High CO% (misfire, glowing manifold, lumpy idle, hgf sorted)

[madbilly]

Hi everyone,

I had some problems on the motorway yesterday leading to me getting a tow to my destination (helpfully my parents' shed where I was going anyway to change my belts and pulleys - now I've got more jobs to do!).

After about 3 hours of driving, including a stop at a services about halfway, I started to get a serious misfire, not continuous, seemingly independent of revs or load, eventually leading to the check engine light (CEL) coming on. After the CEL coming on the misfire seemed to go away, but I needed a "comfort break" anyway so stopped at the next services.

When I got out there was a "singed" smell, so I popped the bonnet and saw the exhaust manifold glowing red where it enters the turbo. Is this normal?

I started the engine again (no problem that I recall) and ran it for a bit to pump coolant round, and after a few minutes the manifold looked normal again so I turned it off and went off to do what I stopped for in the first place! When I came back I started it and the idle was all over the shop - misfiring, backfiring and one or two "pings". I decided I needed help so I called the AA. They weren't much use TBH but I'll give him the benefit of the doubt. On starting the idle was terrible again, but when he revved it ran better, didn't misfire and then idled smoothly, if slightly hunting. His diagnosis was that the misfiring could have led to fuel burning off in the exhaust manifold which made it glow red, but didn't do any tests to work out what might have caused the misfires (and I didn't do the paperclip test) and decided I needed a tow home. Three and a half hours later the "angel" arrived to pick me up (nearly twice as long as they first said it would take)!

So, next day, I started it and the idle was still terrible. I then checked the fault codes and got only code 42, "electronic spark timing". I started down the fault code diagnosis flow chart by clearing the fault codes then starting the engine and idling it for a minute to see if code 42 was set and it wasn't (though the idle was terrible again until I revved it and then it idled okay). So can only assume that code 42 was set by the misfiring on the motorway yesterday and the terrible idle isn't setting a fault code - both could of course be caused by the same faulty component, or I've got two or more faulty bits of kit. I didn't have chance to do anymore today due to social engagements but I'm erring on it being an ignition coil problem, or should I be thinking cam angle sensor/hall effect sensor (CAS/HES)?

I gather a whole CAS is very hard to come by, but the HES is much easier to come by? I know the coils can be replaced individually, but is the "direct ignition circuit" referred to in the manual the same as the "ignition module" that's part of the coil pack along with the coils and base plate? Or is it ANOTHER part that could be faulty?!

Any help greatly appreciated , as I've only got this week to diagnose fault, order parts, fit and test them as well as changing all the belts and pulleys that I was planning to do this week. First thing I'll do is go through the diagnostic flow chart in EMK 3 on the Electronic Spark Timing and check the coil resistances. Any other suggestions?

The AA man also decided I might have a blown head gasket due to the steamy exhaust, but since then I've checked all the usual signs:

Water loss? None.
Water clarity, smell? Clear, doesn't smell.
Dipstick? Only oil, no gunge or condensation.
Oil filler cap? Only oil, no gunge or condensation.
Start engine from cold, how quickly does it pressurise? I started it with the expansion tank cap off, no obvious pressure after a minute or so, then replaced the cap and drove the car from where the AA man deposited it to the shed where I'll work on it, then removed the expansion tank cap and there was a small hiss but nothing else, so basically barely any pressurisation in about 5-10 minutes of running.
Steamy exhaust? Yes, but it's a fairly cold day and I didn't run it for long.

So, does it sound like HGF? I don't think so, but any other thoughts?

Thanks,
Will

PS 1/4/12 Mods please move this if you think it would be better off somewhere else.

[rip]

Steam & carbon dioxide are the only two products of complete combustion of a pure hydrocarbon. (Petrol has lots of impurities & engines rarely burn fuel completely).
For some reason Elans produce more visible steam than most cars, so this is not a sign of HGF unless it is excessive.

If you can run Elanscan, this may be able to help solve your problem.

[madbilly]

Checked coil resistances, 0.3 ohms across the primary, 5.5k ohms across the secondary of one and 5.4k across the secondary of the other. Is 5.4k a bit low?

