'95 LeSabre Dies Without Warning

[timrice]

Yesterday I got the timing light on it after it died. When I cranked, no light.

Dug the ECM / PCM out of it tonight. I checked the ground wiring there near the battery where it attaches to the frame. It does look pretty cruddy. There are three black wires that are under that nut. The nut and top ring terminal are rusty-colored, but snug. I'll clean that up tomorrow and maybe take it for a drive. I have one of the two main connectors detached at the firewall - maybe I can check resistance to ground from that connector before doing any cleaning. If it's more than a couple tenths of an ohm, that'd be abnormal in my opinion.

Found the fuel pump relay, too. After I clean the ground post up and if need be, try another ECM, then I will pop a new fuel pump relay in there while all the guts are exposed (though I doubt it's fuel-related at this point). As impatient as I'm getting with this, I still don't want to try more than one thing at a time so if/when I do actually fix it, I'll know what it was.

To answer pcmos' question, I read the codes out w/my brother-in-law's "X-Ray" scan tool. ($$, so I hear) Can't just stick a paper-clip in the ALDL connector like the older cars. The only code we've ever read out is P0341 - camshaft sensor circuit. It doesn't always set a code when it dies. It doesn't always turn the check engine light on either. So that code was what led me to camshaft sensor replacement early on in this ordeal. I feel foolish for blindly reacting to the code. (But not quite as foolish after professional mechanics did the same thing later--in their defense, their electronics guy was on vacation, and to their credit, they didn't charge me for it...) I hate "false codes." As complex as these controllers are nowadays, there's bound to be some occasions where they lie. I think the 341 sets because of the engine stopping, instead of the engine stopping because of the 341 code.

Could it be that the ECM is losing power / ground, the engine dies, and then the ECM wakes back up sometimes (because of this bad ground theory or because the ECM *is* whacked), and because the engine is dead at that point the ECM doesn't see the cam position in the right place like it expects it to, and therefore the 341? And maybe likewise, on those times the ECM *doesn't* wake right back up, those are the times it doesn't turn the CE light on and set the 341?

ECM / PCM P/N for this car is 16183247.

More news as it breaks. (News, not the car.)

[pcv]

Timrice,
I do not see anything about replacing a crank sensor. Next time the vehicle stalls, have a gallon or two of water in the trunk, pour it around the crank sensor, located next to the crank shaft pulley. If the vehicle starts immediately, replace the crank sensor.
Good Luck.

[timrice]

pcv: I did try the "crank sensor cool-down" thing. (See post #12, 6/8/04) I didn't use as much water as you suggest, admittedly, so maybe I didn't give it a fair chance to work. You probably know of it, but I saw in another forum where a guy was suggesting to someone else to use their windshield wiper fluid pump to put the water on the sensor as you're going down the road. He put some extra hose on the outlet, duct-taped the hose in place, filled the reservoir with ice water, and drove down the road. When the car would start running rough, he'd hit the wiper spray switch, and if it cleaned up the rough running, he knew he needed a new crank sensor. I thought that was a pretty cool diagnostic method, though it seems out of the ordinary.

So far today, I have checked my grounds - didn't find anything I'd consider high resistance, but I thoroughly cleaned the ground post on the frame next to the battery and all four ring terminals that come off it. Those babies shine. Haven't got a new ECM / PCM yet - I put the original on the RH floorboard and hooked it back up. It's sitting outside idling. If it dies, I will get a new one and install it.

One thing I noticed...maybe it's not significant, maybe it is: the oil pressure indicator moves around at idle. It will sit at 50 psi for a couple seconds, then it will wiggle between 50 and 60 psi for a couple seconds, then drop back to 50 or so. It seems erratic. For all I know, it's normal.
This fluctuation smooths out when I drive the car.

Maybe cleaning up the ground is all the ECM needs to do its job. Got my fingers crossed. I will go change the oil in the other car while I wait to see if this one dies again.

[timrice]

I am not totally convinced it's fixed just yet, but:

Yesterday after cleaning the ground terminals up, it idled for five hours straight, and I then drove it 25 miles, and no failure.

And today it got me to work and back--no failures.

I will be fairly convinced if it goes the rest of the week. It hasn't logged even two good days in a row in two months.

