Lean combustion?

[KenS]

Checking through several of the last week or so plots also shows the point where the initial O2 sensor voltage collapses (at 35degC coolant temp) is consistant with the idle being slowly reduced and crossing the 1200rpm point and that point also corresponds to the IAC being in a very similar position around 42-46 counts (but this is where the ECU believes it has placed it, maybe not actual). I suppose the IAC is programmed to slowly walk down in relation to the coolant temperature, hence all 3, coolant temp, rpm and IAC position are all tightly linked. Is there something special about 1200rpm? Could the IAC be skipping counts around count 44? But cannot think of a reason for the mixture to go lean, if anything would think the revs would change or hunt.

rpm-iac.jpg

[peteoddcar]

Chas
To answer the screening question
Screens between engine block and chassis, ( we have no body as such so that leads to confusion constantly!) are only grounded at one end otherwise they become conductive paths between chassis and engine block and will actually induce a voltage to the signal at crank that’s being protected by the screen.

The whole point of the complex coil pack grounding arrangement is that the coils were designed to be mounted on the engine block.
Hoping that the chip swap does the job - you seem to have eliminated any hardware and system problems.
What bugs me is why did it suddenly go wrong at the MOT
Did they maybe plug it into a diagnostic tool?

[dapinky]

What bugs me is why did it suddenly go wrong at the MOT
Did they maybe plug it into a diagnostic tool?

Extrapolating from that thought..... the original PROM chips (used by Lotus, but now NLA - or at least, only cheap Oriental versions of the chip which have, shall we say, "mixed" reviews) are the type that can be re-written with a different PROM set-up....... so, someone with the right kit can burn whatever values they want into the old chip to give it whatever parameters you want.... more basic versions of the 'same' chip are single-write not re-writable.

Basically, this is how Alan did the original Mountain chips, but Ron has gone with a more modern (available!) chip set.

My thoughts are that at some point, the PROM values may have been affected electronically (if the MOT garage has plugged it into something?????).

Is there someone local who would be prepared to do a temporary swap of ECUs to see if the problem moves with the ECU or stays with the car?

[KenS]

Been looking at that 35degC point in the plots. I think it must be a point in the engine warm-up period where the ECU changes its phase of operation and is common to all cars. this change causes a minor hiccup in the normally smooth running of injector pulse widths etc, but normally it is so minor it is of no significance (see lower plot). Plotting the 'O2 Sensor (integrator)' for Chas54 plot and a reference plot, shows the integrator in Chas54 Plot drifts off and looks like it saturates, unluckily at that 35 degC point, whereas the reference plot shows its integator is still in its operating range and survives the minor upset at 35degC. So is this showing the integrator in Chas54 ECU is suspect (leaky capacitor?) ? Any comments or knowledge on the significance of the O2 integrator trace?

integrator small.jpg

[Chas54]

KenS

Ive just been looking for significance of the O2 integrator figure and it appears that 128 counts is the 0% point and above that fuel should be added and bellow that fuel should be dropping off.
I think this must control the injector band width . Or at least on a Scanner Dan video I just watched , this was what i think he said . It would make sense as the fault i get is a lean sensor .The point you have highlighted then is a lean period and it should be corrected by the injector band width increasing. Ive got to go back to a previous , on the road scan to see if this is common to my my fault occuring.

Ill have a look when i get home tomorrow . It definitely looks like another clue. Thanks

Chas54

[KenS]

That would make sense, if the integrator value was used to proportion the fuel injector pulse width, with the ramping up of the integrator increasing the fuel volume to prevent the 'lean' condition. But it goes odd again when looking at that injector pulse width, the ECU is reducing the pulse width instead of increasing it even prior to the integrator hitting that upper limit !

integvswidth.jpg

Could this mean that instead of needing to increase the fuel it tries to decrease the air to get the air/fuel ratio correct correct? That could lead to the IAC position in doubt ? Not sure how to split this problem, these closed loop feed back systems get too tricky!

[chrism]

Still perusing...I noticed a difference in injector pulse width vs O2 sensor volts between a Chas scan (above) and a recent scan of my own - healthy, I think - m100 (below) and I'm wondering if the Chas lower pulse width might be significant?