[madbilly]

Decided not to do the EST diagnostic process (EMK 3) because at idle the engine's running fine. Haven't taken it out for a long run again yet, going to do that tomorrow.

My HT leads are red (though very dirty) and say "R-16 E-E Sumitomo 1989" on them. They look good quality, but the 1989 makes me wonder, are this the original HT leads?! Anyone else had some like this?

Number 4 lead looks like it might have very slight perishing, and has a resistance per metre of 22.1 kOhms whereas the other leads are 18.4, 17.2 and 17.8 kOhms per metre for no. 1-3 respectively so maybe number 4 is struggling. Performanceleads.co.uk are making up some leads for me, I should get them on Thursday.

I don't suppose the fact that the 1&4 coil has a slightly higher resistance and the 1 and 4 leads are the higher resistance ones is intentional is it? I suspect it's a coincidence.

I know there's a lot of wisdom out there so if anyone's got any thoughts on any of this that might help me please let me know!
Cheers,
Will

[madbilly]

Okay, I took it down to a garage yesterday for them to case their eye over it. They couldn't suggest anything definitive but were fairly convinced by the smell that it was running rich! Obviously lots of sensors that could be causing this, including CTS, so I hooked up ElanScan and compared the cold and hot tables in EMK 4-8 and 4-9 and there was nothing out of range. I've attached the scan in case anyone can look at it and point out anything that looks out of the ordinary, plus one after I'd tried adjusting the mixture; here are my thoughts:

After the first run from cold I went into service mode to set the timing and CO. I couldn't get the CO to go below 5%, and I'm not sure my adjustments had any effect at all - the second log I attached shows CO pot still inputting to the ECM at 2.0V, same as in the first scan, so maybe the pot's broken? I'll check it tomorrow.
At ~1250-1260s in the first scan I momentarily went full throttle twice to try and "blip" it, but what actually happened was a drop in revs, nearly stalling. Another symptom of running too rich?
The spark advance seems to be all over the shop, is this normal?
In the first scan, during my first start to check the comms, I start off with 67 knocks and go up to 75, weird! When I next start the engine it's at 0.
What are DRP interval and Voltage Comp?

Other things I've checked. All the injectors read 2ohms across the coils. Pulled off and put back on most connectors and vacuum hoses.

So, lots of things are pointing towards the misfires being caused by a rich mixture rather than ignition system problems, so why did code 42 come up on the motorway? Well, I think it was a symptom rather than the cause. If the engine misfired sufficiently to cause a stall, then the car would recover from that stall in a "bump start" fashion because I was travelling at speed with the clutch engaged, and I recall that the CEL light came on and STAYED on when I changed down a gear during a misfire (I had plenty of misfires, and it was several before the CEL light stayed on) - the transient disengaging clutch could cause a REAL stall to occur which was subsequently recovered from by bump-start when the clutch reengaged, during which time the ignition module and ECM could get out of sync, leading to code 42 being set. After that the car would have run in bypass mode until I stopped, which would probably be a rich mixture? Whould that make the exhaust manifold red hot where it enters the turbo?

One other thing. The second time I tried to set the CO after I'd tried the analyser out on another car, when the fans first came on I was away from the car and there was a really loud, high pitched screeching - is there an audible low oil pressure alarm on the Elan, and is this what it sounds like? If not, any other ideas?

Another other thing. The garage yesterday did plug in a "Multiscan" diagnostic tool with the matching connector they had. With the ignition on but the engine off the CEL light was flashing code 12, but inverted, so it was blinking off rather than blinking on. Other than that and making a very audible relay click on and off at the same time as the CEL light blinked the Multiscan provided no other information. BUT, when I started the car after we'd unplugged the tool the revs just went up and up and up and I panicced and turned the ignition off; the same happened when I started the engine the next time but only up to ~2500rpm. Is this what happens if the ECM doesn't know where the IAC valve is?

