Car Problem

We have a 2000 BMW 323i, with 180,000 miles on it. I’ve been doing a lot of work to get it running properly, including new plugs, and replacing a fuel injector and cleaning the rest. It’s no longer throwing any codes, but when I look at the fuel trim, it’s running very negative (ranging from -12 to -25, depending on RPM) on Bank 1 (cylinders 1-3). The other bank is much better, running between -3 to -6). This would indicate that it’s running fuel rich, but the O2 readings before and after the catalytic converters (one for each bank) are looking normal, and indicating that the cats are in good shape (i.e., flat output on the outlets, indicating that they’re eating the stuff they’re supposed to).

If it was both banks, I’d suspect a fuel overpressure or something, or maybe clogged intake, but it has a new fuel pump and air filter, and in any event, those would cause problems in both banks, not just the one, since they’re common to both. Any ideas?

[Saturday-afternoon update]

Welp, I hope the problem was a bad oxygen sensor on Bank 1, because I broke the wires on it when I pulled it…

[March 31st update]

Well, I replaced the oxygen sensor that I broke, and it’s showing the same thing. -18% at idle. So it wasn’t a sensor issue.

Hard to know what the problem could be, since the two banks differ only on the exhaust side, not the intake. All I can think is a bad fuel injector on one of the three cylinders, but only way to test that would be to swap them out, and it wouldn’t narrow it down to a single one, unless I did them one at a time, which would be a royal PITA, because it’s just as much work to change one as it is to change all of them, given that I’d have to put it back together each time to test it. But maybe I could start with the old ones, and leave the new one (on #1) for last.


[Friday-afternoon update]

A new weirdness: I was seeing a misfire on cylinders 5 and 6 (bank 2). When I pulled the coils, they came out wet. There was a bunch of water in the spark-plug wells, though 1-4 were dry. I stuffed paper towels down into them, and soaked it up. No idea how it got there. But the car’s running pretty well now. I’ve giving up on the fuel-trim issue for now, though maybe I’ll take it over to my local Bimmer mechanic and ask him what he thinks.

28 thoughts on “Car Problem”

  1. If you don’t hit upon an easy fix soon, and if it’s not a hard to do so on that vehicle, you might want to look at swapping the sensors on bank 1 and 2, to see if you get the same results, or, it’s a sensor issue.

    1. Yeah, that’s a good idea. I could look at the sensors at the intake. They’re replaceable, so they shouldn’t be hard to swap. It’s just that you have to take a few things off (like the cabin-air filter) just to get at them.

  2. Do I have this right – the ECU thinks it needs to run Bank 1 extra rich, but the end result downstream is normal exhaust O2 and cat behavior?

    Sounds like the ECU is successfully compensating for a physical problem with Bank 1 fuel injection. (Via longer injector pulses?)

    Could be a partial clog in the Bank 1 fuel line, could be a marginal ground (or driver circuit?) reducing power to the Bank 1 injectors.

    Though I once ran into a similar problem where it was a bad pulse damper in the fuel line, leading to “water hammer” that at certain RPM’s coincided 180 out-of-phase with the injector pulses. Ran the side of the engine closer to the failed pulse damper dangerously lean for a while – the pressure pulses damped out enough in the additional tubing length across the engine so the other side of the engine was OK.

      1. There is a rival to Scotty Kilmer who calls himself The Car Wizard, who has a video about a guy who drove all the way from Colorado to Kansas to have this mechanic with a Z-Z-Top beard cure his car.

        The problem was with a high-performance Audi, and I think the simple solution was to replace an oxygen sensor, even though the code didn’t indicate “oxygen sensor.”

        Look for this video and listen to this guy who has a pretty high opinion of himself because he offers more detail and cautions than I am offering here.

        Scotty also warns that in replacing oxygen sensors, you want to go “OEM” rather than the slightly cheaper substitute.

      2. I couldn’t find that Car Wizard episode, but I think it was an O2 sensor acting goofy but throwing every other code but an O2 sensor code.

        Do you have a scan tool that let’s you watch the waveforms on the fuel trim — that helps diagnose a condition with the O2 sensors.

        With two banks, you also have to identify the correct bank to pull the O2 sensors.

        O2 sensor can be a pain to replace crawling under the car. An O2-sensor wrench socket can help — not expensive, and maybe a long wrench handle because they can bet baked on.

        Scotty Kilmer complained about a Honda requiring unbolting the front passenger seat to get at the pigtail. Even on a Taurus this was a problem because I had to work “in” a long screwdriver to press on the connector tab that I could pull the connector halves apart.

        If there is a reason you need to reconnect the new sensor before screwing it in, there is a trick to “pre wind” the cord so it isn’t all twisted once the sensor is screwed in.

      3. This is an inline six, then? (Else “one fuel rail for all six cylinders” would still have two halves with some sort of cross-connect.)

        So, what is NOT shared between the two banks, but is shared within a bank? As mentioned, I’d look into a possibly separate Bank 1 injector driver circuit and/or that circuit’s ground connection.

        1. Yes, it’s an in-line six. I don’t know why they have two separate banks, but as far as I can tell, they’re only separated on the exhaust side. Everything is shared on the intake side. There is only one pressure-relief valve on the fuel rail.

  3. Your O2 sensor voltage, once warmed up, should not be flat, but somewhat sinusoidal-ish as the car flips from lean to rich to back again. “Flat” suggests bad O2 sensors.

    My bet, however, is a dirty or warn injector on the rich side that is leaking fuel inot one cylinder. If you have the right software you should be able to shut off one injector at time and see the difference….If my theory is correct (and I’d bet it is), you will notice the bad injector quickly.

    1. The pre-cat O2 voltage is varying, as it should. The post-cat is steady, as it should be. I just cleaned all the injectors, and replaced one of them, a couple weeks ago. I’m going to try swapping the sensors and see if the problem shifts to the other bank. If so, then it’s the sensor.

  4. Welp, I hope the problem was a bad oxygen sensor on Bank 1, because I broke the wires on it when I pulled it…

    Isn’t that more or less the story when removing an O2 sensor? I’ve heard from mechanics plan on having a spare on-hand for everyone you remove.

    1. I’ve read a lot about it. That guy’s problem was positive trim, which is usually a vacuum leak. Negative is a lot harder to run down. I’m going to start swapping injectors tomorrow. I can’t imagine anything else that would cause a problem on one bank, but not the other.

        1. No obvious problems. My concern is that if it’s running rich on a bank, it’s a) washing the front cylinders (or at least whichever cylinder is the problem) with gas, inhibiting lubrication and b) getting reduced fuel mileage. That’s why you look at fuel trim: To identify issues that aren’t obvious merely from performance.

  5. I’d go with the bad fuel injector theory. Sounds like one is leaking by. Depending on what is more valuable to you (time or money) two more new injectors might be money well spent.

  6. Have you tried testing the exhaust temperatures with an infrared meter? If you are truly running rich (or lean for that matter) the exhaust manifold temps should show this discrepancy.

  7. I have to say, you’re making me reconsider my plan to convert my 1974 Lotus to fuel injection…

  8. I had a funny problem with O2 sensor once and it turned out to be a head gasket. A small leak had developed from a coolant passage to a cylinder. It’s simple to diagnose, most shops can do a pressure test on the coolant system.

Comments are closed.