smog check failure...

[Robert Bondi]

I just relocated from Austin to Albuquerque. In TX, cars older than 24 are emissions exempt. Now, I'm in the only county in NM that tests emissions and no E12s are exempt - (exempt is 1974 and older).

I'm failing HC 5X over the limit at idle, but near passing at ~2500 RPM.

I've got a modified 77 530i. I'm running 81 528i FI and Euro exhaust
manifolds sampling 3 cylinders for a heated O2 sensor. There is no cat.
I was initially baffled because I know
the O2 sensor functions because I have an air/fuel meter; however, it does indicate lean conditions at idle.

After some researching, including the Mike W. FAQ, I highly suspect an exhaust leak around my manifolds is giving the sensor a falsely oxygenated environment. I guess that causes the ECU to send in more fuel that can't be combusted. From excess valve noise, I already had suspected an exhaust leak for a while.

How does that logic sound to smog experts? I also thought about checking the O2 sensor ground and making sure its ambient air vents are free and open.

Robert

[socalfiver]

How old is the oxygen sensor? A failed/failing oxygen sensor will ensure smog failure. New generic (clamp your own connector onto supplied wire) Bosch oxygen sensor is $9 and may fix it.

If you do have an exhaust leak, fix it before you burn a valve.

[Robert Bondi]

Sensor is probably about 8 or 9 years old. I've got a new one. I'm planning to replace the sensor, all the plugs, and hopefully also find the exhaust leak.

[Mike W.]

I think you're on the right track with both the sensor and the leak. What kind of standards are you having to pass? And don't forget just retarding the timing can clean things up even if it does turn it into a dog to drive.

[Robert Bondi]

I think you're on the right track with both the sensor and the leak. What kind of standards are you having to pass? And don't forget just retarding the timing can clean things up even if it does turn it into a dog to drive.

Mike,
I think I'm being held to 1977 non-catalyst standards. Despite all the manuals I own, I'm having troubling verifying the tail pipe requirements. The failed emissions report shows these limits without units:

high and low RPM:
500 (HC)
5.00 (CO)

It seems strange to me that high and low RPM have same emissions
limits.

[Mike W.]

I think you're on the right track with both the sensor and the leak. What kind of standards are you having to pass? And don't forget just retarding the timing can clean things up even if it does turn it into a dog to drive.

Mike,
I think I'm being held to 1977 non-catalyst standards. Despite all the manuals I own, I'm having troubling verifying the tail pipe requirements. The failed emissions report shows these limits without units:

high and low RPM:
500 (HC)
5.00 (CO)

It seems strange to me that high and low RPM have same emissions
limits.

Those are really high limits, I mean really high. My Bav with carbs had to pass tighter than that back in the 80s, so you shouldn't have any trouble at all. The numbers would be 500PPM for HC and 5% for CO. The most recent unloaded standards I had to pass with an E12, 528i, was 220-2.0. You might want to check the FI temp sensor too, and of course verify it's going into closed loop once you get the new O2 sensor in there.

[Robert Bondi]

Mike,
Thanks. Yeah, I had been thinking that I might actually pass the emission test open loop as a 530i and the O2 sensor might just be overly enriching the mixture. They are loose limits.

I'm dropping the exhaust today to start the troubleshooting. I'll locate
or buy a new FI temp sensor.

Thanks for the tips.

[Robert Bondi]

I spent a lot of time going over things, but really didn't find anything obvious. I replaced the O2 sensor, put in new plugs, inspected the exhaust forward of the resonator for leaks, put in new exhaust gaskets, verified reasonable T response for the FI temperature sensor, performed ignition timing, and a few other smaller things w/o finding the problem.

I had little luck tuning the car once I had it running again. I think that's probably because I stumbled across a pretty serious vacuum leak that's hopefully the main cause of this grief. A mysterious oil leak reappeared around the valve cover area of cyl 5 and 6 despite having wiped everything down. I used my new air compressor hooked up to a port in the intake manifold to verify a leak at the rear of the valve cover large enough to feel with my hand.

I tightened the valve cover nuts until I feared stripping them, but w/o stopping the leak. I've order a new valve cover gasket. In the past, I've always put the valve cover gasket on dry - it shouldn't need any kind of silicone to seal, right?

[Mike W.]

I've order a new valve cover gasket. In the past, I've always put the valve cover gasket on dry - it shouldn't need any kind of silicone to seal, right?

Correct, silicone will just make a mess. If I try and reuse one for some reason I'll use some Permatex #3, but not silicone. It works, but at what a price in later cleanup. But a new gasket will seal dry.

[Robert Bondi]

I fixed several more things that should be affecting emissions. Since I had to pull the valve cover anyway, I checked all the valves and found many to be very tight. I adjusted all to be a loose 0.012".

