Lean combustion?

[chrism]

Chas54 results not okay til you got on the road imo.

The lhs shows your current scan plotted against distance and I think it shows the sensor drops out well before you're mobile...

Your latest trace overlaid with your 'cad park out.ecu' shows that your sensor seems to predictably start up, break down then start up again after ~600s presumably when the exhaust triggers it.

But I get a gut-feeling something other than a loose pin/connection is rearing it's head to cause it to switch off the moment it gets going. Can't think what tho'...definititely not the sensor.

Yep substituting all those wires is the next step...if nothing then it'd have to be ecu. Could it be the 1499 memcal addition? I wonder if it would it be worth trying an EScan with the previous promid - even before the wire sub?

[KenS]

I agree, it looks the O2 sensor is slow to warm up, but throughout that period 200sec to 600sec it is in closed loop, that doesn't seem to make sense, how can it not be receiving O2 sensor values but stay in closed loop? Maybe as Chris suggests, there is a parameter that is holding off that O2Sensor but allowing it to stay in closed loop, maybe it is a true lean mixture, but I still would have thought it would have come out of closed loop. Then later, when driving, if the O2 Sensor data stays low for a period, it does cause it to drop out of closed loop which seems more valid. But even these don't look correct in terms of time to drop out and come back in and why doesn't the Lean flag get set if it really is lean? Looking at the plots in detail they are lacking in some resolution. If you click on 'cursor' then slide cursor across the waveform and watching the 'time' in the far left window, it increments at best at 0.2sec, but at times its as long as a second so this is why the plots look a little ragged and maybe missing some of the O2 voltage pulses? Plot below compares your latest 'PL to cad.ecu' with one of mine (increments at 0.1sec, occasionally may miss one). I think this may be due to the ElanScan interface, especially if its an FTDI type and very reliant on the USB port settings.

compare.jpg

If you think its getting to the point of needing close scrutiny of plots I can send my ElanScan interface (Bluetooth) to see if it makes any difference, if it doesn't give 0.1sec increments it might suggest the ECU is struggling to putting out the data fast enough? If it does give better resolution data it may help to see correlation somewhere.

[Chas54]

Chrism

Thanks for looking again .

I think it was idling well before i set off , at about 489secs idling the O2 sensor stared working . Somewhere round about 800 seconds it stopped working the stated working again .


I changed to closed loop after buying the complete ECM from Bell and Colvill for £160 in April 2021. I still have the old ECM complete but to use it I would have to go back to the old CO pot. And all the hassle of setting that up.

With the failure of the O2 sensor to produce a voltage at some point , i keep thinking it somehow has got to be related to temperature although sometimes the sensor is producing a low voltage at 50mph .
It looks a lot like 82.3c is too cold for the sensor even though the exhaust gas is hotter .. I know the reaction on the O2 sensor heated plate operates at around 300c.

I am in a mind to run the car hotter and see if i still chill the O2 sensor when dropping speed . The engine runs around 82c mark at the moment and this is the thermostat setting. I can insert my radiator grill cardboard blank and run warmer and then do an elanscan . It would at least prove if im clutching at straws.

I tend to agree with your suspicion that its not just a loose terminal as the predicability of when it occurs would deny that. I cant think its alternator output voltage as its 13.5 and really steady.

You could be right and its actually the ECM but its out of my expertise how to test it.

Kens . Thanks for your offer but i am not sure on using a bluetooth device . My elanscan is using a homemade FDTI type through a USB interface i do remember having to adjust port setting to get it to work . A gentleman on "you tube" showed how to make one and set it to work, my only mod was to insert a dapinky switch so i could change to service mode whilst still logging and viewing the elanscan live figures .

Thanks again for your valuable thoughts

Chas 54

[KenS]

Is it worth trying a jump cable from the body of O2 Sensor to a good chassis earth? Could the earth path through the manifold to block be engine temperature dependant?

[chrism]

Ken's jump cable earth is a great idea and easy to do. The sensor heater has an earth return tho' - which you've tested to be working - otherwise the resistance wouldn't be correct...does the sensor body have to be earthed too? Anyway easy enough to try. If that works could a fix be as simple as cleaning up the turbo plug threads? (could it be that the simple heat expansion of the plug could be Ken's 'temperature dependency'? )

Perhaps a check that the loom earth that plugs into the black sensor m/plug connector is okay by m/metering between that and the bonnet earth strap on the driver's side?. I seem to remember checking that, too, when I had 'troubles'.

If you can get a 1499 chip into your old ecu then that might be another easy test to see if you can get a decent O2 volts trace and eliminate your current ecu from the picture? (I've got a spare but I think I'd prefer to fit it myself, having done it a few times...and you probably not...)

Can't think of much else atm...

