What exactly is your roommate experiencing? The battery leaking? The engine dying?
I need more info.
GiantRat, I had exactly the same symptoms with my motorcycle (2001 Yamaha YZF-R6). The first time it happened, the tach needle started jumping all over the place and a few moments later the engine stalled. And I coulnd’t restart the engine because the battery was dead. When I got home I realized that the battery had leaked. I thought it was due to a faulty battery, so I replaced it, but then I had the same problem with the new battery too.
Then I started checking the atlernator. Winding resistances nominal, voltage output nominal. That was strange. I took the motorcycle for a ride and when I got to the highway I notice the tach needle to start moving erratically. I immediately turn off the lights and slow down and manage to bring the motorcycle back home. I grab my multimeter and measure the voltage across the battery terminals: 9 volts and dropping and the battery is hot to the touch! I remove one cable from the battery terminal and the engine stalls.
The problem was the rectifier. I later found that this is a very common problem with this motorcycle. The battery leak was because when the alternator stopped producing power, all the electrical power was provided by the battery alone. That caused the battery to overheat and spew electrolyte.
Sorry. No, no battery leaks - I replaced the battery, harness and wiring when I had the earlier problem before selling him the car. The engine will die (only at low speeds, I believe) and sometimes just sit dead for 10 minutes before miraculously being fine again. This is so similar to the previous problem that I had that I’m inclined to suspect some association between the two (although I suppose their could be a fuel line obstruction).
I just spoke to my roommate on the phone and he says that the stereo and fan are still working, so I’m wondering if it’s an electrical problem with the engine not getting enough juice to keep firing (but still enough for accessories), or if it’s more likely an obstructed fuel line or something. And I’m wondering if I even have a right to wonder, being a less-than-amateur mechanic.
BTW - you should be billing us for all your consultations!
GiantRat What you just described to me does not sound battery related. If it was battery related, I would expect the entire car to be dead.
Let me see if I have this straight. Your roommate is driving the car, and it dies, and may stay dead for up to 10 minutes. During that time the lights and radio work is that correct?
Will the starter crank the car during this time? (the WRRR sound when you turn the key to start)
If the answers to my questions are lights and radio work, and starter cranks, but the car does not start, then I am very sure it is not a battery/alternator issue. Much more likely it is a problem a heat related open circuit probably in the ignition or fuel injection system. What happens in these conditions is as a component heats up, it fails when the temp exceeds X degrees. As long as the temp is <X the component is fine. When it does fail, the engine dies, and once the engine has cooled off, the system regains normal operation. Things like control units, relays, RPM sensors, engine coolant (ECT) sensors, harness connectors, fuel pumps, and yes even a fuse are some of the parts that can fail in this manner. Anyone of these failures can cause a stall for a period of time and then a restart and run fine condition.
So which part is it that is failing? From the information given, pretty much impossible to tell.
I once drove a customer car for two weeks* before the problem got bad enough and lasted long enough for me to be able to diagnosis it. Did I mention that this type of problems give technicians gray hair?
The good news is with a 10 minutes failure window, there is time to do some diagnosis before the car heals itself. You roommate can either find a good shop to do some serious diagnosis, or you can start throwing parts at the car.
*No not non-stop smart ass, to and from work and on weekends
Completely unrelated, but Rick is obviously an automotive GENIUS and I’m tearing my hair out, so, have a hijack.
Oh Great God of the Horseless Carraige, aka Rick: in an earlier post, you said:
Can you point me in the right direction for finding this sort of stuff, or failing that, could you possibly point me in the right direction for finding the O2 sensor on a '02 Ford Focus?
See, I’ve got a new one in hand, and that’s what the computer sez is busted (well, it’s reporting that the mix is rich, which means the O2 sensor is a likely culprit, according to a mechanic acquaintaince who hasn’t answered the phone for two days)… but I can’t seem to find the old one. If I knew where to begin to look, it’d help… :smack:
It has to be in an exhaust component. The upstream sensor (the one you’re looking to replace) is somewhere in the manifold or piping up to the front of the catalytic converter. The downstream sensor is usually at or near the rear of the converter.
You may discover that finding it is the easy part. On some designs, access is such that a special wrench is needed to replace it.