New valve cover gasket made no difference. I still had a big leak in the same area - might be just a little warpage. I used some sensor-safe Ultra Blue silicone for about 1/4" of gasket perimeter around affected area - seemed to work.

I was very surprised to also find a massive leak at oil filler cap. Has anyone else noticed that the old metal style has a circumventing path around the gasket because of the crimp-on cap design? I verified the path with the air compressor. It's hard to believe the design. I also fixed that with a bead of Ultra Blue.

[Robert Bondi]

I now fear I may have poisoned my new O2 sensor. In the middle of tuning, I ended up with it somehow stuck around 800 to 900 mV.

My best guess is my very small usage of Dupont silicone spray w/ teflon to ease assembly of vacuum hoses. The can doesn't claim sensor-safe although it shows automotive usage. I could have sworn the NAPA brand silicone spray I was previously using claimed sensor safe.

I'm resigned to blowing another $36 on another sensor, but how do I know all the silicone (if that's the problem) is gone so I don't poison the next? Would a spirited open-loop drive followed by an oil change be enough before putting in another sensor??

[Mike W.]

I now fear I may have poisoned my new O2 sensor. In the middle of tuning, I ended up with it somehow stuck around 800 to 900 mV.

My best guess is my very small usage of Dupont silicone spray w/ teflon to ease assembly of vacuum hoses. The can doesn't claim sensor-safe although it shows automotive usage. I could have sworn the NAPA brand silicone spray I was previously using claimed sensor safe.

I'm resigned to blowing another $36 on another sensor, but how do I know all the silicone (if that's the problem) is gone so I don't poison the next? Would a spirited open-loop drive followed by an oil change be enough before putting in another sensor??

I don't think so. I think if it was contaminated it would read low, not high. Try forcing it lean, pull the oil filler cap to introduce a vacuum leak. Try clamping on the fuel supply hose to restrict fuel supply and force it lean. Maybe play with the AFM while it's running.

[Robert Bondi]

I don't think so. I think if it was contaminated it would read low, not high. Try forcing it lean, pull the oil filler cap to introduce a vacuum leak. Try clamping on the fuel supply hose to restrict fuel supply and force it lean. Maybe play with the AFM while it's running.

Mike,
You were correct. The O2 sensor wasn't destroyed, although I did repeat the problem again. I thought everything was fine after the car was idling maybe 20 minutes or so. Then, I ran a fast idle (~3000 RPM) for about
30 sec, but after letting it back down to ~1000 RPM, the sensor went rich again. It wasn't stuck, but it had certainly shifted high and rich.

I'm also confused about the charocal vapor canister. I did another leak check with compressed air and found air leaking out the bottom. Upon inspecting the bottom, I found a foam piece covering many holes in the bottom of the can as if it's meant to pass air through the bottom. Huh? Unless there's a check valve in there, how is this possible?

[Mike W.]

I don't think so. I think if it was contaminated it would read low, not high. Try forcing it lean, pull the oil filler cap to introduce a vacuum leak. Try clamping on the fuel supply hose to restrict fuel supply and force it lean. Maybe play with the AFM while it's running.

Mike,
You were correct. The O2 sensor wasn't destroyed, although I did repeat the problem again. I thought everything was fine after the car was idling maybe 20 minutes or so. Then, I ran a fast idle (~3000 RPM) for about
30 sec, but after letting it back down to ~1000 RPM, the sensor went rich again. It wasn't stuck, but it had certainly shifted high and rich.

I'm also confused about the charocal vapor canister. I did another leak check with compressed air and found air leaking out the bottom. Upon inspecting the bottom, I found a foam piece covering many holes in the bottom of the can as if it's meant to pass air through the bottom. Huh? Unless there's a check valve in there, how is this possible?

I think that's how they work. For testing purposes, FI, not the evap system, they say to disconnect the hose from it and just leave it that way. Maybe it can't draw the fumes in if air isn't passing by it?

Back to the original problem, any chance the AFM has been tweaked a bit? Got a spare to try?

[Robert Bondi]

Back to the original problem, any chance the AFM has been tweaked a bit? Got a spare to try?

Mike,
Yes, I adjusted the AFM spring back in 2002. I just went through the
procedure again last night, which resulted in more loosening. I had noticed during previous tuning sessions that pushing the door in a little was good for idle smoothing and increase. Adjustment was made so that the door is naturally opened for optimal idle (open loop) - worth 17 teeth of loosening on this round.

I think I'm close. If the sensor stops getting inexplicably fouled up at various open throttle positions, I'll probably have it. Everything looks good before the sensor fouls up to the high side.

Also, I removed the sensor to inspect. The flutes look very clean - almost new, so it's not that I'm choking it with soot when it goes rich.

I'll try it again tonight.