[pcmos]

Yeah I was actually the one that posted that message on the other thread about using the windshield washer pump to get some cool water on the sensor. Actually I was talking more about using this method for diagnosing a bad cam shaft position sensor, not the crank sensor, just because the crank sensor is monitored more carefully by the PCM. Usually even intermittant failures on the crank sensor will trip a code in the computer, but the cam sensor is not monitored as carefully. I actually was able to diagnose my cam sensor on my old 92' LeSabre by using the washer pump to dump some cold water on it while driving, but this is because when it was acting up in the driveway one night I poured some cool water down behind the pulley's and got the car to run. Since there are a couple of sensors that were hit with the cold water, I decided to rig up the washer pump to get it more specific, and also to use it and drive the car at the same time so I could be sure. First time it started acting up, I gave it a quick shot, and boom, the car started running perfectly normally within a quick moment. This was enough to convince me to take a gamble at replacing the cam sensor and I never had a problem again.

[rustbucket]

Sounds good so far. In case get back in the parts replacement mode, Under 400 RPM(that is when cranking) the spark is controlled by the crank sensor, the cam sensor and the ICM alone.(not the ECM). Above 400 the ICM tells the ECM to take over the spark, and the ECM tells the ICM when to spark.

[pcmos]

Yes, and if you read the manual you will see that after cranking and above that 400 RPM mark, the PCM monitors the presence of a signal from the cam sensor, but it will only set a code if that signal is gone for a full 5 seconds. This is the loop hole in GM's code that allows the cam sensor to fail in an intermittant manner without setting a trouble code. The other problem is that the cam sensor produces a simple on-off series of pulses, and unlike the crank sensor, the PCM doesn't compare this signal to a known reasonable signal. So if the sensor misses a pulse or two every few cycles of the engine, the computer will still try to use the reading anyway, which is what caused the driveability problems on my old 92'. I can only assume that the computer was using the faulty signal. I suspect this because if the computer were generating a random cam signal to use, it would only get it right about 1/6 of the time regardless of engine RPM, it would run terribly at all RPM's in this case and I can't imagine that the computer could be doing this for more than 5 seconds and not count that as a loss of signal from the cam sensor. Since no code ever set, and since the engine was having trouble only at higher RPM's, I made the assumption that the PCM was actually trying to use a poor or intermittant cam signal. As the engine RPM's increased the percentage of bad pulses in a given time period increased and the problem became noticable, but at lower RPM's the cam sensor didn't have to get so many pulses correct in the same period of time because the engine and sensor magnet weren't turning as fast. One way or another, however, replacing the cam sensor most definately fixed the problem, and my gas milage went back up to like 25mph highway.

[pcmos]

I'm sorry, I don't mean to blab about the cam sensor, I just realized that this is completely off topic in this post. But I am curious to know how this problem turns out with the "95 LeSabre dying without warning". This really doesn't sound like a cam or crank sensor problem to me at all, just because of how quickly it dies out. Sounds more like a power distribution / ground problem, but it will be interesting to see if cleaning those terminals helped out.

[timrice]

No apologies needed as far as I'm concerned, pcmos. I'm glad to learn all the extra details. I know some people get frustrated if they have to read more than a couple lines, but without the details, what's the point of even sharing a forum, I figure.

Now I must say that rustbucket has me worried again with what was said in Post #23 about the PCM having nothing to do with spark control during cranking.

The ICM *was* replaced about a month ago by my local shop in their first effort to fix this problem. I guess there's nothing that says that I couldn't have gotten a bad replacement. But I'd have thought my mechanic would have immediately load-checked the one he installed after I brought it right back in complaining of the same symptoms he originally told me were cured by replacing the ICM on my prior visit. He seemed awfully sure the original ICM was bad. Yet after I took the car home from the shop with the problem unresolved, the symptoms continued through last Friday. If the original diagnosis of a bad ICM was *truly correct*, I wonder why the shop didn't say the replacement ICM was bad too and that I needed another one?

Here I'm thinking I have it narrowed down to the PCM or its ground, but based on what rustbucket says, how can the PCM be at fault when it dies and won't start back up right away? It seems reasonable to me that if the PCM didn't have a stable ground it could cause the engine to die, but if the engine doesn't even *need* the PCM to start, how does it add up that the PCM and its ground should be suspect at all? If I unhooked the PCM and put it in the back seat, could the car still be started?

