Summary: original linear aka wideband aka AFR (air-fuel ratio) upstream oxygen sensor (marked 2R83 9F472 AC DENSO3120, has grey connector), Jaguar part C2C029250 / C2C 29250 / C2C29250 faulty. Replaced by X-Type front generic, box marked 234-9016. Looks like 234-9030, DOX-0428, DOX-0430, Bosch 15846 etc would all be OK and differ only in wire lengths. STR only needs short wire length so any will probably be OK. 4-wires: 2 blacks (heater), blue & white. You can buy sensors without a connector but I chose not to because many sellers are terrible at correctly describing what they sell.

DO NOT GET THE WRONG TYPE OF SENSOR! The commoner ones are the older, voltage-type, and used for the downstreams, whereas a current-sensing type is needed. Confusingly there are two meanings of Universal: one means it is a linear sensor but more often it just means it has no connector so you have to add your own.

Torque: 40Nm. 22mm hex.

Picture of connector attached. Note the colour!

---

I got DTC P1646. From JTIS / codes PDF it's about bank 1 oxygen sensor 1 - which means upstream (front of cat, close to the engine). P1647 would be bank 2, same position.

Quick aside: all P1 codes are car make, model, even year specific (so vital to use the right data).

Its 30A fuse was OK and as the Electrical Guide shows a common relay with bank 2 I decided the sensor was next to check.

OBD live data surprised me. Open loop on both banks! I expected the good one to be closed loop and then a clever PCM might have treated the bad one like the good one. Oh well, it's quite an old car and I know the PCM is already having a tough time in an STR. Unsurprisingly, fuel usage (MPG) was worse. Needed the car, though, so I hope the cats are OK. A little disappointed that the PCM wasn't doing better as it has the downstreams that should let it make sure it's not running too lean. Interestingly, no signs at all of restricted performance.

I couldn't see any real point in clearing the code. The PCM is good.

The Electrical Guide describes the sensor's location and connector colour - WRONGLY.

Each sensor 1 (upstream) is at cat top (2 is centre) and connector is GREY (aka GRAY). (Sensor 2 has black connector.)

Access is good once you remove the engine undershield (aka under tray / belly pan / "air deflector"). The sensor's connector is slotted onto a metal bracket. I left it there while pressing in the tab so I could pull the mating rear connector out. Then prised the O2 sensor's connector off the bracket to use a meter.

Checking the heater (2 black wires to pins 1 & 2) I found an open circuit. Quick compare against bank 2 which was 1.1 ohm (aka 1R1) when cold. So the heater had finally blown after 12 years 120K miles. I've had the car 7 years; this is the first sensor to fail. (The DCCV failed a while ago but is strictly speaking an actuator.)

So, I knew the old sensor had to come out. You can't get a socket on due to the sensor body but a ring spanner was OK once I cut the wires. There are special tools that don't require that (net search: oxygen sensor socket / spanner).

Oxygen sensors can be in "tight" (English euphemism, the word "extremely" being implied!) due to the heat more or less welding them in. Mine came out after a bit of wire brushing, penetrating oil (not WD40) and not much effort.

On the new sensor I was pleased to see some sort of probably anti-seize grease on the screw thread, protected for shipping by a small plastic cover. (Looked like molybdenum rather than copper grease.)

As I wondered how the PCM would cope I decided not to clear the learned fuel trims and other values and instead started the car while watching OBD live data. It went closed loop within 30 secs and as it warmed up I could see trims adjusting, and O2 values changing as ECT went up. Let it get plenty hot then switched off and let it cool over night.

The MIL (malfunction indicator light aka EML aka CEL) is supposed to go off after 3 warm ups (with driving?) so I decided to see. I think it was the 4th when the MIL went off.