[KenS]

Yes, I think the Chas injector pulse widths do look a bit unusual.
My thoughts which are more for discussion since I may not have a true picture of how the ECU behaves and am learning as we go along!
The Chas plot initially has injector pulse widths around 3ms, reference plots are nearer 2 to 2.5. Then when the Chas O2 sensor voltage dies (or goes lean at 140sec) the pulse width has dropped slightly but still relatively high, still not far off 3ms. If it has gone lean, the ECU should have increased the pulse width even higher but it hasn't. The O2 Sensor Integrator implies the the ECU knows the mixture has progressively got leaner, but has not increased the injector pulse width. Maybe the ECU has a limit on the max pulse width allowed for the conditions at that time in the warm-up.
So if the ECU is commanding a relatively high pulse width but still not getting enough fuel in, I think it must mean that either the fuel pressure is too low or multiple fuel injectors are not opening as expected or there is too much air getting in (air leak or IAC). So still lots of possible contenders! We know normally the system has enough authority to achieve good working O2 sensor voltages with one completely dead fuel injector, so a fuel injector issue seems low on the list. Chas has run checks on the fuel pressure/delivery with nothing showing. Also, at idle I wouldn't think the fuel demand should be too high. That leaves too much air. How does the ECU monitor the volume of air? The MAT and MAP data plots do not look unusual. Chas has run smoke test with no evidence of a leak. Leaves the IAC ? But I cannot think of a plot check or test that could help this along. I guess a failed ECU is also a contender, but since the plots seem to have some clues in them I think that is low on the list. Again any comments or discussion.

[chrism]

Also a comparison of injector pulse width vs IAC position:-

shows a difference. Chas' in blue and mine in black...

[Chas54]

Although i did my smoke test , i am not sure i haven't got air passing the the throttle butterfly plate when its shut . If i block the IAC port i still see some swirling smoke at the plate edges. Combining this evidence with the fact that ive closed the idle air bypass screw all the way in again when checking in service mode to get a 950 rpm idle when hot. In service mode the IAC sits at 20 counts . I am wondering if i just have too much air passing the plate when i lift off the throttle although my event occurs at other throttle openings. Did this yesterday and today ive been on a slightly longer run . I am having to rely on the Lotus for transport at the moment as the modern hybrid has gone into the dealer for a software update. I cant really take to many bits apart at the moment.

It had a few ECU light on events , one when approaching a roundabout slowing down , the engine stopped . I switched off and restarted straight away and the light was off and stayed off, the engine running smoothly.

Still my mysterious loss of O2 feedback voltage is still there and my next task is replacing the wire . The short end of wire at the ECU plug putting me off today .

Thanks for all your suggestions and advice .

Chas54

darke.zip

[chrism]

Well Chas54 from your latest 'darko' scan today you must be sorting some of your problems cos now your O2 sensor voltage is showing closed loop whilst it's swinging, and dropping to open loop when it's not. I think this is different?



Also the heater voltage is starting the process within the exhaust heat level startup of ~600s and it's not dropping out so hugely

I suspect your CEL light might have flashed on during those dropouts - but we'll never know for certain.

Getting there...I'll await your signal wire replacement tho'...

[Chas54]

After much contortion and an extensive search for some soldering flux, ive managed to place a temporary wire from the ECU D7 pin to the O2 sensor socket(wing side) . Its external at the moment across the dashboard and out the door then under the bpnnet. soldered at both ends for good joint. If this proves effective i will remake the joints and thread the cable through .

Ive looked under the wheel arch liner again and stripped some tape but cant find a plug anywhere on the O2 sensor cable. Ive retimed the ignition during my idling run today and opened the IAC bypass screw so its now about half a turn open.

The engine ran badly at initial start up and i think this is due to the memory being cleared by my battery disconnection for my ECU plug work.Eventually got to 82c and idling reasonably so i did the timing and IAC set up once warm enough.

The idling elanscan trace of the O2 sensor output looks OK but of course the proof will be out on the road - hopefully tomorrow if its not raining

I had concerns previously about two wires on the J2 plug that are just jumpers across 2 pins but they are on the EMJ4 drawing of the ECM connections as "peak & hold" jumpers pins D8 to D13 and pins D9 to D3 .

Thanks for the encouragement and advice

Chas54

temp O2 sensor feedback wire,.zip

[peteoddcar]

Chas
I absolutely admire your tenacity - the routing should eliminate any EMC issues

[chrism]

sorry...hate to say it but at first it looks like your temp wire scan indicates a step backwards.

Now it looks like the O2 sensor is only being fired up by the exhaust heat again...

However your O2 sensor, despite not being fired up by its own heater, is not saying 'lean' for the first time(!) and a:f ratio is nicely on 14.7 throughout apart from one small hiccup.

I've got a feeling that it ought to run fine now - but there's still the O2 heater supply problem - which might not be a problem after ~10mins of warm up...let's see?

[Chas54]

Ive tried a run out on the road today with my external temporary O2 sensor cable on .
I didnt wait for it to warm up just straight out to the garage to put some petrol as the low petrol light was on. E5 of course.

Ran badly but i was expecting it as i still hadnt done the "over 30mph at steady throttle" to reset the IAC and learned memory. I think the engine stopped and i had to restart it

Once on the road and warmed up it did feel like the car was running better but on the way home the ECU light came up , i recorded a lean sensor fault.

