[scrot]

Hello, I have an '87 Olds Cutlass 2.5L that sometimes runs very rough only when it's first started, and only in the morning when it's cooler. Not every time, but sometimes.
I've checked the obvious like vacuum leaks and loose spark plug wires, but those problems would make it run bad all of the time.
I thought about a sticking EGR valve, but I think that would be a problem more than what it is.
The manual says code 45 is a rich a/f mixture. It also says it could be the O2 sensor, but I thought the O2 sensor isn't used right after start-up.

Doesn't the car stay in "open loop" until the O2 sensor heats up?

I believe the computer ignores the sensor input and refers back to it's memory for the fuel mixture and timing.
I'm thinking it could be a bad coolant temp sensor.
I believe if it is bad, it will tell the computer the engine is colder than it is, and the computer richens the mixture.
I don't get any trouble codes for the coolant sensor however.

What do you think?
Thanks.

[Deuce]

I am moving this to the engine area ... hopefully where it will be viewed by a few engine techs ... BUT ... I had the same problem with a 90 Silverado with a 350 and I changed the temperature sensor at the thermostat housing and the problem went away ...

The sensor was giving false readings to the computer.

[oldbogie]

Originally Posted by scrot

Hello, I have an '87 Olds Cutlass 2.5L that sometimes runs very rough only when it's first started, and only in the morning when it's cooler. Not every time, but sometimes.
I've checked the obvious like vacuum leaks and loose spark plug wires, but those problems would make it run bad all of the time.
I thought about a sticking EGR valve, but I think that would be a problem more than what it is.
The manual says code 45 is a rich a/f mixture. It also says it could be the O2 sensor, but I thought the O2 sensor isn't used right after start-up.

Doesn't the car stay in "open loop" until the O2 sensor heats up?

I believe the computer ignores the sensor input and refers back to it's memory for the fuel mixture and timing.
I'm thinking it could be a bad coolant temp sensor.
I believe if it is bad, it will tell the computer the engine is colder than it is, and the computer richens the mixture.
I don't get any trouble codes for the coolant sensor however.

What do you think?
Thanks.

It could be a defective thermostat that keeps the engine too cool. It should come out of cold start at 176 degrees. The computer is also doing a time check so it'll look at the mixture at a point after start up where the engineers had established that the engine should be warm enough to suspend "choke" and go into closed loop.

I doubt this is an EGR problem, which usually causes the engine to fall lean when it fails.

It could be a fuel system pressure problem caused by the regulator. Higher pressure will result in a richer mixture.

A leak in the exhaust ahead of the O2 sensor will cause the sensor to drive the computer to richen the mixture.

A faulty O2 sensor, damn there hard to take out, rust being what it is!

If this problem is real, you should see a reduction in fuel economy. In my Franken Mouse powered S-15, I watch every tank full because I have flex couplings in the exhaust ahead of the O2 sensors and they are a known leakage point, when mileage takes a 3-5 miles per gallon drop, I know it's time to replace them.

Bogie

[scrot]

OK, thanks for the replies.
It happens imediately after starting up, so I know it's not the thermostat.
I don't have a flex coupling. Even if I did, the O2 sensor isn't an input to the computer at this time, Correct?
I'm going to check/replace the coolant temp sensor.
I imagine it could also be a sticking idle air control motor.
Any others suggestions?
Thanks.

[T-bucket23]

I would check the O2 to see if it working, check the temperature sensor that talks to the computer, and the Idle air control in that order.
It could also be the MAF or MAP depending on what type of system it has.

[oldbogie]

Originally Posted by scrot

OK, thanks for the replies.
It happens imediately after starting up, so I know it's not the thermostat.
I don't have a flex coupling. Even if I did, the O2 sensor isn't an input to the computer, Correct?
I'm going to check/replace the coolant temp sensor.
I imagine it could also be a sticking idle air control motor.
Any others suggestions?
Thanks.

Yes the O2 sensor is an input to the computer, however, it's supposed to be held out till the engine hits 176 degrees F at which point it should go into closed loop operation. But there's also a timer circuit that assumes the engine is warm after a certain period of time that will attempt to put the engine into closed loop regardless of the temp sensor.