[dapinky]

I changed to closed loop after buying the complete ECM from Bell and Colvill for £160 in April 2021. I still have the old ECM complete but to use it I would have to go back to the old CO pot. And all the hassle of setting that up.

2 thoughts on this (or possibly 3?)......

If you didn't actually remove the CO sensor, then it will still be in the same place that it was - so may help.

The only actual difference between the old and the current ECM is the Memcal chip - so maybe a bit of 'Mix and Match' may help to establish if the issue is within the Memcal, the ECM (plug/wiring/internals) or elsewhere.

I think that I'd fit the old ECM in its entirety as a first approach, it may not run right, but it will give you a 'baseline' scan to work with....

Then open up the unit and swap the 'new' Memcal unit into the old ECU - it may be that once you open up the ECU the actual Memcal chip may not be fully seated, or has a duff connection, and simply swapping it across solves the issues.... after that, try the new ECU with the 'old' chip to compare against the "new baseline".

I know it is a lot of faffing about, but it will help to narrow down possible failure points further.

[Chas54]

These all sound good suggestions.

I was thinking the ECMs were different but if its only the chip, I could easily swap the "newer" chip into the original ECM. Im not sure I still have the CO pot. I think i took it off and put it somewhere safe and then moved house since . I didnt leave it in the mount .

Ive ordered a warmer thermostat at 88c due to my feeling the problem is when the coolant temp drops and its around the 82c mark.. I can idle in the garage and without the air going through the radiator the O2 sensor eventually jumps into life at around this 82/83c mark. Still its strange the sensor starts to produce a voltage just after starting the goes off until it reaches 82c.

Im also changing the petrol , ive got other old cars that will accept any old petrol and any water may be visible in it. Ill refill with some fresh E5.

Im also going to try the direct connection to earth with a large crocodile clip round the bottom of the O2 sensor and to an sound engine earth ( may try back to the battery if that doesn't show a result)

Weather has turned a bit now though so i may have to report results a bit later in the week -- unless something miraculous occurs.

thanks again guys for all the wise suggestions

Chas54

[chrism]

Totally agree with dapinky - and not so much faffing about as removing wheelarch splash-panels, splicing wires, digging into connector pins etc...

Plus I never removed my old CO pot when I went closed loop. My (very green and nervous) thinking was that the idea of piggy-backing the old memcal with Rons 1499 gave me a 'fall back' if the 1499 misbehaves - not that it ever has .

[Chas54]

Ive tried the earth directly to the bottom of the O2 sensor and to the battery (cold start elanscan attached ) just idling though.
Had it been the sensor earth i think i wouldn't see the missing O2 voltage . But still not there until after 82c temp.

Ive swapped the 1499 memcal chip into my old ECU unit . The memcal unit i took out was dated 09/05 and was marked as 9930 .

I messed up my start from cold as the engine running was all over the place , then realised not having swapped the retaining bracket 10mm set screws had left the ECU case with a very poor earth..
Clamped it up and the running improved and did an elanscan but from warm.. (attached)

Ive delayed taking it on the road as the weather has changed . Also wanted to do a scanned start from cold .

Having the self inflicted problem with the ECU earthing has made me wonder if the whole problem is caused by this .

Reading EMJ2 p2 there is a statement that says the ECM controls the ground circuits through "quad driver modules"

With a suspicion that the temperature has some interaction on my problem , and the CTS being a variable resistance allowing a 5v supply to earth via this resistor? should i think about getting a new CTS from somewhere ?

Any advice appreciated

Chas54

warm start od ECM with newer chip.zip

EARTH TO O2 SENSOR.zip

[dapinky]

Whilst there is a published set of resistance values for the CTS dependent on temperature (in the Wiki), they are not a dear item to buy..... unless you really want to buy a Lotus part with the Elan part number.

If you want a Lotus part, then get the Esprit one - it is far cheaper.....

.... but an aftermarket one is cheaper still...... https://www.rockauto.com/en/moreinfo.ph ... 48&jsn=621

[chrism]

It does seem from your latest scan that the O2 sensor struggles until a certain temp is reached. This time it seems to perk up when the t'stat opens but I'm thinking is this a pure coincidence that the exhaust heat could be supplying the 'perk' just as the t'stat opens?

We're blessed with 3 temp sensor controls; the dotted line comes from the dashboard sensor - that's working okay - and we have to assume the fan sensor's okay from earlier scans(?) - which leaves us with the CTS which only the ecu 'sees' - and nothing else is obvious to show it's 'dud'.

Personally I think I'd throw a new one in since you've now got a shiny new O2 sensor and some earthing betterment that seems to have sorted something out .

I wonder if there could be an earthing issue at the CTS too? I haven't had the misfortune to remove mine (a local pal has, cos he thinks he might have broken the top connection) but I wonder if there's an earth through the mounting/threads that's being difficult - indeed, could it be loose? Is it easy to check?

