Lean combustion?

[chrism]

Chas54 a preliminary look from your latest graphs...

1. your spark advance was ok on warm up , went silly on the run - but to be expected? (I have to admit I've never been interested enough to take a mobile EScan...)
2. your O2 sensor took ~240s to start firing , mine's there in ~50s and yours is still a bit erratic imo. Yours stopped working entirely as you got going and accelerated...more about this later
3. your a/f ratio got weaker as the O2 sensor stopped when you were driving and especially as you accelerated more - to be expected again?

As Ken noted before, IAC posn was okay as long as the O2 sensor was 'going' but went silly again when you became mobile; could it be 'stuck' or could a hose be leaking as you accelerate, weakening the mixture?

Your O2 sensor heater circuit resistance readings of 206ohms and 260ohms bother me. I've measured 3 sensors - Denso at 13ohms, Bosch at 4ohms and Walker at 3ohms
Using I=V/R then P =I*2xR this gives heater wattages of ~13W for Denso, 36W for Bosch and 48W for Walker. (I kept the Denso in the car...)

Using the same maths your heater wattages were 0.74W at 206ohms and 0.65W at 260ohms (a sidelight bulb is ~5W...) - would that be enough wattage to spark up your O2 sensor from cold? (doubt it?)

Sooo...with respect...are your resistance measurements right or are they showing that both your sensors are 'fecked'?

Or have you got a dodgy IAC/air connection? (as well or instead?)

Or could it be something else that we haven't even noticed yet?
(Ken + Ron and Matt are the current EScan gurus... )

[Chas54]

Thanks Chris
I don't know how I .managed to mess up the resistance, I should have spotted that was to high.
I've dashed out to the one I had in the sonic bath that is a Bosch item. Measuring again at 15 ohms.
I'll have to do the one I just put in the car tomorrow.
Thanks for the info.

Chas 54

[chrism]

Yep, that's more like it...

[KenS]

Getting tricky! Is it agreed that the O2 sensor is looking as if it is telling the truth, although it is slow in starting up/warming up?
I guess the measurement of the heater voltage was at the O2 sensors 2 pins for both the +12V and its return so that there is no doubt that the heater is passing current.
I did look more at the injector pulse widths on this latest plot, especially just before the 1200 second point, the revs increase with the throttle position, the pulse widths briefly increase then fall below their idle norm and then it drops out of closed loop and the revs also drop. I'll have a look at the Closed loop manual and see if there are any notes about what can result in drop-out of closed loop. Wonder if it would help in trying to get the same plot from my car, run at idle till warm then give some throttle. I'll try to find time tomorrow since it might help understanding this.

in-work1.jpg

[Chas54]

Thanks that interesting.
I swapped the other sensor back yesterday evening and the raised to idle cold with my voltmeter across the O2 heater circuit terminals. I found a bit of stuff wire to poke in the cable seals in the plug . Good voltage 12+ not any sign of fluctuation . Tried moving off my driveway and fault appeared again near a bump on road . No change in voltage.

I went back on the driveway and put the wire prod across the earth and 02 feedback to ECU cable. The voltage corresponding to the scanned voltage on elanscan until I passed over the same piece of bumpy road near my house, voltages dropped to zero and ECU fault light operated.
Returned to driveway and cleaned the female terminal connections on the loom plug a bit more vigorously with a round magneto file. Tried again and no fault this time or loss of feedback voltage.

A short run afterwards did produce the fault light again for a short period.
Returning home the car was running better but as there is still a glitch somewhere, I think I'll take the inner wing out again and try a little more abrasive terminal clean up. I. Will also check the idle air valve but I renewed it a couple of years ago.
I've taken to not clearing the faults now because I'm concerned I keep clearing the BLM and that might be adding to my problem.

I'll attach an elanscan file of my yesterday evenings event for perusal. I'm trying to work out wether to go for another run from cold now or set about my checking while it's cold enough to get my hands on the IAC valve.

[Chas54]

Road bump elanscan

tighten O2 clean and tighten sensor1 terminal.zip

[Chas54]

KenS

I have a feeling the point at 1200 secs was when i slowed down to turn right across oncoming traffic. I was in 5th gear just before then down to third at a very low speed close to stalling .
So i'm just wondering if it was me causing the issue you described with the injectors.