I know most of my posts are essays and not that appealing to read, but if someone can give me some feedback on whether it looks like I'm on the right track or not, or can answer some of my questions I'd be grateful.

Thanks,
Will

[racerstev]

This sounds like a DIS problem to me. What EST does is switch the timing from static to DIS control,
on a Isuzu with the same setup static timing is 10 degrees. If the EST dosn't switch it could
lead to the turbo and manifold glowing red because of the retarded timing.

If the DIS modual itself is going bad it could account for all your problems. The modual sits
under the coils and is prone to getting coolant on it from a overfilled or overflowing coolant
bottle.

Steve

[madbilly]

Thanks Steve, I'll check the ignition module is behaving correctly tomorrow.

Took the CAS out today and found the peg was sticking out one end, like it had pushed the spring to one side. Put the peg back in, refitted etc, and the wobble in timing marks I can see when using the timing light is much less, down to less than a degree. I haven't had chance to roadtest it yet to see if that gets rid of the intermittent misfire. On my latest scan (attached, spark advance is still all over the shop so I assume this is controlled by the ECU, since I've sorted the CAS wobble.

I've also got the new HT leads from Performance Leads in. Ditto on the roadtest and misfire. However, in cosmetic terms they look good! Definitely fatter than my old leads though - it's a bit cramped! So Performance Leads can now provide sets of leads for Elan SE's, though I need to let them know to make the lead for no.1 cylinder a cm or two longer. They cost ~£57 in total delivered.

On the CO pot, I took it out and tested it and it works fine. I checked the connections to the ECM harness and they're fine. I wound it down to minimum resistance (which I believe should be the leanest mixture) and put it back in. None of this has made any difference to the mixture, I can't get it below 5%, and in ElanScan the voltage still showed 2.0V. I've looked at other people's scans with open loop systems and those with closed loop and the CO potentiometer value stays ast 2.0V on all the open loop systems I looked at, and only varied on the closed loop systems, so I think in ElanScan this variable is misnamed?

If I increase the idle speed in service mode I can get the CO % to go down, so I can pass an MOT, but have a silly high idle! Does anyone know if the having the idle at 950 rpm is for anything other than comfort and fuel consumption?

One more thing I'll check tomorrow is the crankcase breather, as EMK-6 suggests this and I tihnk I read a mention of this in another topic somewhere on the forum.

120323 from cold.zip

[madbilly]

PCV valve and oil separator and their hoses all fine. Changed oil and filter, got CO down to 4%. Checked CO pot signal was getting to ECM properly, it was, and turning the pot from min to max gave full range of volts 0.66 to 4.2 (actually ~0.63 to 4.15 but it's close enough).

I decided the screeching noise I heard the other day was the alternator belt (somewhat obvious eh?!) - it's been slipping since, but I really can't see how this would cause it to run rich.

My thoughts are now turning towards the fuel system. I checked the MAP hose to the fuel pressure regulator, it's clear and has no leaks, so perhaps the diaphragm itself in the regulator is getting worn out. Another possibility is the injectors may be leaky. I'm taking it to a garage for them to check tomorrow as I don't have the kit to do that.

[madbilly]

Well, the garage I took it to couldn't test the injectors or the pressure regulator - they said it was specialist, is that normal?

Anyway, they cleaned out the throttle body which apparently got the emissions down to ~2.5%. They said the rest of the difference between that and spec would be down to mechanical wear in the engine - piston rings and valve stem oil seals probably.

[madbilly]

Thanks for the help rippers and Steve, I'm pretty sure I've now sorted the misfire and I'm certain I've not got HGF. What I'm focussing on now is the CO%/mixture. I know 2.5% is within the MOT requirements but it's not near the spec of 1% where it should be. I know I can fit a 9930 ECM (if I can find one), fit a mountain chip or go closed loop, or a combination of the last two (and I'm certain to do this in the future) but atm they would be curing the symptom and not the cause, and I want to make sure the car is running as it's meant to before I make any mods.

As far as I can work out, high CO% emissions could be due to overfuelling OR it could be due to something else.

