Let me preface this by saying this is my sister's car, not mine, so some of the observations are secondhand. Check engine light came on last Wednesday, I was told about the issue on Friday before she left New Hampshire to come back home for the weekend. I said to drive slowly and that I would look at it when she got home. This drive apparently went fine. Car was showing a code 51 at this point- fuel mixture lean. I figured this was probably an O2 sensor issue, as the car had head gasket problems about a year ago and the sensor was not replaced. The "old" sensor looked to be the same as the replacement I bought, and didn't look like it had been in there more than 5 years or so. Installed the new sensor, changed the air filter, which was filthy, and went for a test drive. The car seemed fine, although I didn't really push it because of its age. CEL came on after about 10 minutes. However, the code 51 was now accompanied by a 21- O2 sensor lean. I drove back home, which was also uneventful despite the CEL being on. I reset the codes, but the code 21 returned. My sister tried to take the car back to NH today, but couldn't get on the highway. She said the car would not go faster than 45 and was shaking badly. She got back home on back roads and here the car sits. It's still showing 21, the 51 has not returned. The CEL does not come on until the car is warm, I was expecting that it would come on steadily on startup at this point. My current thoughts: -I damaged the new sensor on installation or it was defective from the factory (unlikely). I twisted the wires up pretty good installing it, and I'm worried that I pulled something loose internally. It didn't seem like it was enough to damage them and they didn't feel loose after I installed it. -Grounding issues. I forget which ground the O2 runs to, is it the one on the driver's side of the head? -Something else, like an exhaust leak in the downpipe between the manifold and the cat. I want to test the new sensor without removing it from the vehicle if possible. I think I remember seeing that they could be tested off-vehicle by heating with a blow torch and measuring across the output for a voltage change. What I'd like to do is warm the car up and then unplug the sensor and measure across the outputs, I should see an AC voltage here if the sensor is operating, and I expect it will probably be 7-8V ("RMS") on my cheapo multimeter. Is this a good test and are these reasonable expectations for the output? Any help's appreciated here!