Regards
Chas54

[chrism]

And yet again a direct correlation between O2 sensor and a/f ratio in your latest scan!!...

But what could be causing the O2 sensor to keep dropping out if all your connections seem okay?

And why didn't exhaust heat keep the sensor 'swinging' this time? (or would it have done given a longer run-time?)

Methinks a good idea to wrangle those connectors behind the w/arch then another EScan over that bump. Perhaps if this throws up the same problem it'd be time to try a different sensor to finally exclude it from the issue?

ps Ken's idea for an EScan running it up from idle then raise the engine speed is brilliant I reckon. I'm gonna do as you did when the sun comes back and I can leave it outside overnight to get cold, then idle, then drive away as you seem to have done in order to see if I can replicate your dynamic data - which I've never done before (suspecting all the sensors would interfere with each other and give meaningless individual data...)

pps cable connections inside the car? lead? new ElanScan_beta-USB + protocol? (probably not in hindsight...)
perhaps the 'above the speedo' cable to the ecu or it's placement in the blue ecu loom?

[KenS]

I have carried on looking at 'idlng then switch to service.ecu' arounf that 1200 sec point where the engine was idling then reved to about 2800rpm at which point it dropped out of closed loop. I have run the same condition on my car and it stays with running O2 sensor and closed loop throughout.

Your plots where the O2 sensor stops as soon as the revs are increased, then since the O2 sensor has stopped, it drops out of closed loop:

plot1.jpg

And equivalent plots from my car, but with 2 periods of reving just to be sure!:

plot2.jpg

So it does seem odd that the O2 values die immediately as the revs increase, so far I cannot find cause of that. Dropping out of Closed loop soon after does make sense, since the ECU needs to see a changing O2 sensor value.
Extract from Closed Loop Service Manual:

manual.jpg

I'll carry on looking at your other plots, hopefully we can find something that make sense of this.

[KenS]

A quick thought - are all the earth cables connected at the drivers side back corner of the engine? Everything suggests that they are providing a common 0V to everything but its any easy one to cross off the list. Mine has 3 solder tag connections on one bolt to the plenum and one thick braid connected by a 2nd bolt to the cylinder head.

[RonR]

Just noticed your O2 sensor voltage doesn't go completely to zero very often but hovers around 10mV, suggesting a high resistance connection with occasional complete disconnections.

Here's what I would do:
Unplug the ECU connectors, check for corrosion, squirt a little WD40 in there and plug / unplug several times.
Remake the engine ground connection using Copper loaded grease.
If the problem persists the only place left to look is Main Harness connector MHC3 - somewhere behind the dash....

Edit - Just spotted Ken's post - he beat me to the earth connection!

[chrism]

For me, pleased that Ken has established that O2 sensors don't switch off when revved, that closed loop is maintained when the sensor is swinging okay and that a/f ratio stays constant when the sensor's active as well.

Also we've established that your sensor is dropping down (if not totally out as Ron's pointed out). It looks very much like a connection, perhaps corrosion, issue and seems probably between the w/arch and ecu (when you've managed to totally remove those w/arch connectors from the issue).

If not Ken + Ron's earth cables it might be the signal cable then (loose/corroded?). I believe yours is an SE->1499 conversion. This is how it's likely to be wired up according to the LEC conversion advice:-



0. Remove the glovebox, dash binnacle ("half moon" thing above the instruments) and driver's side front wheel arch liner. Pass the end of some suitable signal wire from the passenger footwell behind the radio and over the back of the instrument panel. now take a length of 3A or higher rated supply wire and tape it alongside the first wire. Pass the pair down to the right of the steering column and into the right end of the driver's footwell.
1. Take about 8" of thin steel wire, bend it double and push the bent end from the wheel arch through the bulkhead grommet alongside the main harness. From the driver's footwell, pass the two wires through the loop, then use the loop to draw the wires through the grommet into the wheel arch.
2. Splice the new wires into the RHF harness. The signal wire connects to the wire on RHFHC pin 8, and the supply wire connects to the wire on pin 4.
3. Disconnect the battery, return to the passenger footwell and unplug the ECU harness from the ECU. Connector pin D7 can be located using the markings on the connectors. The wire from this pin should be spliced to the footwell end of the signal wire.
4. Return to the open binnacle cover. The "IPHC" harness connectors are on the left of the opening. Inspect the large black connector, disconnect it and identify the wire on the IP harness side of pin A15. Splice the binnacle end of the supply cable in here.
Congratulations, you now have the necessary wiring! Check for good continuity from O2 sensor connector pin A to ECU harness connector pin D7, from O2 sensor connector pin B to IP harness connector pin A15 and from O2 sensor connector pin D to any earth point. Investigate and correct if necessary. Replace ECU


Chas54 does this help to 'wrangle' the signal wiring if you didn't wire it up yourself? (So sorry, I see from an earlier post that you wired it up yourself...) It's from the SE->closed loop instructions in Wikipedia (I think). Tuga's done a video of it too...

