whilst messing with the plug i caused a voltage drop . Ive half convinced myself now its the on board female socket that is a problem. At the moment ive cleaned the inner terminals with a round magneto file and cleaned out with an electrical cleaning agent. Tweaked the connection terminal tubes to make a tighter grip It seemed slightly better but then the voltage still dipped . Noticable with a thin wire slipped up the heater +ve terminal on the new sensors plug. ie it may occasionally be losing voltage still.

I think i will change the socket on the car but cant locate a supplier of the original type yet. I could use a different type plug and socket but then have to change the sensor one as well. If it comes to that i may use a cheese block in the short term as it will make for greater confidence when connecting meters to the terminals ..

Incidentally , my Fluke meter has fused probe cables and one fuse has its own resistance whist showing no sign damage and still ringing through OK. Rated at 0.5 amp ive replaced it with something that fits.
Its more than 20years out of its annual conformity test and i only use it on low voltage stuff nowadays.

i am attaching a couple of elanscan .
One cold start and idling during my messing about with the sensor supply and plug .

More telling is the cad park one , the car played up with the fault light comming on. I stopped in a layby and the sensor had lost its 12v +ve supply when i inserted my thin wire into the sensor male socket (up the side of the wires seal ) . I tweaked the connection there and it seemed OK but then further along the road brought the light up again .

Back home Ive managed to re create the voltage loss by squeezing the plug which is why i'm thing of changing the socket.

Still have this phenomena where the sensor doesn't produce anything worthwhile until coolant at 82c

If ive got an errant wire strand somewhere I can see it will be hard to detect .

cad park out.zip

Chas54

[Chas54]

socket tweek.zip

socket adjustment

[chrism]

The socket you're gonna change, is that the female at the wing/engine bay that the O2 sensor lead plugs into?

I'm wondering if a breaker can provide that plug chopped with a bit of lead on it so's all you would have to do is splice it in?

Your scans are showing that you're still losing closed loop when the O2 sensor drops. Other than that nothing else making sense that I can see...

[Tuga2112]

if i was in your position i would be thinking of de-pinning the connector and adding new pins.

you need to find out what is the stnadard used in the car (i believe Ron has a record of the known ones. and theres a post in the forum)

i have boxes like this in my garage to do that kind of fixes.

i ended up with that stuff when i tried to do the O2 sensor wiring "properly" but back then i got frustrated with being unable to de-pin the harness at the wheel arch and just added a new wire leaving the depinning task for later (i think i even pointed that out in the video)

https://www.ebay.co.uk/itm/366568293214 ... R-L6oaWIaA

[Chas54]

Tried tightening all the terminals in the O2 sensor socket a bit more and the results seemed OK until i got on the road.
Still get a feel of slight hesitation when dropping speed to 30mph and the ECU light comes on again. Warmed up on driveway and then short run. returned with the ECU light on and it cleared when i got onto my driveway.. At the end of the run i tried manipulating the plug but couldn't produce a voltage drop . I unplugged it the plugged it back in.

Looking at the elanscan trace, again the sensor keeps stopping as no trace is produced. Then starts again .

So ive ruled out the sensor and I am going to try de pinning the sensor and onboard plugs and creating an external heater power source direct from the battery to the sensor, also a separate feed back cable to the ECU D7 pin.. Ill probable include the earth cable to a common ground . This i hope will determine if its the ECU fooling me .

Also as im getting lower on petrol , i may syphon out the tank and refill with some fresh E5. I cant rule out Tesco having sold petrol with water in..

I'm getting a bit stuck for ideas after this though.

Thanks for all suggestions
Chas54

PL to cad.zip