If the tightness checks out I think relacement'd be my next move (even tho by all accounts it's a pig to get at).

Even so, the scan shows sickness you're in a:f 14.7 and closed loop so I reckon the car should drive better? I'm waiting on a road test ?

Here's a pal's S2 traces when he was grappling with (we think) a broken CTS as well as O2 sensor troubles (also now a suspect EScan cable but it's driving well...)







Similar immediate crashdowns?...

[dapinky]

I wonder if there could be an earthing issue at the CTS too? I haven't had the misfortune to remove mine (a local pal has, cos he thinks he might have broken the top connection) but I wonder if there's an earth through the mounting/threads that's being difficult - indeed, could it be loose? Is it easy to check?

If the tightness checks out I think relacement'd be my next move (even tho by all accounts it's a pig to get at).

The CTS screws into the aluminium of the housing, but some of the new units come with a smear of dried red paste (some sort of sealant, I suppose) - which suggests to me that the Sensor-to-housing connection isn't 100% important. Clearly, it is immersed in coolant/water so there will be some connection that way.

I'd be careful with trying to tighten a non-leaking one as it is a taper thread which shouldn't need to be tighter than a fluid seal.

It is a quick and simple job to change - as long as you approach it logically......

1) do it with the engine cold - keep the header tank cap on to prevent any fluid loss when you remove the old sensor
2) mark the current position of the CAS unit- a bit of masking tape along the bracket marked with the exact position is easy to do, and saves a lot of potential grief if you forget!!!!
3) unplug the 2 connectors on the CAS, remove the holding bolt and pull it out
4) you can now see and get to the CTS - undo the plug from the top of it, then unscrew it - a 3/4" ring spanner fits (19mm is close enough, albeit not 100% correct) - you will be lucky to get a socket anywhere near doing the job!
5) if there is no water system leak, then no water will come out when you remove the old unit - but it is worth loosening the cap once it's out to flush any debris out of the housing, then screw the new one in..... just enough is good enough - do not over tighten it - it's better to do it up a bit later than risk stripping a thread now.
6) replace the wiring plug, refit the CAS, and put the 2 plugs back on..... line up the marks and do up the clamp bolt.
7) have a cuppa to celebrate.

[KenS]

Just a few thoughts that might lead to some discussion.
The CTS is the temperature sensor seen by the ECU so the value seen in ElanScan should reflect the temperature the ECU is working with, so I don't think the CTS is generally bad unless there is noise on the sensor signal that we cannot see in the ElanScan traces, so no harm in changing it if low cost and easy.
Been looking at the detail in the plots again, there is possible match with the O2 sensor starting up again with the point at which the IAC reaches its warm-up stable value of around 24 (about 470seconds in) :

1-9-2026.jpg

Also at that point the Manifold pressure reduces a bit more (more vacuum).
Could this 'O2 sensor' drop-out actually be a valid lean mixture? But then, since it is in Closed loop, why doesnt the ECU try to enrichen it? Looking at the Injector pulse width and duty cycle traces during this 'dead' period it doesn't look like the ECU is trying to increase these to get the air/fuel back to normal. Could this be because the ECU map is working at its limit and won't allow any further fuel increase? If so could this lean be caused by an unmonitored vacuum leak? And then its only when the IAC is closed well down that the total volume of air is low enough for the closed system to kick back in? This might explain why on a run there were instances where increasing the throttle caused the O2 sensor to 'die' and drop out of closed loop, again where there is too much air for the ECU to accomodate with fuel adjustment.
I admit my knowledge of the engine management is not great, so again very happy if anyone can confirm the reasoning or shoot it down! Not sure how to run any tests on this, maybe closing down the Idle Adjustment screw? but not worth doing unless there is a real chance this could be related.

[Chas54]

Thanks for the comments, advice and the Rockauto link .

Ive ordered a Rock auto CTS although with postage and vat it goes up to £22 but still far better than all the Lotus dealers at £45.

Ill probably fit it at the same time as I fit my 88C thermostat.

Due in the next few days ..

Chas54

[KenS]

Another thought/question - does anyone know if an ECU rule is to stay in closed loop during the warm-up period even if the O2 Sensor voltage stops oscillating (e.g. 'lean' period) exceeds 4 or 5 seconds? Then what determines the time of that warm-up period?

[chrism]

Here's my pal's O2 sensor vs closed loop. Well more than 5s...unless closed loop is a target value only, rather than actual?

and here's my Denso low wattage sensor taking ages to warm up.

More importantly you can see my Denso's dotted manifold pressure drops like a stone, Ken. Your Chas54 trace took hundreds of seconds to drop. Yes, it looks like there's a correlation but I'm wondering if there could be a blockage instead of an air influx? eg air filter? rodent's nest in the trunking? etc...

