R33 GTS25 NA - Stalling when vehicle is at operating temp

[MBS206]

Have you confirmed the 2 pin coolant temp sensor on the motor is working properly?

I've had a very similar issue when I forgot to plug mine back in many years ago.

[joshuaho96]

On 4/16/2025 at 6:33 AM, GTSBoy said:

Nah, it's different across different engines and as the years went on.

R32 era RB20, and hence also RB26, the TPS SWITCH is the idle command. The variable resistor is only for the TCU, as you say.

On R33 era RB25 and onwards (but probably not RB26, as they still used the same basic ECU from the R32 era), the idle command is a voltage output of close to 0.45V from the variable resistor.

R33 with RB26 seems to go off of learned voltage for idle. Seemed like whatever the baseline voltage was once ECU first gains power will be the idle voltage. With ignition on unplug/replugging the TPS would relearn the idle TPS voltage.

Edited April 17 by joshuaho96

[GTSBoy]

I dunno. I just go off what I know works. On RB20, the idle switch meant something to the ECU and the potentiometer was ignored.

In Nistune, with the switch unplugged, you could bridge the terminals in the loom connector and see IDLE come on and off. Not so by moving the pot.

The R32 RB20 ECU and the 26 ECU both look like this.

Sure there is the "throttle sensor" (pot) on pin 38, and also sure, the idle switch is also directly wired to the TCU, whereas the pot is only directly conected to the ECU. But I am sure that the throttle position from the pot is passed to the TCU across the data bus on pins 21, 22 & 31. Maybe the ECU likes to know throttle position, but it sure as hell doesn't use it to determine the idle condition.

Meanwhile, on the later engines, like the 33 25DET and my Neo, you remove the TPS and move the pot to-from the 0.45V position, and IDLE comes and goes in Nistune. No throttle switch on the ECU diagram. Just the pot.

[dbm7]

19 hours ago, MBS206 said:

Have you confirmed the 2 pin coolant temp sensor on the motor is working properly?

I've had a very similar issue when I forgot to plug mine back in many years ago.

Same thought crossed my mind ~ depends on how one connotes 'stalling'...ie; gets a rough/stumbling idle as it get warm until it stalls... or... idles ok and simply falls-over when it gets to temp...

...you can test the whole circuit with a couple of resistors ...unplug coolant temp sensor, bridge terminals with a 2K7 resistor (ECU will do cold start), or with engine warm bridge with a 330R resistor (ECU will consider engine to be at normal operating temp)...it's a quick way to check wiring integrity/ECU response when you don't have a multimeter handy ...

[dbm7]

Duh... to answer my own silly question, it's actually described in the FSM...

...400 pages away at the end of the manual, for RB25DE/DET signal descriptions, it cites the TPSwitch signal action, is dependent on the TPSensor value ~ this tends to infer the builtin POT voltage signal is the primary, and the switches are fallback/secondary should the POT fail/TPSensor signal lost (and switch alone with no TPSensor signal allows for base idle speed setting).... makes sense... they (TPS units) used to fail/wear the POT with time, they're not exactly built to last ~ having the switch as a redundancy gets around this...(or, it's less likely both signals would be lost as they're on different power rails)... and of course wrt RB26DETT, you have to electrically disconnect the IACV solenoid from the harness, to defeat idle air control...  

[joshuaho96]

21 hours ago, GTSBoy said:

I dunno. I just go off what I know works. On RB20, the idle switch meant something to the ECU and the potentiometer was ignored.

In Nistune, with the switch unplugged, you could bridge the terminals in the loom connector and see IDLE come on and off. Not so by moving the pot.

The R32 RB20 ECU and the 26 ECU both look like this.

Sure there is the "throttle sensor" (pot) on pin 38, and also sure, the idle switch is also directly wired to the TCU, whereas the pot is only directly conected to the ECU. But I am sure that the throttle position from the pot is passed to the TCU across the data bus on pins 21, 22 & 31. Maybe the ECU likes to know throttle position, but it sure as hell doesn't use it to determine the idle condition.

Meanwhile, on the later engines, like the 33 25DET and my Neo, you remove the TPS and move the pot to-from the 0.45V position, and IDLE comes and goes in Nistune. No throttle switch on the ECU diagram. Just the pot.

I should try the experiment you're talking about, the throttle switch is still there carried over from the R32 and it's still all wired up but after I did the whole intake manifold refurb and had to recalibrate the TPS I managed to somehow get the idle switch reporting activation at 0.22V, then when I adjusted it to 0.45V for idle it decided the engine was permanently no longer idling which caused some very weird behavior, closed loop idle was disabled so it would basically be at the whims of the cold start valve and whatever the base timing table was at. Then just unplugging/replugging the TPS with the ECU live caused it to relearn the idle TPS position and decide 0.45V was idle. Presumably there's nothing in the TPS that allows for the throttle switch to "recalibrate" like that, not easily at least.