A W212 that drops into park while the car is still rolling is reporting a supply the ignition switch stopped delivering. The selector moves to P because the unit behind it lost terminal 15 for a moment, not because the driver touched anything. The scope on terminal 15 shows that moment.
The driver reports that the car jumps into park as it slows at a junction or while manoeuvring in a car park, and that it restarts normally afterwards. At higher speed the same fault would be dangerous, which is why this complaint gets attention as soon as it appears. On the road it shows as a single event in a drive rather than as a habit, so the owner often arrives with no pattern to describe.
Ask whether it happens when the car goes over a bump and whether the engine has to be warm, because each of those answers describes a different joint. A fault that follows a bump is a connection and one that follows temperature is a joint opening inside a unit.
Read the ignition switch address and the gearbox address, then write both lists as the tester prints them. A switch that dropped terminal 15 stores an entry about supply, while the box stores an entry about an implausible selector position at the same moment.
Compare the time stamps where the tester shows them, because two entries recorded within the same second describe one event seen by two units. That pairing is the strongest evidence you can put on the sheet for this complaint.
Terminal 15 is the ignition supply the switch delivers once the key is turned, and it wakes the units that need to know the car is running. When it drops for a fraction of a second the units see an ignition cycle and react as if the key had been turned off and back on.
On this car the selector unit reacts to that cycle by moving to park, which is a designed response and not a gearbox fault. The unit that drives the selector wakes on that terminal and reports its state to the gateway within a fraction of a second. That is why the box tests fine afterwards and why replacing it changes nothing about the complaint.
Put the scope on the terminal 15 pin at the switch output and set a slow time base so a whole minute of driving fits on the screen. Drive the car over the surface where the fault appears and watch for a notch in the trace that a meter would never show.
A trace that falls to zero and returns within a few milliseconds confirms the switch, and the length of the notch tells you how long the units were without a supply. Save the screen, because that capture is the evidence the repair is planned from.
When the supply blinks, the selector unit restarts and picks the position it considers safe, which is park. The restart is quick enough that the engine keeps running, which is why the driver feels a gearbox event and not an electrical one.
Confirm the sequence by watching the selector position value and the terminal 15 trace on the same screen where your tester allows it. A position that changes within the notch is the proof that the supply caused the shift.
A selector unit with a worn contact can report park without any supply fault, so the trace is checked before the selector is condemned. A position value that jumps with terminal 15 steady puts the fault on the selector and clears the switch.
A box with a hydraulic or electrical fault stores entries about ratio or about a solenoid, and neither appears here. That absence is worth stating in the sheet, because it is what keeps the repair away from the gearbox.
Measure the permanent supply and the ground at the switch connector with the plug fitted, because a switch that is fed poorly produces the same notch as an internal fault. Compare both values with the battery posts and write them down. Write both values again with the key off, because a switch that is fed poorly behaves differently in the two states.
Measure the ground in millivolts against the battery negative post while the car is running, since a ground that lifts with vibration gives a notch on terminal 15 without any fault inside the switch. Clean the point and repeat the capture.
Repeat the drive over the same surface with the scope still connected and confirm the trace stays clean through the whole minute. A capture without a notch is the result to show the owner, because the complaint is intermittent and a road test alone proves little.
Write the time stamps of the paired entries and the length of the notch on the sheet, so a return visit compares like with like. Tell the owner which surface brought the fault out, since that detail is what makes the next test short.
Start with the battery and the charging system, because an ignition switch that drops terminal 15 for a few milliseconds is often a symptom of a supply that sags under load. Measure the battery voltage during cranking, then watch the charging voltage at idle with the headlights and the rear window heating on. A charging voltage that moves by more than half a volt while the engine idles tells you the alternator or its regulator is the source of the resets, and the gearbox is only reporting them. Once the supply is stable, repeat the road test with the scope on terminal 15 and on the selector signal, and compare the two traces on the same time base. The moment the two events line up is the moment you know whether the switch commands the shift or the gearbox decides it on its own.
| Symptom | Likely cause | What is measured |
|---|---|---|
| Jumps into P while slowing | Terminal 15 notch from the switch | Trace on the terminal 15 pin over a minute |
| Fault follows a bump | Ground lifting with vibration | Millivolt drop at the switch ground |
| Position value jumps with supply steady | Selector unit contact | Selector position value with terminal 15 watched |
| Same event in two units | One supply event seen twice | Time stamps of the entries in both addresses |
The selector unit sees an ignition cycle when terminal 15 blinks, and it moves to the position it considers safe when the supply returns. The box and the mechanical parts are not involved in that response.
No, because the gearbox is reacting correctly to a supply it lost for a moment. The scope capture on terminal 15 is what identifies the real cause before any part is chosen.
Not reliably, because a meter shows an average and a notch of a few milliseconds disappears into it. The scope is set to a slow time base so the whole minute of driving fits on one screen.
A selector position that changes inside the notch on terminal 15 puts the cause on the supply. A position that jumps while terminal 15 holds steady puts the fault on the selector instead.