Also, if any IAC problem was causing the a/f weakness would/should the O2 sensor switch off as your latest EScan seems to imply? (Just asking...)

[Chas54]

The cabling to the O2 sensor already was in place when I carried out the conversion so I had no need for threading cables through etc. It looks like maybe I have a faulty cable or connection between the plug and the ECU.

I'll try the agitation and spray on the ECU plug , also a continuity test to see if the cable is breaking down somewhere, it's not the worst job in the world to replace the cable if warranted.

From memory, I seem to recall all the earths being on one terminal at the rear of the plenum, I will have to check these out though as it's about 5 years since I had the plenum off.

At that time I replaced the IAC valve and refaced the valve seat. There was a mark on the valve face and seat from wear I guessed.

I will check out and clean the earth connections as they are no suspect.

Thanks for all the advice guys , it's really appreciated.

Regards Chas54

[KenS]

Don't let this divert the route you are going, it just seems a good thread to ask the question - anyone know what a good O2 sensor plot would look like for a true lean mixture? It doesn't look fast enough to show pulse by pulse variations unless the plots are aliasing, if one injector failed shut would that be identifiable in the plot as 1 in 4 pulses missing (don't think so) or would the trace be attenuated or go flat near 50mV? I am tempted to get a plot of idling and pull one of the injector plugs and see what it shows, but worried about loosing the injector connector clip - maybe if I have a spare 1/2 hour!

[chrism]

Ken, does this help? I can send the whole scan if so.
A pal's S2 had O2 sensor and temp sensor probs. He's now running okay but data is dropping out in a lump like Chas54's
I have other scans of his where 'O2 sensor lean' was flagged up but 14.7 was also claimed tho'

[Chas54]

Chrism

On your friends scan, does the air fuel ratio drop?

I've noticed on mine that the fuel air ratio is at the stoichiometric perfect of 14.7 to 1 but drops slightly during an event to 13.9 to 1 which is actually going richer not leaner. Would this suggest an ignition issue??

I have checked through the ignition system , coils are newish but I swapped them round to test for any alterations. Cleaned the plate and earth's and the insulation of the ignition module.

Incidentally, I don't have a plug socket under the wheel arch, the O2 sensor is cabled directly into the loom. I'm guessing the plug is the other side of the bulk head.
At the moment I'm re-attaching the inner wing liner securing fittings as some are barely functional..

Earths next then the drivers foot well and binacle to find the next plug to check

[Chas54]

I think the only way to get at the earths with any satisfaction is by removing the power steering header tank.
I can see all the earths described , one at the cylinder head by the 2ndry throttle body. Three others attached at the plenum and there is a bradded earth below the master cylinder on the top chassis cross member.

I have taken the binacle cover off to look for more O2 associated connection blocks but its a bit of a muddle .. I am actually thinking i can prove if the O2 sensor feed back to the ECU is good by using a jumper lead from the O2 sensor feedback plug terminal and the ECU plug D7 pin . Draped over the windscreen initially for testing . This would prove if the signal is lost due to fault in wiring or if its still there just not registered.

But thats for tomorrow now , I cant really have both doors open in the garage so it has to be outside on the driveway and we are getting some rain.

[chrism]

Now 17/08/26

Before 01/03/26


His A/f ratio doesn't drop now when the O2 sensor does (and I've verified it with the cursor) although (as you can see) he drops the O2 sensor out in a big way in a scan this week. He's yet to do another very soon to verify/compare.

But as the 2nd graph shows a/f ratio did drop at the origin of his troubles in March '26. At this time he had to renew a broken coolant temperature sensor (apparently top broke off - sensor should've been okay, I think) then we found out that his oxygen sensor was 'duff' too...I think both were 'dodgy' at the time (but you only find out afterwards, don't you? )

Perhaps when the O2 sensor struggled and the CTS wasn't hooked up the ecu was overfuelling to compensate which resulted in the lower graph?

