|
|
I've now had five or six incidents, two in the same day and the others separated by at least a month, where my Saturn has suddenly lost power. In those cases, the engine RPMs drop to around 100 and stay there, but the car doesn't stall (normally it idles around 1000). Pressing on the accellerator doesn't do anything. If I turn it off and turn it right back on, I get the same problem. If I turn it off, let it sit for a minute, and turn it back on, it's fine for another month or two. The time it happened twice in the same day was in West Virginia, and both times happened while going up long steep hills. It happened last night going up a hill that's big by Ann Arbor standards but small by West Virginia standards. It also happened once or twice in traffic jams on level roads. Any ideas?
11 responses total.
Fuel pump? Cloggy fuel filter?
Going up a hill, says Jim, gravity could be moving the gas away from the intake. But that does not explain stalling in traffic. More details? Saturn made only fuel-injected, not carbureted, he says. Jim has no experience with new cars. (post 1987) 1991 was the last carbureted Subaru, last car with a carburetor.
Something loose in the gas tank occasionally drifting over the intake? When I was a kid somebody had slipped a plastic bag with rocks into the gas tank, and after a while the bag worked loose and would get sucked over the intake.
The fact that it's happened both on long hill climbs and idling in traffic suggests to me that it might be heat related, but this is a tough one. Has the 'CHECK ENGINE' or 'SERVICE ENGINE SOON' light ever come on during these incidents? If so, the computer has stored a 'trouble code' in its memory, and reading that might provide a clue.
I asked Jim if his car ever talks to him and tells him to service the engine soon. "Yeah, but not in English, or Japanese, or German." (He is laughing too hard to explain.) 'It says it all the time, in fact I don't think I've ever been in the car when it didn't say that'. 'Which of the cars?" "Any of them". (The engine on one has been making loud descriptive noises since 1986 when it came here from Montana.) He has not yet serviced it - he was waiting for the engine to die so he could change the engine and it will not die, with 140,000 miles and one bad cylinder. He feeds it oil, curbside oil. He cannot imagine coddling a car with a computer.
I bet it's computer / computer sensors related. For it too run at 100 RPM the spark must be retarded to TDC (Top Dead Center) or after. Intermittent problems like that are tough to isolate. The computer *might* keep track of engine / sensor problems, but probably not long enough to make it worth taking in. You can call and ask, however. About your only real choice is to keep driving it till it becomes common enough for a service person to experience.
Re #5: Carburated cars without computers are definately easy to work on. Fuel-injected cars without computers are a nightmare. The FI system on my '75 VW had dozens of electrical connections and nearly half a dozen sensor inputs spread out all over the engine bay. With no computer diagnostics to give a clue what wasn't working right, troubleshooting became quite a chore. Re #6: I'm not familiar with the computer systems on Saturns, but on my Ford any fault where the computer realizes something is wrong (a sensor that goes out of range, for example) triggers it to save a trouble code in non-volatile RAM. As long as the battery isnt' disconnected, that code will stay there until it's cleared, and it can be read out at any time. Ford systems have two types of trouble codes. "Hard codes" are problems that the computer senses at the time you tell it to do a self check. "Soft codes" are problems it's noticed in day to day running and stored in memory. The codes don't always tell you exactly what's wrong, but they tell you where to start. (A code indicating that the oxygen sensor is reporting a lean condition all the time, for example, could be the result of a bad sensor, a loose connection, or a fuel system problem.)
The mechanic I had taken the car to in West Virginia after this happened twice in one day had said he didn't have the "software for connecting to the computer," for Saturns this old (1994), and suggested taking it to the Saturn dealer. Since the car had been well behaved from then until a few nights ago, I hadn't gotten around to that yet. According to the Hanes manual the reason the mechanic didn't have the software for reading my car's computer is that there is no such software. Instead, you short a couple of contacts together and it displays the diagnostic codes by flashing the service engine soon light. In my car's case, the service engine soon light flashed 32. The Hanes manual says 32 means, "EGR System Fault. Vacuum switch shorted to ground on start-up, switch not closed after teh PCM has commanded the EGR for a specified period of time or the EGR solenoid circuit is open for a specified amount of time. Replace the EGR valve."
Yeah, the Ford system on my van is the same way. The only "diagnostic tool" you need is a bent paper clip. ;) It's possible that an EGR system problem could cause the problem you've seen, I suppose. An EGR valve open at idle will cause the car to stall or idle rough; one that's closed all the time (or plugged up with carbon) can cause pinging and overheating. Whether replacing the valve will actually fix things, I can't say.
So how does shutting the car down for a few seconds not fix the problem but shutting it down for about a minute does? I don't think EGR valves work that way. That sounds like how computers work. It could be going out in the weeds. A computer in a car has to have heavy filtering on the power supply so it is possible that it is staying up if you shut down for just a few seconds. Shutting it down for a minute allows the filters to drop below the boot circuit's threshold voltage and when power is applied the CPU reboots.
Could be a thermal problem in the valve itself or the wiring connected to it, as well. The EGR valve is part of the exhaust system, which means it heats up and cools down pretty quickly. One of the rules of fuel injection system diagnostics is you check everything else out *before* you assume the computer's bad. This is simple economics; in the case of my VW, a rebuilt computer was over $250, and wasn't returnable if it didn't solve the problem. That made it the most expensive part in the whole system. Engine computers rarely fail, but the wiring to them often does. (I remember hearing about one particularly frustraing case where the engine would die more or less randomly, but there were never any codes stored in the computer. Turned out the computer's ground connection was loose.) Did I mention that a lot of times the first step in troubleshooting a fuel injection problem is to tighten every ground in the system? Many sensors are grounded though their mounting brackets and those screws do come loose.
Response not possible - You must register and login before posting.
|
|
- Backtalk version 1.3.30 - Copyright 1996-2006, Jan Wolter and Steve Weiss