By the way, no failures again today. That makes two straight work days, plus almost 6 hrs engine time on Sunday. This streak is unprecedented. The last thing I did was clean up the frame ground post and ground terminals. Yet I do not dare claim victory prematurely.

[pcmos]

Well, let me double check my shop manual for my 1992 and my 2000 LeSabre's and make sure I am correct on this. But I believe the PCM unit is required to initiate the startup sequence. I am pretty sure that it is the job of the PCM to turn the control over to the ICM before it even fires the first spark, and it may also be the case that the PCM is designed to take a few moments to reset if it senses a bad power supply. Once again I will double check this later, but I am also pretty sure that there is a ground post for the ICM module itself under the alternator, which you may want to clean up as well. Frankly I think you got it fixed by cleaning those PCM grounds, but one must be cautiously optimistic in this tricky business.

[rustbucket]

I said the ECM does not control the "SPARK" below 400 RPM. It still controls the fuel injectors and probably the ICM gets the supply voltage through the PCM in some way. So it couldnt START without the ECM working, but the timing light should still flash (of course if all the power and grounds were right). When your timing light test did not work, that leads me to the cam sensor, crank sensor and the ICM, but not the ECM. (again assuming all the connections were good)

I just did a test on my Lesabre. I unplugged the ECM at the battery, and put on the timing light, and cranked it, and the timing light was flashing. (I was wondering if what the manual told me was correct!)

[timrice]

When I used the timing light during cranking last week, I got no flashing. So maybe this thing is caused not by loss of the PCM/ECM's ground, but possibly by the loss of the ICM's ground? I cleaned four ground terminals on the one fender-mounted post--maybe one of them *was* for the ICM, I'm not sure. I did not realize the ICM ground might be under the alternator, nor have I checked anything under there. If that's where it is on *my* car, then the ICM ground has not been messed with at any point in this story.

If it's an ICM or ICM ground thing, I don't see how the car can possibly be truly fixed at this point. Based on rustbucket's test (very cool of you to do that and share the results, by the way), if you can unplug the ECM/PCM altogether and still get a spark during crank, then it sounds like a non-functional ECM/PCM could *not* have been the cause of my missing spark.

Only thing I can figure is that even though I might not have realized it, the ICM ground does run to the fender post and I actually did clean it.

So I'm still confused: *IF* it's really fixed (and that remains to be seen), then what explanation allows *all* the facts in the case to fit together and not contradict one another? Guess I need to know just exactly where all four of those grounds actually go...

I don't want to drive anyone crazy with all my questions and handwringing, but I sure would like to be able to go back to the shop and explain to them exactly why this happened and what fixed it and be able to back it up.

No failures today either. Cautious optimism is the status.

[BReynolds]

Tim,
I am a student at Nascar Technical Instutite, I have a 1987 Buick LeSabre with the same problem and I have been informed that the most likely cause is the Crankshaft Position Sensor (CKP), this is what produces the signal to the ICM which controls the ignition and may not produce a good signal if it is dirty or misadjusted. The Camshaft Position Sensor (CMP) will just make the car back fire and lower fuel econmy because it only controls the sequential fuel injection. I have not put a new CKP in it yet because I was just informed of this today. But i'am going to put a new one in tomorrow and let you know how it works. If you find anything else please let us know.

Thanks,
Brian

P.S. There are no TSB (Technical Service Bullitins) from AllData on this problem.

[timrice]

During the recent two weeks that my car was in the shop, the technicians had put a portable two-channel graphic-readout meter on both of the crankshaft sensor outputs. Both signals were clean, at least during any time they were being monitored. At no point have I seen any crankshaft sensor-related codes, though that may not mean much one way or the other. I do have a known-good crankshaft sensor left over from a diagnostic effort on my old/sold '89, so if it comes to that, I could put it on for little or no additional cost as long as the same sensor works for a '95. I don't enjoy taking that pulley off to access the sensor, though.