So, he's renewed his CTS too; I wonder if that's significant? I then checked your coolant temps to see if there was a prob. Your 17/08/26 scan shows as normal but your 18/08 and 19/08 scans seem to start mid-way up the temperature scale so probably all okay here?

Another cold scan would verify...

Yes those w/arch fixings are a waste of space ...

[KenS]

Chas54 - Sounds like the earth connections are in place and almost certainly OK which is good.
I have run a test on my car as a reference, I disconnected Fuel injector #2 while idling. The O2 sensor voltage dropped flat to about 50mV for about 4 seconds (!) then recovered swinging between 100-900mV. I guess it initially sees a weak mixture, then demands a higher injector pulse width/duty cycle to get it back to ideal. I then reved to 3000rpm and again the O2 sensor dropped for a couple of seconds and recovered. Throughout it stayed in closed loop and Air/Fuel ratio was solid at 14.7 (but I think RonR says this is a target value when in closed loop, not a measured value). Unluckily the monitoring PC went to sleep just after that giving a static trace until it woke-up, but that dosn't matter and can be ignored. Also the O2 sensor does/should respond to every pulse of exhaust gas, so at around 1000rpm we should see 33 oscillations per second on the O2 Sensor Voltage (think that correct!), but due to the sampling of ElanScan we don't see that detail and aliasing displays it as a slower frequency, but good enough to know when its working. That explains why when running on 3 injectors the O2 trace is a bit more irregular.

pulling injector-sml.jpg

Zoomed in traces for Closed loop/Air-Fuel ratio/Injector pulse width/pulse duty cycle/O2 Sensor voltage and the full .ecu files are attached as a zip file for future reference.

So if your connections all turn out to be OK then next likely is that both your O2 sensors are suspect? Have you washed/cleaned both? It's hard to get an authoritive answer but the sensor can be contaminated and affect the rate at which its output can change for a given O2 level change. Google recons petrol should not be used to clean them. From my run today it seems that a fueling issue would have to be quite severe to throw it out of closed loop, I don't think one erratic injector would be enough.

Chris - I guess we shouldn't go too far with you Friends traces in this post or it might get confusing, but I think the recent plot (17-8-26) sort of looks OK, it implies it's in closed loop (a solid 14.7 Air/Fuel ratio), and the O2 sensor voltage is dynamic 100mV-900mV, I think the dead bit from 220-1200 seconds is where Elanscan has stopped logging, maybe PC gone to sleep or comms lost? The old plot does show something quite wrong with the O2 sensor since its voltage is only up to 60 or 70mV and the Air/Fuel ratio is 14.5 implying its in open loop - because of the failed O2 sensor or too cold.

Wow these issues do keep the minds of us oldies ticking over! They always poke the corners that we haven't quite been down before!

[chrism]

So...because your O2 sensor stayed in closed loop you've effectively proven that poor fuelling can't have caused the sensor faults, but rather the O2 sensor's caused the fuelling faults then Ken? That's very noble of you to interfere with your own car to get traces to prove it

And who'd've thought that 2 sensors on the trot would be 'duff'?

I'm not sure Cha54's cleaning would need to be necessarily blamed; I absolutely scrubbed my current (in-the-car-and-working-well) Denso when I thought it was broken (it turned out to be the wheelarch loom... ) with carb cleaner then soaked it overnight in white spirit following up with a toothbrush the next afternoon. My thinking was that it had failed anyway so I could afford to see if it could be cleaned up (and I had a shiny new Walker one from RAuto). I put it back in when I realised that its heater resistance of 13ohms gave it about 13W of heater energy which might lighten the load on the alternator that wasn't designed to run an O2 sensor, being a '90 SE. Also I wanted to see if the reduced wattage would make the warm-up slower and found that it took about a minute longer (90s compared to 30s) to start 'wiggling' - so I left it in!!! :-

- Denso is black, Walker is blue

Yes now you mention it I get a strong feeling that my pal's last trace was due to leaving the l/top to 'time out' on the car seat whilst he polished something shiny instead but glad his latest trace showed that we 'sussed' it

ps Chas54 I think you've got to try another sensor now...?