Back home onto the driveway and stationary the idle was good but at about 2022 to 2040 the light came up again and i felt the revs drop off. Looking at the elanscan trace shows the sensor output dropped off.
Time 2027 secs, High O2 sensor voltage The injector pulse width shortens to reduce fuel , but then i can see a slight increase in manifold pressure which would pick up the fuel pressure to maintain the injector diff . The O2 volts rise indicating a lean mixture and the voltage then drops to virtually zero, rich again . It looks to me like the injector bandwidth is doing as it should (but not all the time) .. I am wondering if the periods of O2 sensor showing very low volts are true but the system is just giving it to much fuel and not responding quickly enough . Ive got no idea why ..
I haven't run the car on the road with the earth and +ve 12v supply jump leads directly onto the O2 sensor . I think ill try this so all 3 terminals and cables at the O2 sensor are guaranteed sound and under no influence of any other interference. If i still get the O2 sensor voltage low drop outs, i think ill have to start looking at fueling.
Although I still have the old open loop memcal to put back in the ECU to try.
I think the o2 sensor has its +ve supply from the alternator output possibly to ensure the engine is running and prevent the plate from overheating if no gas is passing it. I'll have to warm the engine first then swap to my jump lead and have a quick method of disconnecting ..

Again thanks for all the suggestions and sage advice.

Chas54

temp o2 road run.zip

[KenS]

Chas54 - I agree that it looks the O2 sensor is reporting correctly and the drop outs are true lean conditions. A contender in the air/fuel calculations is the MAT sensor. Although its values look sensible I wonder if it is sluggish or has a small but significant offset. Its a relatively easy one to take it out and try cleaning, compared to the other investigations. At least that would cross off another sensor.

[Chas54]

KenS

I will give it a go I think. I had considered it but as it wasn't showing anything sudden I just assumed it's OK. But as you point out, it's an easy one to clean up and possibly eliminate.

Thanks
Chas54

[Chas54]

MAT sensor cleaned and now all three O2 sensor cables are on jumper cables from under the bonnet by the drivers side hinge and through the door then over thedashboard .

I have two jump starting cables attached to the battery and threaded through behind the passenger seat and into the foot well . The _+ve croc clip on the seat for ease of disconection when engine stops.

They are soldered at the wing socket cable socket (wing side) the originals are cut and tapped away.

Immediately i started the engine i attached the +ve cable , the other two being already connected .

Noticeably the engine was running better and no throttle blipping to keep it going. I let it warm a little then straight out onto the road . Every time i cast a look at the sensor output it was clearly working. No ECU came up until just before arriving home. Whilst out i stopped at a road works and then in a lay by. Still running well and indicating a voltage.

When i arrived home with the ECU light on the sensor was indicating zero voltage and the idle not flawless. At about 1530 to 1550 secs , i blipped the throttle and on the revs dropping the manifold vacuum improved from 0.42bar to 0.36bar and the O2 sensor started to produce a voltage .

On the road it had felt so much better i almost believed id done something to improve the running. But on inspection of the elanscan trace , i can still see places where the O2 sensor stops producing a voltage or runs (or runs very rich)

This is a bit at odds with my initial thought that it might be the throttle plate passing , so when i blip the throttle it shut to a better fit , which explains the vacuum improving but prior to that it should have been running lean not rich.

Another thought is today’s improvement may actually be the better earth. I believe the O2 sensor earth is via the ECU J2 plug D6 pin. On the manual drawing it shows it being common the the inlet manifold earth and i'm considering running my earth from the O2 sensor direct to this manifold earth .

Ive attached todays elanscan so would welcome any comments or thoughts .

Chas54

O2 sensor on 3 external jumpers.zip

[Tuga2112]

your issue is certainly very strange in nature.

and apologies if you already done this. but i am not keeping up with the thread.

i am wondering if the O2 sensort plug (at the car side) itself has one of its pins not crimped properly or the copper got hold of moisture and corroded.

de and re-pinning those connectors is not the end of the world. i have done it on the MR2 just a couple weeks ago on exactly the o2 sensor.

[KenS]

A lot of the old features are still there in the plots, I don't think much has changed. The O2 sensor still drops out as the coolant reaches 35degC and after that when its 'working' it generally operates in the range of 0V to 800mV instead of 100mV to 900mV. It still reverts (randomly?) back to open loop when driving but slightly less prevalent (I think its actually limp mode because 'MALF RMP Limited' is set). I think the O2 Sensor is giving correct data, it does show 800mV (rich) when throttle/load is increased ( at 646sec), but it does look a bit erratic. Those couple of 'rich' blips are also shown in the Air/fuel ratio plots, even though normally the closed loop value remains solid at 14.7 (so is it a target value when in closed loop?)

21-9-26.jpg

21-9-26c.jpg

Then when the throttle was blipped at 1537 sec, again the O2 sensor did respond, showing going from lean (0V) to rich (900mV) briefly.

21-9-26b.jpg

There must be more clues in these plots, there is a lot of data there. I'll go and have another poke around the plots! What we want is an M100 ECU Simulator/model so we could run fault conditions and see their exact effect!