That is true.

 

If an assumption must be made, yes I would also agree with your assumption. I understand that the ECU has a pre-set program for open loop operation that includes fuel trim, ignition advance, etc. This program sees the engine temperature at startup, and uses a time factor to determine approximately how long it will take to warm up and reach closed loop operation. It has to be an approximation because the engine will warm faster if driven and slower if it's only idled, and the open loop program has to work adequately for all warm-up scenarios. Once the engine actually warms up and closed loop is achieved, the open loop program ends and the ECU starts taking live readings from the sensors to determine fuel trim, ignition advance, etc. Also, I could be wrong about some/all of these details but this is my understanding of how the engine operates during warm-up (open loop). This prescribed closed loop program is a fairly common industry approach for fuel-injected electronic controlled engines.

 

Yes, in open loop that is true, it won't care about the exhaust leak since its running off a preprogrammed timing table, once the CTS reaches temps and the timers expire, the computer reads sensor information in closed loop.

 

I am not all that sure ....

The ECU has a 'lerning' ability. During closed loop the ECU will adjust the fuel trim. Both short term that is erased every time the ECU is powered off and long term that is stored in another memory that is not erased.
An exhaust leak will affect the closed loop fuel trim, and the ECU will store bad fuel trim data.

It is possible that the long term fuel trim is active in open loop also.

I don't know if anyone has reverse engineered the ECU source code because that is the only way to know for sure.

 

This is true, at least for the 94-96 PCMs. This is why Tunercat has a version of the vehicle file definition with a BLM-locking function programmed in. If you flash a calibration file with that feature, then it will not keep learning in open loop. But otherwise, it does. For a 96, there is no option I know of to lock out BLM in open loop, because Jet DST does not offer a definition file with that feature (and that is the only tuning software now available for a 96). As for earlier vehicles with completely different ECMs, I don't know if they have BLM enabled in open loop or not.

All that said, I wonder how much an exhaust leak really affects how much O2 the sensor sees? Any time there is positive pressure in the exhaust (basically any time the throttle is open), then an exhaust leak would be letting gasses out, not letting air in. That wouldn't affect the O2 reading at all. I suppose if we're talking about a leak where at the exhaust port meets the header flange, then it can have negative pressure at certain times in the valve cycle at certain throttle settings and certain speeds; and that would allow air to be sucked in and affect the O2 reading. But I doubt any stock manifold ever has that condition, and even long-tube headers probably rarely do on street cars. And it would have to a whopper of a leak in a single primary to have a serious effect on the O2 reading.

 

 

The engine runs fine in open loop with no learned valued because it is set to run to a targeted a/f ration rather than trying to seek stoichiometric ratio. IOW, it makes best power safely with no learning. The BLM learning it does in open loop is toward efficiency and emissions, not toward more power or better throttle response or anything we would care about at WOT. So in fact, the learning it does may slightly compromise power and safety as it heads toward leaner mixture. Again, that's why the BLM-locking feature in Tunercat exists: so that tuners can lock in a target a/f ratio that is both safe and ensure max safe power levels at or near WOT.