I think I may be done chasing down an elusive, intermittent P0171. So far this year, I:
- replaced gaskets from the LIM up
- checked vacuum manifold and the miscellaneous vacuum lines (i.e. brake booster vacuum hose and valve)
- replaced cracked exhaust manifold
- replaced the front O2 oxygen sensor -- twice
At first the gasket and exhaust manifold work seemed to have licked the P0171. And I replaced the o2 sensor because the one in while the maifold was broken was no longer acting right. GOt a Delco sensor from RockAuto.
But the code (after a little time) came back.
If I coasted down a very long hill using engine braking I could get it to set pretty consistently. When I drive overland to visit my dad, through the Laurel Mountains on US Rt. 22, there are a couple mountain passes I could reliably get the code to set and CEL to light. One, the hill leading towards Blairsville; the other, the hill from Cresson to Altoona past Horseshoe Curve. And over the course of the weekend visit, the CEL would go back out (although you could easily find the "stored code").
The service manual was a help, in a bass-asckwards way. Some things that cause P0171 are actually designed-out of the Riv, and a scan tool lets you look at the MAP and MAF sensors and reckon what the O2 sensor should be doing. When you get into it, if you read the fine service manual then the P0171 is not *that* touh to track down. Nuts of that -- the sensors seemed to be reading right, and once in a while the O2 would sit at the high end of range instead of pulsing.
Hmmm... Or as some of us would say, WTF? (I wouldn't say that)
Problem was pointing toward a bad O2 sensor. I had told Rick Wakefield about this some time ago and wouldn't you know he had a NOS Denso sensor (among other things) to sell me. Yep, I bit. Although I am too embarrassed to say how long it was between when I bought that NOS sensor from him and when I put it in (yesterday). Only Rick knows for sure...
So I finally got around to removing it (the Delco O2) this past weekend, and Hmm-mm... remove it and find that the Delco sensor was not seating properly in the exhaust O2 bung hole. It had sooted all over where the gasket was supposed to have compressed to seal it up. I may have found the problem. For those of you not familiar with and curious about exhausts here are some details.
What is obvious is that exhaust is hot gases moving through a pipe. What may not be obvious is that it is rapid bursts of gases, not a continuous flow. After each cylinder fires, the gas pulse moves thru the exhaust manifold. It pulses back against the other (closed cylinders and gets reflected back. Path of least resistance is out the tailpipe. As it moves along that path, it cools down. Gas that is cooling off *shrinks.* Plus, its moving, which causes a lower pressure. Matter of fact, it's definitely moving faster than the air outside the crack, so a crack in the exhaust will blow a little at the peak of a pulse and suck the rest of the time. (you knew exhaust sucked, right?!?!?) This causes you to get the Venturi effect, sucking colder air into the exhaust through any cracks. The Venturi effect states that in a situation with constant mechanical energy, the velocity of a fluid passing through a constricted area will increase and its static pressure will decrease. Flow through the exhaust isn't constant, over the course of a second or three you can see that it oscillates in time with the ignition pulses. But over very small time slices (during a pulse, fraction of a second) it is relatively constant. So in *that* context those pulses increase somewhat in velocity as they travel (assuming pipe diameter is constant, which it isn't but bear with me), and that causes enough gases to be sucked INTO the exhaust from any cracks to make a difference - at least, to make a difference to an oxygen sensor, which although it's bolted like a bung into an exhaust is actually a precision instrument and is going to measure the oxygen in the atmospheric air that gets sucked into the exhaust from, say, the area around the sensor that's not sealed.
Oh, it's all so obvious (not).
Exhaust gas is first pushed by the pistons during startup. Each piston's pulse travels down through the exhaust pipe and actually pulls the gases that are behind it by vacuum. The car's combustion does most of the work to get the hot gases out *but* a significant portion is moved by the lower-pressure-cooler-gas zones between hot pulses -- this is generally called "scavenging." The scavenging process is a cylinder's volume of air travels through the pipe, afterwards which the next firing cylinder's air follow it, and the process is repeated over and over again, making these hot pulses that suck a good bit o wind behind each one. It's this vacuum that would be pulling in the relatively cold air from the engine compartment past the front O2 sensor into the exhaust - not only disturbing the scavenging but also spazzing out the sensor - flatlining it like I was seeing on the scan tool and when done long enough (coasting at high speed downhill) sucking enough air to set the P0171 (system too lean) code...
The confirmatory clue for me was that when I pulled the front O2 sensor, it came out way easy and the area where the O2 sensor's metal gasket should have been touching the bung port surface was black with soot. Exhaust was blowing out past the O2 during pulses and it was sucking wind in between each pulse. Coasting downhill, the pulses get smaller - so much so that it set a System Too Lean code. Even though the system wasn't actually too lean. The ECU then lights the idiot SES light, and goes into a more or less default mode where it uses various programmed-in assumptions to control combustion.
Removing the O2 sensor I also noticed that it was similar to but not exactly like the Denso that Rick sent me to install. The Denso has a longer cord, which enabled me to mount the sensor wire away from other hoses and hard parts and places the connector where it easily can be seen, and a slightly different geometry. I cleaned the rim around the bung hole (it was covered with a fairly thick layer of soot so I stuffed a corner of a shop rag in then sanded the rim clean with emery cloth) and installed the Denso.
So far so good; I'll know for sure after this weekend's drive when I go to visit my Dad. But as it stands I think the P0171 initially was there because of the exhaust manifold crack, then the gasket leaks, then the wack sensor.
Or maybe it was just bad gas? That happens around here...