The plot thickens...
I've just checked my fuel trims and O2 sensors.
If I had a vacuum leak, the fuel trims would be very high to compensate for the leak, correct? Here's mine. Short term (instant) on top, long term on the bottom:
Mine's completely normal. Normal operating readouts oscillate constantly between around 5% and -5%. If I were to have a vacuum leak, the long term and/or short term would be way into the positives.
However, look at my O2 readings on a cold start after the SAI has turned off. Top two are pre-cat, bottom two are post cat:
Post cat O2 readings are fine - bang in the middle. Not running lean at all.
Then I drove the car for a while, stopped, and checked it again:
As you can see the pre-cat sensors are still working normally, but now the post cat results are way, way too lean. The two lines should be in the middle.
Now take a look at this - these readings are whilst driving steadily at 30mph on a quiet industrial estate:
Then when I stopped the car:
They're all over the place!
Reading up more and, apparently, misfires can cause a lean result. You'd think that they'd cause a rich result due to the unburnt fuel, but car's don't have fuel sensors, they have O2 sensors, and they'd pick up the unused O2 in the cylinders and cause a lean result. Which is kind of what's happening, except mine's happening constantly, not just spikes - it's constantly lean in both banks.
So, lots more info, but I'm still unsure. I think I've ruled a vacuum leak out with the fuel trims being normal, but what's causing the O2 sensors to be constantly lean? The car's just passed an MOT emissions test, so it's not the CAT's.
Interesting...