Possible causes of overfuelling (I think) that I haven't checked:

- Leaky injectors
- Fuel pressure too high (faulty FPR, blocked fuel return line)
- Faulty injector driver circuit (results in injector pulses being longer than the ECM intends)
- Faulty MAP sensor (from ElanScan it seems to read the correct range (=barometric when engine stopped, 0.35bar at idle when warm), but how would I know if this was working correctly, and how would it affect fuelling, eg would reading too high cause overfuelling or underfuelling?)
- Mechanical valve timing incorrect (as in either the intake or the exhaust camshaft is a degree or so out, but I would have thought this would manifest itself in ways other than a rich mixture)
- Spark not strong enough? (I've got new plugs and leads, haven't changed the ignition coils because I think they're okay)

FYI, unburnt HC reading was 403 ppm, how does this compare to open loop cars that are bang on the CO% spec of 1%?

BUT, if it's not overfuelling, and another fault is causing high CO%, then I could actually be running the engine lean which will end up knackering it.

So I need to get a proper lambda/AFR test done. As far as I can work out the AFR on an Open Loop Elan should be 14:1, is that right? That's what ElanScan shows the ECM thinks it's doing at idle.

If I do find out I'm running lean, and adjust the CO pot to get back to spec, the only other things I can think of that could cause high CO% NOT due to overfuelling are:

Leaky exhaust? (not sure how this could cause high CO but apparently it can)
Secondary throttle is stuck slightly open?
EGR valve is stuck slightly open?
Clean out the entire intake manifold?

All thoughts and help appreciated.
Cheers,
Will

[dapinky]

Will,

IIRC from when i was still 'open loop', my CO2 readings were always in the region of 2.5 - 3 %, whatever i did with the pot.

On one MOT (with a failing HG), i couldn't get it within the limits without inducing an air leak (disconecting a vacuum pipe) in the system, and when i did, the HC levels were far too high (HG!).

I think that your numbers seem okay (or at least, as normal as they can be expected on a car of the age/mileage), but WILL improve by going to a closed system.

[theelanman]

Dont forget your CTS as it too wil give you overfuelling
(coolant temp sensor)

[madbilly]

Dont forget your CTS as it too wil give you overfuelling

I think the CTS is working fine. On my scans I posted it starts off same as ambient (MAT), you can see the thermostat opening at ~85 degrees and closing again at ~80. On another scan (not posted) I had the fans coming in at 104 or 110 I think.

Dave, I know 2.5% is okay, I just to figure out what's preventing me getting down to 1%. Once I've figured that out I'll go closed-loop, happy in the knowledge that I've sorted any problem, rather than mask it by going straight to open loop.

I've made serious use of the search facility on the forum, checking for anything else that people have found reduces emissions or causes overfuelling.

[madbilly]

Hi everyone,

I decided to have another go at setting CO% and idle following the garage reducing the emissions by cleaning out the throttle body. In service mode the idle was ~2000rpm, probably due to the throttle passing much more air at idle due to the aperture being cleaned out. I wound it back in to 950rpm, but when I came out of service mode the car stalled. I tried to start it a few times and had difficulty unless I had a little pressure on the accelerator pedal. I then wound the idle screw out a bit and tried starting again, and this time it was fine.

That's basically how I left it, the issues I have are that when starting the idle drops to ~500rpm for a few seconds then climbs up to a normal 950rpm, I presume this is the IAC valve being moved to get the right idle speed. An CO% in service mode is now back up to 4.5%! So, what did I do wrong when setting the idle? If nothing, then do I have a faulty IAC valve? I get no errors in Elanscan.

What could cause the emissions to be higher with a narrow throttle idle setting, and lower with a wider throttle idle setting, but with the same idle speed? I would have thought that at the same idle speed there's the same air going in, the same MAP and therefore the same fuelling, same mixture, same emissions, but that doesn't seem to be what's happened? I think I'll hook up Elanscan again and watch the IAC valve position and see if there's a difference with narrower and wider throttle idle settings.

I know this is a ramble, but writing it all down helps me think about it!

Cheers,
Will