2G stuck in limp mode

[tschuhly]

Hey all,

I am trying to figure out what steps I should go through to figure out what is causing my issue.

Some background is that the car has been put together and running since about mid April. Towards the middle or end of May it started to act like it had a misfire or something. I replaced the spark plugs with the higher horsepower ones but that didn't do anything.

I decided to check for codes and I got the knock detection circuit fault code. I found that my sensor was loose so I tightened it up and no luck. I then was driving it and as I took off from a light, it sputtered a little and when I gave it slightly more gas it bucked hard once and then turned off. I got it towed and later found that my ECU 20amp fuse had "blown" or broke (I'm not sure if it was just old and brittle or if it legit blew). I replaced it and my knock sensor with a new one jic and my car started up except now it won't go over 4500rpms.

It doesn't have any codes but is still running a little weird and is still in limp mode. I tried to go into my ECU pinout settings to tamper with them to see if it was because of my AEM boost sensor being pinned to my MDP pin, but it is still in limp. I will say that for a little bit I had forgot to unplug my stock MDP so there were probably 2 conflicting signals going to the ECU so maybe that did something? I haven't driven it yet because I worried it will leave me stranded again.

I attached a log of it starting, idling and revving a little. Any guidance would be appreciated!

 

[tschuhly]

I swung by ECM tuned and met the legend himself! First of all, what a nice guy. Second, he said my ECU has nothing wrong with it so it's probably something in the harness. The clicking that he hears he suspects is the MPI relay, but he said that last plug that's not plugged in is the plug that powers the ECU. So the fact that the ECU is trying to trigger the MFI relay means that there's a wire that's powering the ECU when it shouldn't be.

Just found that connector B56 is causing the issue and tested voltage on each pin. Here is the list of reading and what each pin is labeled as.

Pin 81 - sensor impressed voltage - 4.5v
Pin 80 - backup power supply - 12v
Pin 76 - front O2 - 4.5v (also has my afr gauge spliced in)
Pin 73 - MDP - 4.5v (also has my boost gauge and sensor spliced in)
Pin 72 - IAT - .62v
Pin 92 - sensor ground - .6v
Pin 90 - MAF - .2v
Pin 87 - closed throttle position switch (idle switch) - .6v
Pin 86 - vehicle speed sensor - 9v
Pin 85 - barometric pressure sensor - .25v
Pin 84 - TPS - 1.08v
Pin 83 - coolant temp sensor - .6v
Pin 82 - ignotion switch - 12v

I am now definitely thinking that when I accidentally tried to jump my fuel pump and accidentally hooked to the engine speed reference plug on the harness that it melted some wires somewhere

 

[steve]

I swung by ECM tuned and met the legend himself! First of all, what a nice guy. Second, he said my ECU has nothing wrong with it so it's probably something in the harness. The clicking that he hears he suspects is the MPI relay, but he said that last plug that's not plugged in is the plug that powers the ECU. So the fact that the ECU is trying to trigger the MFI relay means that there's a wire that's powering the ECU when it shouldn't be.

Just found that connector B56 is causing the issue and tested voltage on each pin. Here is the list of reading and what each pin is labeled as.

Pin 81 - sensor impressed voltage - 4.5v
Pin 80 - backup power supply - 12v
Pin 76 - front O2 - 4.5v (also has my afr gauge spliced in)
Pin 73 - MDP - 4.5v (also has my boost gauge and sensor spliced in)
Pin 72 - IAT - .62v
Pin 92 - sensor ground - .6v
Pin 90 - MAF - .2v
Pin 87 - closed throttle position switch (idle switch) - .6v
Pin 86 - vehicle speed sensor - 9v
Pin 85 - barometric pressure sensor - .25v
Pin 84 - TPS - 1.08v
Pin 83 - coolant temp sensor - .6v
Pin 82 - ignotion switch - 12v

I am now definitely thinking that when I accidentally tried to jump my fuel pump and accidentally hooked to the engine speed reference plug on the harness that it melted some wires somewhere

I assume you met Tom at ECMTuning, I'm glad he also thinks this isn't an ECU hardware issue.
The MPI relay or relays since you have a 97 GST should click when you turn the ignition to the RUN position and the when the ECU sees the engine turning over, then again about 5 seconds or so after the engine stops turning and the same delay after you turn the ignition off. What it shouldn't do is chatter, switching on and off like a pissed off squirel.

Once again I'm not sure what or how you were testing by pulling a connector off and on. It's generally a bad idea to do so while the ignition switch is on.

A couple things:
I assume you no longer has a front O2 sensor and only the Wideband is connected to pin 76.
If you have two things attached to the MDP you need to disconnect the boost gauge on only connect the MAP sensor.
How are you measuring the voltage at the sensor ground, actually what are you using for the negative (ground) voltage reference for all these measurements. Each should be measured from the ECU ground point.
Mitsubishi MAFs output a frequency alternating between roughly 0 and 5v. So a voltage measurement doesn't tell much.
The VSS is much the same in that its a series of pulses as the transmission turns and it varies from 0 and battery voltage.
TPS voltage on a 2G has a wide spec for the output voltage 400 to 1000mV depending on where the internal IPS switches. Your is a hair outside spec at 1080mV that could be due to adjustment of the TPS or measurement offset due to your choice of negative point.

