A bit late to this thread as I found it on a google search for other stuff, and joined so I could answer. I may be able to answer at least one of the questions.
Titania lambda probes are often 0v/5v switching, but not always. However the scantool representation of them can be significantly different than what you'd see on an oscilliscope.
For example My Volvo S40 has 0v/5v switching Titania lambda sensors. On the oscilliscope observing their signal directly they go to nearly 0V for rich and 4-5volts when lean. Yup, low for rich. In the scan tool showing live data, in this case ScanXL Pro, they show as low for rich and high for lean still, but the scantool signal waveform is -
a) truncated, so you only see 0V to 1.275V portion of the actual waveform, and,
b) isn't sometimes an accurate representation of the finer details of the real signal due to the update rate of the scantool, even running single PID updates.
Have you tried doing a running compression test and better still with a pressure transducer viewable on a scope?
How long was the LPG kitted to it and which make model was it? Active or passive injectors?