The O2 sensor is the device that determines and sends a signal to the CE light by way of the computer if it senses a rich mixture. It should have enough authority to step the mixture leaner, but if it can't it sets code 45 which means it's maxed out.

The question is not only why is the mixture rich but why are you getting a code when the O2 sensor should be ignored by the computer. Assuming the computer hasn't gone south, one has to consider the problem is either a sensor is out of whack, a connector or wire is having problems.

The IAC could be out of adjustment, but the MAP sensor should pick that out as a disparity between idle speed and manifold vacuum.

A failure of the temp sender in the hot position would make the computer think the engine was warm when it wasn't and it would be in closed loop at start up but the engine would show lean characteristics when it was cold, i.e. backfire.

Where I think a good possibility is that a failure of the sensor in the cold position would cause the computer to fuel the engine in "choke" for a rich start up mixture. In addition to the temp sensor, there is a time out circuit for the O2 sensor and if the computer counts down and goes to closed loop, but the engine temp sensor is reading cold, the "choke" condition would prevail into the closed loop function at least long enough to get a code. In that case the O2 sensor would try to lean out the mixture, but since it's still in cold start mode on injector timing, I doubt the O2 has enough command authority to lean the mixture enough for it not to be rich. At that point the computer seeing a continuously rich mixture would set code 45. One has to keep in mind that the O2 sensor is simply a little clean up circuit, in cruise and under certain conditions, it trys to keep the mixture chemically correct for efficient burn, but it's very limited in what it really can do. The engine would act similarly if the cam was falling out of time, the ignition was out of time, a spark plug wasn't firing, or an exhaust valve was burnt. So these conditions also need to be checked out as they would either allow excess unburnt fuel to pass through.

My reference to an exhaust leak ahead of the O2 sensor is to mean that a failure of an exhaust gasket, a loose or holed manifold/pipe would add air in the exhaust which would cause the O2 sensor to correct by richening the mixture, while this usually doesn't set a code, you'll see fuel economy decrease. The same can be said for a leak into the intake side. This also will add excessive air to which the O2 sensor will attempt to correct with more fuel. Again these usually don't cause a code, but mileage suffers. None of these conditions occurring in a carbureted engine would cause a carb to correct the situation so this is new and unique to EFI motors.

Lastly you can't rule out a failure of the O2 sensor, though this usually shows as a specific code. But I've been involved in electronic systems self diagnoses for many years and if one thing is true in complex systems, its each component points to another when a failure occurs (kind of like kids getting into trouble), so the code shown isn't always what the real problem is.

Bogie

[scrot]

OK. Thanks a lot people, and oldbogie. I'll just go ahead and replace the coolant temp sensor and see what happens. It's less than $20.

[scrot]

I took a shot in the dark and replaced the coolant temp sensor since it's cheap and easy to replace and I believe it's an important sensor. Unfortunately, it didn't change anything. No biggie. It didn't work, but at least I now have a fresh sensor (and a spare) that should last for the life of the car.
I then spent some time looking thru the manual, looking for other possibilities such as the suggested MAP SENSOR, and MANIFOLD TEMP SENSOR.
I went under the hood and decided to remove the EGR valve to check for it sticking open. It was ok, so I cleaned and lubed the shaft with anti seize and reinstalled..
I then jumped the fuel pump relay to pressurize the system without the engine running to look for a leaking TBI injector.
There it was!

The TBI fuel pressure regulator, which is mounted up near the TBI injector, was leaking gas around the diaphragm.
It made the area wet with gas inside the air cleaner, causing the vapors to be sucked into the engine.
I bought a new one ($25). The new diaphragm is a little thicker around the circumference, which I believe is a design change to fix this problem. In the TBI housing the diaphragm is sealed by sandwiching it between a groove in the housing and the stamped dome that is screwed to the TBI.
I installed it and briefly checked, and it's as dry as a
This should fix the symptoms. I'll see in the next few days.
Thanks for everyone for the help.