(ps would love to know how you 'pinked' your trace - really useful )

[Chas54]

I can see from Chrism trace of the denso sensor that my manifold pressure drops a lot slower which could be indicating a vacuum leak but the IAC is fully open allowing loads more air in. Also i am having to feather the throttle to get it to keep running initially, letting even more air in. My throttle is fully shut at 0.4v . I tend to start the engine then use the throttle to keep it running whilst backing of on the throttle untill i se the 0.4v. it will idle reasonably then.. It adds a bit of confusion to the initial start up traces. I think i might take the IAC valve of to check if its at all stiff.

I gave the car a run today with the elanscan running.

Idled OK in the garage but somewhere around 90secs the O2 sensor stopped again and didnt restart until around 520secs. The O2 sensor lean operated almost as soon as i got on the road. I stopped in a layby to look what was happening and could see the O2 sensor was working again.

Went to garage and filled with some fresh E5 . The ECU light had cleared by the time i got home.

Looking at the initial O2 lean at about 670/671 secs, speed 11mph . Reaching a T junction before getting onto main road. O2 signal was giving 0v and the IAC jumped jumped from 22 to 31 . The throttle was fully shut and the lean fault activated. Manifold pressure can be seen dropping from 0.63 to 0.29 when the throttle shuts ..
This would seem to indicate a sudden rush of air into the plenum and the IAC being the source . I will look at the air intake for any blockages but I’m pretty sure i would have seen something in the upper parts so i may have to check the air cooler. Its out the way and could perhaps be damaged .

idle to waarm then out onto road and petrol fill.zip

I should have both the 88c thermostat and the CTS on Friday so i may possibly have them installed by Sunday. .

Funny enough, the car went a lot better after the fill up. I'm going to have to check the pump and suction pipework now i think.

Attached today's run with petrol fill up .

Does everyone else have a continuous output from the O2 sensor when viewing their elanscan traces , or are there any breaks like in mine ??

Chas54

[KenS]

Chas54,

Does everyone else have a continuous output from the O2 sensor when viewing their elanscan traces , or are there any breaks like in mine ??

It should be continuous, one of my examples:

my start+service mode.jpg

30-5-2026 cold start service at 600sec.zip

( I switched to 'Service mode' at 610 seconds to 710 seconds which caused the apparent hunting of the revs)
In this example the O2 sensor took about 35seconds to get up to temperature and start working. Chrism's example of his Denso O2 sensor took about 100 seconds to get up to temperature. We have found the start-up times do vary between manufacturers and part numbers, almost certainly related to their heater resistance and hence Wattage. Your new sensor is quick to initially start-up around 30 seconds but the subsequent long drop out is not normal. It is also odd that that initial drop out goes down to about 50mV but later when driving, the drop-out goes even lower, very close to 0V, this was noticed by RonR. These upsets of the O2 sensor while driving are still sufficient to cause it to drop out of Closed loop, again I don't believe they are normal, but I guess there are not so many plots available to prove that. Is it obvious when driving that it is not running optimally?

warm+road.jpg

More inspection of your latest plots, there are a few interesting points including around 800 sec that I want to poke around! Big question is if the O2 Sensor data is valid or corrupt, should its voltage ever go to 0V?

[Chas54]

2ND run after CTS and thermostat change and faults cleared.zip

Thanks for looking Kens .

I can feel the car stuttering a bit just before the ECU lights up when on the road. But at other times it runs and feels OK.

Ive got my 88c thermostat and new CTS on Thursday and they are now both fitted .

Ive attached two elanscans of my warm up then road test , the 2nd being after a battery disconnect to clear the faults . It ran poorly after the fault reset and although the ECU fault light came up and stayed up, no faults were recorded??.

The 1st run although i had the gap in the O2 sensor output , and the ECU light came on just before i reached the main road , It did clear and the car felt good but on returning i could see the O2 sensor wasnt working and gaps in the trace .

The 2nd run after the battery disconnect was bad the O2 sensor didn't perform at all , just a few spikes .

Ive really got to strip the looms taping off under the inner wing and follow the earth from the O2 sensor back to the plenum earth.. Something is affecting my O2 sensor.

Again i'm grateful for all comments

1st run after CTS and thermostat changed.zip

2ND run after CTS and thermostat change and faults cleared.zip

Chas54

[chrism]

Ive really got to strip the looms taping off under the inner wing and follow the earth from the O2 sensor back to the plenum earth.. Something is affecting my O2 sensor.




I found mine had unpinned from the m/plug under the black plastic shroud bottom lhs. I was under the impression that you'd already checked this though?

Also thought you'd substituted wires back to the speedo m/plug area from this w/arch m/plug? And from O2 sensor to the w/arch area?
Better than unwrapping the tape?