V2Rocket

Been playing with my 944 lately, one thing I notice is that even when hot (after a decently long spirited drive) the exhaust smells pretty rich when idling.

Lambda readings are within 2% of my lambda target (since I have an NA the WBO2 is mounted in the standard O2 sensor location in the 2-1 collector pipe).

When I first installed the VEMS/WBO2 I had to calibrate the sensor with "free air" so it should be OK. I believe it to be still good because...I had set my idle lambda to 1.00 and it worked fine but wasn't super smooth, when I got home I lowered it to 0.96 and the idle smoothed out very well.

Anyways...with good lambda readings, why is the exhaust smelling rich? No sulfur/rotten egg smell, just extra petrol in the exhaust gas.

KVDR

Do you have a cat?

V2Rocket

No, a terrier

Yes, I have a CA-legal aftermarket replacement catalyst installed...about 4.5 years old. Passed my last smog in ~February this year very easily.

KVDR

Oh you!

Mine smells like gas sitting in the garage. I need to check my evap system this winter.

V2Rocket

Mine does that too, but this is definitely gas smell from the exhaust pipe while idling.

odonnell

Mine does the same idling at 13.5 AFR, which is where it seems to be the happiest with semi sequential injection. I just ignore it, it was only an issue when I was working still tuning my warmup enrichment and idle was a 11.5 sometimes... You could smell it while reversing.

V2Rocket

still chasing this...
target AFR is 0.97 lambda, actual AFR at warm idle per wideband readout in VEMS says between 0.95-0.99 so right on target...but it smells rich out the tail pipe while idling. car runs fine and doesn't put out the stinking rich smell as if it were running at 12 or 13:1 when driving though.

is it possible for a "free-air calibrated" sensor to lose its calibration?

Raceboy

It is most likely not CO but NOx emission and that needs to be cured with ignition timing. Please make a log and check what is the timing during idle (it automatically adjusts timing at idle to have it smooth and stable), it may be too agressive which may affect emissions.

Tom M'Guinn

Are you using a wideband to control closed loop operation, or is there no closed loop? Why not target Lambda of 1? .95 is on the rich side -- under 14 -- so maybe it's that simple? Also, do you know the response time for the wideband and controller? Some widebands are slower than others, and some controllers even more so -- so if you are trying to run a closed loop idle, it might help to know how many times a second you are getting a valid reading to base the next correction.

On the other hand, I had an RX7 many years ago that always smelled rich. The smog guy would even warn me in advance, but it always passed like a champ, so that rich smell doesn't always equate to an true rich mixture it seems...

V2Rocket

ECU has closed loop via wideband input. there is no narrowband.
Car idles smoother at 0.97 target +/- 0.02, at 1.00 it gets a little bouncy. I *might* have a very slight vacuum leak in one of the under-manifold hoses but no time lately to investigate.

0.95 is close enough to 14.0 which seems to be a safe/typical idle afr?
It is a very fast response - the tuning software has a live "gauge" panel and it oscillates the WB reading a few times a second.

odonnell

Vac leak won't affect a MAP based tune in this context. It's most likely that your WBO2 needs to be recalibrated, or has a ground offset relative to the ECU. On my MS setup I was seeing that the actual reported lambda value in the software was slightly off from the controller/physical gauge sometimes, due to the ground offset. This is solved by slightly altering the wideband calibration (to the ECU software) until it's a direct match.

V2Rocket

LOL you're right now that you mention that...
MAP is MAP...
Dunno about offset because the WB plugs directly into the ECU.

Humboldtgrin

If you have any bit of a exhaust leak before the O2 sensor that may be a problem.

V2Rocket

still chasing this...hadn't thought about exhaust leak but will have to look for that later.

was playing with timing and tweaking fueling a bit...still smells like its rich at the tailpipe (not raw gas smell from the tank, but rich exhaust for sure). lambda target 0.97 and it fluctuates +/- 0.05 or so. tried going open loop, closed loop, adjusting the VE table, adjusting lambda target, adjusting timing...no effect on smell but definite changes in how the car idles.

had thought there was too much timing (idles best at 20 degrees or so) because the WSM says 5-15 degrees at 850rpm or whatever (i have mine warm idling at 1000)....backed it down as far as 12 degrees, just made it run rough and lower RPM into the 700s despite 1000 target. still no change in smell...added timing up to 26 at idle, it was getting kind of unstable sounding at that point, still smelled the same.

backed it off to 1.00 lambda target, it was fluctuating +/- 0.03 around there, running smooth just smelly...

one key detail that i didnt know about when starting this thread was that my cat is toast - several chunks of matrix fell out of the muffler when i was investigating a rattle... and i had to add VE to make it idle better after that (less blockage = more air flow = more VE), but still lambda numbers on target.

so...

1) how much timing does an 8v 944 typically want at idle?
2) are all "catless" cars this stinky?

kev951

exhaust leak?