A likely culprit? Okay. The most likely culprit? I would say no. There are any number of malfunctions that would actually cause the engine to run rich, or on some designs fool the sensor into thinking it’s running rich. You may well have a problem of this sort along with a perfectly functional oxygen sensor. The majority of cases I see with “rich” codes have a fuel or emission system problem, with the oxygen sensor accurately reporting the situation.
Now, it’s also certainly possible that the sensor is faulty. It’s just that replacing it is a rather pricey way to test it if your guess is wrong.
Oh, great. The new one looks like a standard wrench will fit around it; I haven’t tested that, though.
Well, I originally thought it was the spark plugs causing it to run ragged – and there was no question they needed replacing, because, yikes, the tips were horrid and one of them was apparently not firing all the time – but replacing those hasn’t completely eliminated the problem. It helped a lot, but it still is … well, basically, it sounds like a muscle car, instead of a Focus, but not in a good way.
Everyone I’ve talked to says it’s the O2 sensor, since the plugs weren’t the sole cause. Admittedly, only one of them is an actual mechanic and he only plugged in the sensor and listened to it for a couple minutes; he’s the one who’s been too busy to get back to me and I can’t pester him about it because I can’t afford his services anyway. 
I guess I’ll cross my fingers and hope the sensor fixes it.
At any rate, is there anything I, as a untrained-but-good-at-following-instructions amateur, can do to pin down the problem more accurately? I’m not afraid of working on the thing, I just don’t know enough about what’s going on in there to diagnose it myself.
Thanks for the help!
I’m willing to bet that it’s the wiring to the fuel pump cut-off switch in the trunk. At least that’s were it was in my 80s vintage Taurus that had the exact same symptoms. It took a few days to diagnose this and the repair was easy. Have someone check it out. The rear struts may be going which then causes wear-and-tear on the wiring harness back there.
Yes, it’s a 7/8" hex. I’m talking about an access problem getting a wrench on the old one in place in the vehicle, coupled with the necessity most of the time to use a lot of force to break it loose. An open-end wrench ain’t gonna do it. Box end wrenches usually won’t slip over the wiring connector. You might need a special oxygen sensor socket (has a slit for the wiring to go through) or something like a crowsfoot flare nut wrench. Or if you’re lucky, you might not.
First, look at a spark plug and see if the inside insulator is dark. That would indicate that the engine is indeed running rich, and suggest that the sensor is okay. Alternatively, if the insulator is white, that would indicate running normally or even lean and that the sensor is faulty.
If you have a voltmeter that can meaningfully read in the 0-1 volt range, you can test the sensor. The instructions are a bit complicated and I’m out of time, but if you’re interested I can go over that later.
Crud. Here’s hoping I’m lucky, then.
I suppose I could try making a DIY O2 sensor socket with a spare socket and my trusty Dremel, but then again, that thar’s tool steel… I think it’d win against the Dremel.
See, this is good to know. As it happens, I have one of the spark plugs that I removed, and the insulator is still white. Not pristine-looking, but nowhere near dark.
When running lean vs. running rich, which will cause the rumbling I’m-gonna-stall-on-you-HAHA-fooled-ya bit when idling, and no problems at all when actually driving on the highway, etc. (whenever I’m giving it gas, it runs fine). Because my very first thought when this problem started was, “hey, that sounds a bit lean”, but that was back before more knowledgable types told me it was the spark plugs (which did need replacing, I admit), and anyway was based on my limited experience with model airplane engines.
I’d love it! As it happens I do have a meter, and IIRC it reads fine in that range (I used to repair TVs and VCRs, which have a lot of low-voltage components). And hey, the more I know the more dangerous – I mean, skilled – I am, right?
Until Gary T gets back, since you have a DVOM and know how to use it let’s get started.
With the key on, engine off there should be about .45-.50V on the O2 sensor signal wire, when measured to a good ground (battery or engine). Your sensor probably has either 3 or 4 wires going to it. One of these wires should have the voltage I mentioned on it. That is the signal wire back to the computer. While monitoring the voltage, start the car. After a short period of time, you should see the sensor start to oscillate between 0-.8V (or near that) When the voltage goes to zero, the exhaust has a lot of oxygen (lean) and the computer enriches the mixture. When the voltage goes to .8 the exhaust has very little oxygen, and the computer leans the mixture.
