Alas rough running, odd smell, code 44

[DeanG]

I have had this problem for a couple of years however it is now getting constant enough that I might be a find and fix the problem. It is a North American car.

On starting the car at the local Lotus dealer (they only service Elise/Exige) the car started running rough. It had been sitting about a half hour. It cleared up within a mile or so. Next start after sitting less than 5 minutes no problem. Next start after sitting about 15 minutes it ran very rough and never cleared up in 30+ miles of driving. It feel like it is running on 2 to 3 cylinders althoug this varies. It won't pull over 4,000 rpm. Boost sometimes appears on gauge when I push it as hard as it will go. I recieved a short CEL which was a 44 Lean exhaust. Very strong smell, somewhat electrical in nature, possibly from too much raw fuel in the cat. When I got home I couldn't get the laptop to boot so no elanscan.

[GeoffSmith]

Dean,

It sounds like your oxygen (lambda) sensor is on the way out. The output will be interpreted as being lean by the ECU so it will keep enriching the mixture. An ElanScan check on the error codes would be good and a log will show what the sensor output is doing.

Geoff

[DeanG]

I have had the O2 sensor replaced a few months ago. There appears to be no difference in my intermittant problem.

I am still trying to revive my Laptop. I think it is a mechanical problem with the power switch.

[Doug]

ElanScan should make a big insight

could easily be a sensor like a CTS

[GeoffSmith]

ElanScan should make a big insight

could easily be a sensor like a CTS

It is quite easy to misinterpret codes when doing it manually.

[DeanG]

Here is a ElanScan from today. The car was stationary in my driveway and I was using a borrowed PC. My laptop is officially dead (mother board). I don't understand what ElanScan is telling me. The engine runs very rough.

There seem to be many spikes in the data. The graphs of:
TPS voltage
Manifold press (abs)
show the spikes. I would think that these curves should be smooth. Could this be due to the voltage stabilizer? The 'Battery volts' graph seems odd.

The graphs of:
Spark advance
CO potentiometer/Oxygen sensor?
have very sharp spikes. Since this is a Federal car I don't understand why there is any the CO potentiometer/Oxygen sensor? data.

Within the last 6 months the thermostat, CAS and O2 sensor have been replaced.

What is 'Op. temp. - temp.'?

[GeoffSmith]

Dean,

I cannot see anything unduly worrying in the TPS graph (the TPS is the voltage from the throttle position sensor so I can see you've applied a bit of throttle between ~100 and ~120s). I cannot understand why the manifold pressure increases, the spark advance changes and battery voltage drops at ~90s though.

The ECU uses the same monitor channel for open loop cars with a CO pot and closed loop cars with an oxygen sensor - in your case it is reporting the O2 sensor which still looks wrong to me. In your previous run last July, the O2 signal was consistently very high (close to 4V). At least this sensor is oscillating but it should be swinging between ~0.5V and ~4V whereas it seems to settle down at ~1 to ~3.5V. The sensor does rely on an earth contact via the sensor body, manifold, engine etc so I wonder if the earth is a bit poor; you could try an earth strap.

Your problems seem very similar to Mike but his was cured by changing the sensor.

I'm not sure what 'Op. temp. - temp.' is, sorry.

Geoff

[DeanG]

Thanks. I will look at the grounding issues.

What about the wild swings the spark advance? Do you have any ideas about that?

The car is basically undrivable. This was killing me yesterday as we had 70 F and blue skys.

[GeoffSmith]

Dean,

I was about to post NO but looking all the way down to flag #36 (inputs), bit 7 is labelled A/C clutch and rises at ~90s about the same time other weird s**t is happening. Is there any chance your aircon switched on then? If so, this may explain some of the anomalies but your O2 sensor signal still looks bad.

Geoff

PS a good first log is to start the car from cold and just let it heat up until the radiator fans have cut in a couple of times. Make sure the aircon is off though!

[DeanG]

Yes, I switched the AC and lights on about 90 seconds into the run.

Recieved another used laptop in the mail today so I should be able to make one working laptop out of the two I have. Then I can get ElanScans more readily.

[GeoffSmith]

Yes, I switched the AC and lights on about 90 seconds into the run.

[RayD]

Dean,

Have the spark plugs been changed in order to rule them out.

Ray

[Doug]

Dean, can't seem to open the zip file -- can you upload a non-zip version?

I'll try the file at home where I have all the programs too and I can probably open it there.

________________________________

gave it a check at home -- nothing obvious

PS a good first log is to start the car from cold and just let it heat up until the radiator fans have cut in a couple of times. Make sure the aircon is off though!

[GeoffSmith]

Dean, can't seem to open the zip file -- can you upload a non-zip version?

Doug, although this is a site dedicated to the the Elan, Elan log files cannot be uploaded to the site.

They do compress quite well though, so it's no big issue.

[Ellie]

I'm with Geoff on this one.
One of the really handy things about Elanscan is the simple ability to overlay data (open a dataset, then open another - Elanscan asks you if you want to overlay the data). This can be done with multiple datasets, and I suspect that if one were judicious enough to save datasets say every 6 months and compare them, you'd see the oxygen sensor output deteriorating over time.

I have compared Dean's dataset (zip above) with Geoff's 'reference' data and mine both before and after O2 sensor replacement. In my [novice, but learned after the experience] opinion, Dean's O2 sensor is not at its best by a long way. Get a new one (in the States they're cheap - $50 to 80 with free freight) and try again.

The next part is the secondary diagnostics - what's underlying the short life of the sensor. I think 2 or 3 things.
1) sensor poisoning - do not use ANY silicone sealant near the sensor at any time or you will kill it.
2) possible catalytic converter failure (could be a symptom of O2 sensor failure) cat failure causes no boost & poor running.
3) engine sensor problem upstream like MAP sensor or fuel pump - anything giving an incorrect dose of fuel (too much fuel or not enough air) causing sooting of the O2 sensor and restricting its ability to read. I think (on other threads) that the ECU's response to a 'lean' O2 sensor reading is to boost the amount of fuel to the engine to make the mixture richer to prevent the lean burn (bad news for the turbo).

My dead O2 sensor was heavily sooted. I had code 44 (lean O2 sensor) and restricted engine revs. New O2 sensor and I'm a happy boy.

[DeanG]

You folks have convined me to try anoher O2 sensor. I will post when the new sensor is in place.

Having an unreliable Elan is depressing. I very much hope that I can get the car running reliably enough for longer trips.

[DeanG]

In poking around I don't recall having my ignition modue grounded. In the Wiki I read a note that there should be a ground strap. I don't recall having one. I do recall the insulating washers which I assume is to float the module at some ground potential other than that of the chassis.

To what shoud the igntion module be grounded?

[GeoffSmith]

Here's the write up on the coil pack in the Wiki. If you check Page 4 of the circuit diagrams, it looks like it is strapped to the inlet manifold (twice). Now if these were missing, then that could certainly affect the ground reference of the O2 sensor - it would certainly be worth checking Dean.

[Doug]

Dean, IMHO worth a few more runs from cold just to see the data before the change--since you have the time and computer at hand.

I had a CTS go bad and it ended up sooting up my O2 sensor and then I had to replace it, after I found it had gone bad once the engine was running with the new CTS.

Now if the car were running poorly and then you stuck the paperclip in and reved the engine up and let it settle into "service mode", if the enigine improved it could prove that a sensor input is to blame for the poor running (as you may know the service mode locks out the sensors and reverts to the bare ECU and base mode). The O2 is one of those sensors

[DeanG]

It's raining cats and dogs. I have to wait for the rain to subside. Then I have to wait for valintines day to be over. I'll post soon. Thanks folks.