96 850 Turbo - P0172 Won't Go Away

There's not much else that can cause this code. If the air mass meter has been changed, and you're sure there's no vacuum leaks (intake manifold gasket, injector seals etc.) then one other cause can be the fuel pressure (weak pump or defective regulator). And if there's any leak in the exhaust between the engine and oxygen sensor, air can get sucked in tripping this code.

Some garages have a machine that generates smoke. This is connected to the intake, which is slightly pressurized, and then the leaks are obvious where the smoke leaks out.

Also, weak fuel pumps generally give a symptom of extended cranking before starting, especially when the car has not been used for more than

An OBD tool that captures data may help - it certainly can't hurt.

Even the one behind the power steering pump that you can hardly see?

Any unmetered air entering the intake stream after the mass meter will drive the system rich, i.e. intake manifold gasket, crack or split in the intake bellows hose, exhaust crack or leak near the O2 sensor.

Possible cam timing issue.

When checking/replacing the vacuum hose on the fuel pressure regulator was there any whiff of gasoline? Any trace will justify condemning the regulator. Hang a gauge on the fuel rail, the regulator can stick intermittently, and see if you can catch it raising the fuel pressure when the MIL comes on.

Make sure the coolant temperature sensor is sending a valid reading to the control unit.

Only if you can identify the bad reading from a component you already suspect. Typically what you'll see is a collection of all the bad data that were present when the code set. IOW, O2 readings will be rich, long term fuel trim will be cranked all the way lean, short term fuel trim may be adapted lean as well, injector dwell or frequency will be high, manifold vacuum will be low, barometer may be inaccurate, MAF signal will be rich, load signal may or maynot agree with vacuum readings depending on throttle position switch interpretation, and so on and on.

Mixture codes are more typically from the electronics reporting a mechanical problem.

Bob

Can't a failing oxygen sensor cause that code? And surely a flaky engine computer could cause that code or nearly any other code.

The ECU can tell the difference between "too rich" and a failing sensor by the way the sensor works. When cold (or completely dead) the sensor appears as an open circuit, and the .45 volt bias from the ECU is seen on the sensor line as a good indication the O2 sensor has gone bye-bye. When the sensor is working and the mixture is too lean, the output is around 0.1 volt, and when the mixture is too rich it is around 0.9 volt DC. The ECU uses that information to dither the mixture rapidly, slightly rich to slightly lean and back again, hopefully 7 transitions per second or more. The sensors don't usually die altogether, they just get weak and sluggish so the sensor is slow to respond.

The ECU could produce the code by being insane, but that would be a pretty unusual failure. ECU failures are rare enough, and the ones that aren't caused by the ECU getting wet tend to be the power handling areas going bad.

If the code is "real" the ECU is saying that it is trying to keep the mixture under control by slaving to the O2 sensor transitions, but the "ded reckoning" part of the ECU is saying there should be more fuel. This is often called being "out of trim."

This brings us to Mike's observation that the fuel pressure being high can cause the code. The ECU calls for a certain amount of fuel to be injected, but the high pressure forces more fuel through than expected... too rich. The rest of the possibilities seemed to me to be backward - causing the O2 sensor to think the mixture was too lean. But I get confused about things like that lately.

But I do wonder if an injector could be leaking into the manifold or slow to close after a while. I am reminded of the mistake I made in buying aftermarket injectors for our 85 765T. The first set was the wrong ones (they were for a non-turbo) and the second set also failed to make the engine run decently at all... barely idled when cold, perhaps 20 hp max. I eventually took it to the dealer where it was run through the tests and parts swaps. New AMM didn't help - the mixture was still way rich. The injectors flow tested right on the money, but when Volvo injectors were installed the AMM adjusted right up. Something wrong with them dynamically, I guess.

Mike

I have to agree, I don't see how a vacuum leak would result in a "system too rich" error, but it does. I suppose it must have something to do with the way the system reacts to small "test changes" in injector pulse width, or something like that.

Interesting - do both intake leaks and exhaust leaks do that?

Mike

Well according to the book they do, but I've never personally experienced an exhaust leak that set a mixture code. I have seen intake leaks that set codes for both too rich and too lean. And on my S70, the infamous rubber elbow behind the power steering pump was bad, I noticed it without any codes being set.