 

[tschuhly]

Hey Steve,

- I do have my wideband/narrowband sim hooked to my front O2 sensor plug.
-so I would want to unplug my MDP and only use my 3.5bar and that is okay even though I'm not set up for SD?
-i tested the relays with OEM reading and also connecting pins to 12v to hear for a single click.
-i was using chassis ground, so that was something I did wrong.

I tried to go through diagnostics using AI and AI concluded that it is internal to my ECU so now I'm confused. I will post the AI summary here to see if you all agree with its conclusion and hopefully clear up my confusion .

"Here is a complete, scannable summary of the diagnostic testing and results you can send directly to an ECU repair specialist or a peer for review.

Diagnostic Summary: 2G DSM Main Fuse Short Circuit​

The Vehicle: 2G DSM (1995–1999 Mitsubishi Eclipse / Eagle Talon)
The Initial Symptom: 20A Engine / MFI Fuse blows instantly when the ignition key is turned to the ON position.

---

Initial Pin-Voltages (Key ON)​

Initial multi-meter testing on ECU Connector B-56 yielded abnormal reference voltages due to an ongoing circuit failure:

• Pin 82 (Ignition Switch Input): 12V (Normal)
• Pin 80 (Backup Power Supply): 12V (Normal)
• Pin 81 (5V Sensor Impressed Voltage): 4.5V (Faulty – Dropped to 0.01V once aftermarket gauges were unplugged, proving the ECU 5V regulator loop was dead and backfeeding voltage).
• Pin 73 (MDP Sensor Input): 4.5V (Faulty – Backfeeding due to an aftermarket boost gauge splice).
• Pin 92 (Sensor Ground): 0.6V (Faulty floating ground voltage offset).

---

Step-by-Step Isolation Process​

1. Isolated Engine Bay Test (Harness Verification)​

• Action: Completely disconnected the ECU logic board from all vehicle plugs. Reconnected all factory engine bay components (Crank, Cam, MAF, Injector Resistor Pack, Coil, TPS, and O2 sensors). Kept the short-prone ISC motor and aftermarket gauge splices disconnected.
• Result: The 20A Engine fuse did NOT blow.
• Conclusion: The vehicle's factory wiring harness, sensors, and engine bay components are completely healthy and free of shorts-to-ground.

2. ECU Plug Isolation Test​

• Action: Turned the key OFF, plugged ONLY Connector B-56 (low-voltage sensor side) back into the ECU, and turned the key ON.
• Result: The fuse did NOT blow.
• Conclusion: The internal 5V regulated sensor loop, logic chips, and microprocessors on the B-56 side of the ECU board are structurally healthy and not grounded.

3. Pinpointing the Short​

• Action: Attempted to plug in the remaining power distribution connectors. Left Connector B-53 completely unplugged, but plugged Connector B-54 into the ECU.
• Result: The 20A Engine fuse blew IMMEDIATELY upon key ON.
• Conclusion: The short circuit lives exclusively inside the ECU casing on the B-54 circuit paths.

---

Final Diagnostic Verdict​

The vehicle wiring harness is safe. The issue is a hard internal short-to-ground inside the ECU on the B-54 power rail.
Because B-54 handles heavy-duty 12V power distribution, the internal failure is localized to one of two main circuits inside the computer:

1. The MFI Relay Ground Driver Circuit (Pin 38): A transistor or diode controlling the main vehicle power relay has internally melted down to the ECU's ground plane.
2. The Oxygen Sensor Heater Drivers (Pins 57 & 58): A solid-state grounding transistor used to power the high-amperage O2 heaters has internally failed and welded itself directly to ground.

Next Action Item​

The ECU board requires physical inspection by a specialist to replace fried driver chips (such as the M5269L series or solid-state transistors) and repair vaporized copper traces leading from the B-54 pin header.
If you want, tell me:

• What brand/model of aftermarket gauges you are using
• Where they are getting their 12V power from

I can help verify if your gauge wiring is safe before you plug in a repaired ECU."

 

[steve]

Hey Steve,
- I do have my wideband/narrowband sim hooked to my front O2 sensor plug.
- So I would want to unplug my MDP and only use my 3.5bar and that is okay even though I'm not set up for SD?
- I tested the relays with OEM reading and also connecting pins to 12v to hear for a single click.
- I was using chassis ground, so that was something I did wrong.

One or the other, not both. You don't have to be running Speed Density to connect a MAP sensor and log it. We did that long before DSMLink supported SD.

So they test good are they working correctly in the car?

If you use anything other than the same ground point the ECU is using you can create offset voltages due to the resistance between the two points. Best one is right at the ECU connector, second best is where the ECU ground connects to the chassis.

Good luck with the AI results. I'm not here to compete with or explain when it spits out garbage. That's a waste of my time.

 

[tschuhly]

Good luck with the AI results. I'm not here to compete with or explain when it spits out garbage. That's a waste of my time.

Woah, I'm not trying to argue or say your wrong. I literally was just trying to use it to make sure I was doing something that I have never done correctly and figured I'd try AI. If what it's saying is incorrect, then all I'm asking is to let me know.

I'm still planning on testing what you said tomorrow when I get a chance.