There are three possibilities here. (do these tests on a warm running engine)
1) O2 sensor switches 0-.8-0.8-0-.8 regularly and fairly quickly. O2 sensor operation is normal.
2) O2 sensor is stuck at 0V Sensor is reading lean, and sending a request for more fuel. The computer is enriching the mixture. On some cars if the mixture moves far enough, a fault code may be set. On some cars, the mixture can move so far, the car runs like a dog.
3) O2 sensor is stuck at 0.8V The mixture is too rich, the O2 sensor is reporting this and requesting a leaner mixture. Again the computer will try to accommodate this request.
Fault tracing steps:
Condition 1: Nothing to fault trace, this is normal operation
Condition 2: Check for an air leak anywhere between the mass air flow meter and the engine. Remove all plastic trunks and examine for holes. Check for loose/missing vacuum lines. Fix.
Condition 3: Possible bad fuel injector, mass air flow sensor, engine coolant temperature senor bad, or poor connection. Let’s see if Gary can give us the resistance specification for the engine coolant temp sensor (ECT). This is a temperature sensitive resistor that has it resistance fall as the sensor warms up. If it goes bad, or the connector has resistance in it, the system will think the engine is cold (puts in more fuel) than is needed.
Don’t bother to try, either you won’t be able to cut the socket, or it will be too weak to handle the stress. Buy a O2 sensor socket at a parts house. Clean the threads with a wire brush, and use a dab of anti-seize on them when you reinstall the sensor.
My best guess from what you have posted is that you have a problem with one of the following:
[ul]
[li]Mass Air Flow sensor (MAF)[/li][li]Engine Coolant Temperature sensor (ECT) or its connector[/li][li]fuel injector with a bad spray pattern, or leaky[/li][li]Manifold Absolute Pressure sensor (MAP)(If equipped)[/li][li]O2 sensor[/li][/ul]
On preview I just remembered: does the temp gauge on the dash read normal? If it reads low the thermostat could be bad causing the engine to run rich and set a code.
My instructions will be different from Rick’s, because I’m different from Rick. 
Don’t worry about apparent minor contradictions, just use either or both in a way that makes sense to you.
First, find the oxygen (O2) sensor and its 4-wire wiring harness. Then, find the signal wire. According to my info, this is the white one (the others are red, black/yellow, and brown). These colors will be on the vehicle wiring harness leading up to the sensor connector, and probably not on the sensor’s harness. Then, connect to the signal wire. The least damaging way is to insert a thin probe (e.g. a hatpin) into the connector so that it touches the wire’s terminal. Alternatively, pierce the wire (hatpin works here, as well). The insulation is nominally self-healing, but piercing does run a risk of water ingress over time. Meter ground can go to the engine or chassis.
The other three wires are sensor heater (12 V w/ ignition on), heater ground, and signal ground. The signal wire may or may not have voltage at rest. Once the engine is running and the sensor has warmed up (1-2 minutes should do), the sensor will generate voltage ranging from zero to one volt. Usually, readings will be in the 0.1-0.9 V range. Low voltage = lean, high voltage = rich.
Fluctuation is a critical aspect of O2 sensor operation. The voltage will normally cross the centerline value of 0.45 V at least once per second. Some vehicles will show fluctuation at idle, others need to be revved to 1500 rpm or so to show it. A sluggish sensor will fluctuate too slowly, maybe once every few seconds. A stuck sensor may not vary more than 0.1 V.
If you have an oscilloscope, that’s probably more helpful than a VOM. Here’s a typical scope pattern, at the bottom under Lamda sensor (= O2 sensor).
If you have normal range and fluctuation, the sensor’s probably okay. If it’s sluggish, it’s bad. If the reading doesn’t change much, it’s either a bad sensor OR an accurate sensor telling you that something else isn’t working right. If it reads rich, induce a vacuum leak for several seconds or so and see if the sensor responds. If it reads lean, add extra fuel (e.g. spray carb cleaner) to check response. A sensor that responds quickly and that is capable of full range (e.g. 0.1-0.9 V) is almost certainly okay. If the sensor seems okay, it’s time to look elsewhere for the cause of the rich code.