[pcmos]

Well lets assume that something you did under the hood either fixed the problem or improved it, or changed its behavior (being cautiously optimistic). I went onto ebay and found some pictures of 1995 LeSabre's for sale. Lots of people take pictures of their engine when they place a car for sale on ebay, so I got a good look at the engine compartment in a couple of different 95's. From what I can see in the picture, the layout looks identical to my 1992. I think there is a good chance that as long as you do not have the series II version of the 3800 that the wiring would be the same under the hood at least. Ok, so in order to follow a diagnostic process here, I need to assume that the ground distribution is the same on your car as the way it is described in my manual for my 92. I remember working on the ground post on the fender as part of my sweeping goal to clean ALL of the ground posts so that I could completely eliminate a ground problem. Technically speaking a faulty ground could cause an interruption in another circuit that would lead to a cascading failure. Although this is not likely I made the diagnostic assumption when I worked on mine that it would be possible, hence the idea of cleaning literally all under-hood grounds.
Now to start with, it is important to note that ALL grounds, literally everything that is grounded on the car depends on the negative battery terminal making contact. This car WILL NOT operate at all without the battery connected, which means all power is derived from battery power, and even if the ground on the alternator is perfect but the ground on the battery terminal connector is bad, you'll have a problem. That said, lets move to the "fender ground" which is referred to as G103 in my shop manual, G103 is actually fed by one single wire that jumps over to the fender from G104. G104 is located under the coil pack in that general vicinity (poor diagram), but I remember working on both G104 and G103 on my LeSabre. So basically there is another bolt sticking out under or around the bottom of the coil pack just like the bolt you cleaned up on the fender which has a ground connector on it and a little nut to hold it on, called G104. There is one right next to G104 with a bunch of terminals, but I’ll get to that in a minute. One of the connectors on G104 is just a wire that jumps right over to G103 on the fender. Now if my memory and my manual serve me right, the other connector on G104 is the main battery cable from the negative terminal on the battery. Please let me know what you find when you trace that big fat black cable from the battery, I really think it connects to a ground post under the coil pack. G103 feeds ground to a bunch of unrelated accessories including turn lamps and the horn and that’s all. So I highly doubt that anything you did to the ground terminal on the fender affected the problem that you have. I hope I didn't mislead you earlier on that one particular terminal, but I wanted to stress the importance of cleaning all of them
As for the rest, G100 is home to a bunch of different grounds including all of the ones for PCM related components including the PCM itself and the cruise module. Now I remember almost perfectly I was able to find this one under the coil pack, I know this because I was stupid enough to cut off a bunch of the terminals to replace them before I realized I was all out of new terminals to crimp on. So I had to twist all the wires together and limp the car down to my electronics store to buy new ones. This ground post is the most important and is located under the coil pack. You'll know this one from G104 because it will have a bunch of terminals on it. I would definitely make sure you clean this up.
G102 is under the driver’s side fender and feeds a lot of different accessories that aren't relevant. It also supplies ground to the electronic level control system if you have this, and some of the electronic controls for the brakes, but nothing that should bring down the car in the event of a bad connection.
Now the second wire that splits off of the battery on the negative terminal, which is a thin spaghetti looking wire compared to the heavy main ground line, is also important. This wire feeds a number of electronic control modules, including all of the grounds for the air bag system and the ABS system. So you want to make sure that the connector you looked at is making a really good connection, I actually cut the connector out and soldered a new one in because I didn't trust the quality of the terminals inside based on what the outside of it looked like.
Ok, so this is long reply I know, but I hope it helps you narrow down the source of the problem. The thing about it is, if you played with that battery terminal at all or any of the main connections from there to the engine this could be what caused the improvement. Not only that, but we haven’t even gotten into the positive side of things yet, and any terminals that you cleaned or repaired or even just unplugged and plugged back in could have fixed the problem. But to be sure I would definitely hunt down that ground post under the coil pack to clean all of those connectors as well. As I remember on my old 92' if your newer one looks the same, G100 is in a dark crevice behind a metal plate, but once you find it, it is fairly obvious from that point forward. I remember having to work around the accessory belt while I was trying to tighten the nut back on the post with my ratchet, so that should give you some idea of where this is.
Most importantly, however, I am really curious to know exactly what things you cleaned or re-connected that were related to the power distribution under your hood. I really think you must have done something else to cause this change in behavior, other than just cleaning the G103 post on the fender. Either that, or your ’95 model is wired differently from my older ’92. Maybe you can shed some more light on exactly what you altered or changed under there.