My 87 LeSabre (3.8L) is a fine running machine and rarely gives me problems. I recently purchased a scan tool and small scope to help with a problem on one of my Olds and later on hooked up the scan tool to have a look at my Buick just for curiosities sake and was surprised to find out that after 10-12 minutes of running time in closed loop it goes back into open loop without setting any trouble codes. The only time it will go closed loop is immediately after warming up to temperature but always returns to open loop after 10-12 minutes. After that it stays in open loop no matter how I drive it from idling through to cruising. The coolant temperature sensor is well above the closed loop requirement and O2 sensor has a good clean pattern cycling pattern on my scope.

Here'* the kicker though, if I set the scan tool to ALDL mode it immediately goes into closed loop and stays but if I go back to road test mode it goes back to open loop - how can that be?? I've learned that one of the things that ALDL mode does is bypass the closed loop timers which suggests that maybe one of the sensors is giving poor quality data to the ECM and so causing the timers to constantly reset so that it doesn't have a chance to go closed loop again.

So with that in mind I've now tried 3 different O2 sensors, tried a different coolant temperature sensor, different MAF sensor, different TPS, different IAC valve and even substituted another ECM but still end up with the same pattern of behavior i.e. open loop on cold start, warms up to about 185ºF and goes closed loop for 10-12 minutes followed by going open loop again regardless of whether I'm idling or cruising or whatever - it just won't go closed loop again.....unless I force it to by going to ALDL mode. One other piece of information - I did remove the cat converter a couple of years ago and replaced with a straight pipe, you can still get away with that up here in Alberta but I don't think there'* any way the ECM can figure that out in OBD1 or am I wrong? Is there some way that the ECM can tell that I've removed the cat by looking at the O2 sensor data or some other operating parameter? Maybe that'* what I need to do next is replace the cat but the old one was rusted out and I'd rather not go to that trouble and find out it'* not the issue.

Anyway, long winded thread but if anyone has ideas on what the problem might be it would be much appreciated.

The cat should not be a factor. Other than the desire to have your car run closed loop when it should, are you experiencing any driveability issues? I ask because this might lead to why the open/closed loop is not transitioning as expected.

It runs well throughout all ranges, lots of power for a 3.8L, maybe a bit of a hunt while idling and the typical Buick V6 shake while transitioning through about 1200 RPM so no significant driveability issues. Apparently it has been running open loop for some time now unaware to me (no codes) until I decided to have a look. Running open loop probably means it'* slightly on the rich side all the time which may not be such a bad thing I suppose from the perspective of performance but I'm really curious about why.

You seem to have addressed the main possibilities; O2 and ECT sensors. What are your short term and long term fuel trims looking like?

The BLM and integrator are both at 128 so no trim adjustments have been made. I'm at 33 ignition cycles since the last time I reset the ECM so that means only about 6 hours of total closed loop time for the ECM to work with.

Will your scanner read O2 sensor activity? If so, what are you seeing?

Yes, I'm able to see the O2 sensor information with both my scanner and scope. During the short 10 minute period of closed loop operation my scope shows a normal O2 cycle back and forth from rich to lean and the scanner shows lots of cross counts and also has a pair of LEDs showing the rich-lean cycles. I've tried 3 different O2 sensors including a new one and they all show the same healthy pattern that you expect to see but the ECM still goes open-loop after 10 minutes for all 3 sensors. I think the fact that it goes into closed loop no problem in ALDL mode is telling me something but so far can't figure out what that is.
In any case, if the O2 sensor was the problem do you think it would go into closed loop in ALDL mode? As I mentioned earlier, I've read that the closed loop timers are by-passed in ALDL mode but not under road test conditions (road test mode) so it may have something to do with the timers?? I think the way these timers work is that if the ECM gets a sporadic spike from a sensor that is momentarily out of range it will go into open loop and then start a timer. If the spike doesn't repeat itself within the timer sequence (2 minutes or so) it will go back into closed loop but if the spike repeats within the timer sequence, the timer will start over again from zero and so will stay in open loop if the spike is constantly repeating over the timer interval. Is that the way you see these timers working? In any case, the two main parameters other than time are O2 sensor and coolant temperature and I've now tried 3 different ones for each sensor with the same result.
Regarding the O2 sensor activity, since it operates in open loop most of the time, it'* interesting to watch the rich-lean pattern as programmed into the tables - it'* surprising to see how lean it runs when idling and also when driving with a light foot but then goes immediately rich when calling for acceleration so it seems to be running pretty clean under no/low load even without the O2 sensor feedback. It also provides lots of power when asked for on acceleration - very smooth operating engine with no lags or hesitation. As I said, there'* no driveability issue - just a quirk I'm curious about.

With the O2 sensor functioning I am surprised to not see any fuel trim activity. Closed loop essentially means the ECU is using feedback from the O2 sensors to maintain target A/F ratio by adjusting injector timing. The fact it is not doing that may by why the system is kicking out of closed loop. Now, the question becomes what is preventing the ECU from commanding short term fuel trims based on actual feedback from the O2 sensors?

thanks Roy, that'* a very good point - I think what I'll try next then is to put it in ALDL mode where it will stay in closed loop and monitor the trim while I bias the O2 sensor voltage to both extremes - I'll let you know what I find out.

Hi again Roy, think I may have found the source of the problem. While monitoring the trim I also reconnected my scope on the O2 sensor. As I mentioned before, the O2 sensor signal is very clear with many cross counts but if you keep watching it over the 10 minute time period before it goes open loop you can see the the signal slowly but surely starts to develop quite a bit of hash - just before it goes open loop again the amount of hash is really high, you can still see the rich-lean cycle but it'* just barely recognizable.
I went to the GM dealer and bought another new O2 sensor (only $35 for this car, believe it or not!), installed it and watched again - got the exact same pattern.
As I understand it, this hash could be caused by either a leakage of air into the intake manifold or possibly by a weakness in the fuel delivery system. Since the car behaves so well under high load even when operating in open loop I'm guessing I probably have a leakage of air somewhere. This also fits with getting the same cycle every time - fine for the first 10-12 minutes of closed loop and then goes open loop so possibly an air leak from thermal expansion somewhere on the intake manifold? Also fits with the slight rough idle I can feel, most noticeable after warm up.
I've already checked all of the obvious vacuum leak sites - hoses to the EGR, vapor recovery, brake boost, climate control etc. and nothing obvious there. I suppose I should also remove and inspect the EGR to make sure it'* not developing a leak after warm up? In addition, since this only happens after a good long warm up, I'm pretty sure it must be somewhere that is close to the engine, a result of thermal expansion and not a remote hose connection.
I also noticed I have a leak starting in the straight pipe section where the cat converter used to be but is it possible for air to get back up to where the sensor in mounted on the exhaust manifold? I'm thinking not very likely but I don't have any experience about this.
Anyway, let me know your thoughts - and thanks again for your very helpful input.