Speculation is my O2 readings are creating erroneous trims and as a result causing erratic running conditions. That said if I force open loop running it should actually run somewhat better. (Eliminate current too lean running under load)

The only engine control sensor that is ignored in open loop is the O2 sensor. Right?

This would just be an interim trouble shooting step to confirm faulty O2 readings while I'm planning a correction.

Wondering why this was not suggested by anyone looking at my logs - ? If it doesn't run somewhat better in open loop then I'm guessing there are other issues. Is my thinking correct here?

 

Yes if the engine is running like crap and you disconnect the O2 and it goes into its default strategy, and it runs better, then the O2 is probably junk.

 

All sensors report, and ECU responds to all sensor-input, during open loop.

Once engine temperature reaches programmed threshold-point, ECU adds one more sensor, the O2 sensor.

Until ECU makes fueling adjustments based on O2 sensor, fueling is managed by a fixed fuel table.

[O2 sensor is the only sensor on the car that actually generates voltage.

Connect a voltmeter to the sensor, and observe voltage delivery once the sensor is at operating temperature.

Recommended reading:
http://www.autotap.com/techlibrary/u...en_sensors.asp

 

Thanks for replies....

 

Measuring voltage is better than nothing. I also look at "crosscounts", as displayed on a scan tool. This is how many times the sensor switches from rich to lean/lean to rich. A low number indicates a lazy (defective) sensor.


I also had a sensor that became goofy at higher engine speed. It read false-lean, driving the fuel system rich at highway speed, but acted normally at lower speed. I never did figure out why this happened, and only fixed the problem because I got angry and just slapped another sensor in place--after testing EVERYTHING else that could cause a lean O2 reading.



Sometimes, it's better to replace an O2 sensor than to waste a lot of time with an older one.

 

I like the idea of monitoring “crosscounts” as an indicator of the health of the O2 sensor.

I just started using TunerPro. I’m getting triple-digit “O2 crossover” numbers. I assume triple digits is a sign that the sensor isn’t lazy.

I’d be curious to know what the time interval is that the crossover number relates to.

For instance, if TunerPro reports a crossover value of 250, I assume the O2 sensor-value crossed over the magical .450 millivolt threshold 250 times during some standard interval. What is the duration of the relevant interval?

 

I don't know the duration of the interval...and maybe it varies depending on the scan tool used.


With mine, I don't think I've ever seen a higher cross-count number than 60. I get worried when it drops to single-digits.

 

I played around a little more with TunerPro today.

Viewing a datalog from an arbitrary recent run, it looks like TunerPro is reporting crossovers based on about a 15-second interval.

Here is a screenshot of the "Monitors" tab, set for monitoring only O2 crossovers. There may be some way to change the monitoring-interval, but I haven't figured that out yet. Probably if you go into the "scalars" or some such in the menu, you can change the available monitoring interval for crossovers. There is a "time range" slider at the top of the graph which will allow me to select a LONGER montoring-interval. You can see I have the slider pushed all the way to the right, go give the shortest possible monitoring interval.

Anyhow you can see how the graph-line starts low, and then progresses upward until whatever point TunerPro wants to restart the count - in this case about every 15 seconds. I guess a smooth upward arc is indicative of an O2 sensor that isn't cutting-in and cutting-out, or just simply dead or disconnected.

 

You can go crazy, and show all inputs on the graph if you want.

Maybe helpful for pinpointing how the ECU reacts a when specific sensor is reporting something that just can't be (or shouldn't be) right.