1990 AWD High Idle/Power Loss UPDATE: Mystery Wiring

[Mokkan]

First, a big thank you to this community for creating an excellent collection of knowledge. I bought this car partially disassembled and have rebuilt the engine and gotten the car running based almost entirely on information I have found here.

I've finally gotten to a problem that I couldn't find a clear solution to. There are two issues, which may be related:

First, the car idles at 2000+ when cold and 1500-1800 when warm. I can reduce this by a few hundred by adjusting the idle setting screw, but I have never seen this car idle below 1000rpm.

Second, the car stops pulling power at about 4000rpm. In first gear, I can rev all the way to red line, but the power clearly cuts out along the way. I can't go into second (a third issue - transmission needs to be opened), so I shift to third and the car pulls alright until 4000rpm. At this point, the engine will not rev any higher and it feels very similar to a fuel cut out rev limiter.

This is a 1990 Eclipse GSX (turbo, awd) and is completely stock.

Actions I've already taken:

Compression Test:
Cyl 1 - 150
Cyl 2 - 155
Cyl 3 - 155
Cyl 4 - 160

I've preformed a boost leak test and stopped a couple of leaks. Boost appears to be building properly.

Fuel pump is brand new and pressure readings are:
31.5psi with regulator vacuum line
36psi without regulator vacuum line
(measured post filter at the banjo bolt)

Fuel Pump is recieving 6.7v at idle.

Idle air control stepper motor is measuring 30ohms at each of the 4 measuring points.

I'm suspecting the throttle position sensor right now, but could not figure out how to measure voltage on the sensor. All of the guides focus on 1991+ TPS's, so I measured resistance and can provide that for now:

Green-Red to Green-White:
42.3kOhm @ idle
1.9kOhm @ 50%
0.68kOhm @ WOT
Green-black to Green-White: 1.2kOhm @ idle - no connection when throttle opens

Green-Red = power
Green-White = sensor voltage
Green-black = ground

The ECU was replaced recently (with a numbers correct 1990 turbo awd ECU).

So, is there a way to confirm the health of the TPS? Or is there another cause that I'm missing?

 

[troubleshoot]

I would suspect a tps problem as well. They are very temperamental . I never tried to set it with a meter because I have link which makes it a snap. I just log tps voltage and turn it until it reads right. Im sure you just do the same thing with a meter but I dont recall the base voltage setting because its been so long. I want to say about .63v at idle but don't quote me on that. Im sure the info is around here somewhere.

 

[95blackGsTurbo]

Fuel Pump is recieving 6.7v at idle.
Or is there another cause that I'm missing?

I hope that is a misprint otherwise this is going to be the cause of the car not wanting to pull if it isn't getting any voltage. You should have battery voltage back there.

 

[Mokkan]

You should have battery voltage back there.

I was curious about this - it is not a misprint or fat fingering. Is the voltage supposed to be 12v all the time, or does the ECU ramp up voltage to match fuel demand?

 

[undefined0000]

I was curious about this - it is not a misprint or fat fingering. Is the voltage supposed to be 12v all the time, or does the ECU ramp up voltage to match fuel demand?

Well with stock wiring and the car on, you should get somewhere around 12v, but the cars tend to charge betwee 13-13.5 volts, thats what you should be getting back there. voltage stays consistent, FPR is what gets used to deal with maintaining proper fueling in the fuel rail. ECU pretty much tells the fuel pump to turn on when the engine is spinning.

 

[troubleshoot]

Yeah I missed that. Thats definitely not right. My 2g charges at ~13.4. If your fuel pump isn't getting enough or constant voltage id check that out first before fiddling with the tps. Although I do love me a properly adjusted tps...

 

[Mokkan]

Well, I started following the power supply path for the fuel pump to figure out where my missing 6 volts had gone to when I found this:

You must be logged in to view this image or video.

A previous owner has spliced into the MPI pump relay pin 7 (ECU pin 56) with a wire that disappears into the harness. They have also spliced together some ground wires (I think) with bare wires, twisted all of them together and bound them with electrical tape before jamming them back into the harness.

Any guesses as to why someone would want to splice into the pump relay coil control?

Another thing I noticed is that many of the wire colors do not match the diagram on ECm Tuning (http://www.dsmlink.com/images/forums/1GECUPinout.pdf). Do the 1990 models have different wire colors from 91+?

 

[serviceguy]

Looks like a very badly executed attempt at a fuel pump rewire. To answer your question, the ECU actually adapts the the fuel injectors behavior to the supplied voltage and not the other way around.

 

[WES_393]

Well with stock wiring and the car on, you should get somewhere around 12v, but the cars tend to charge betwee 13-13.5 volts, thats what you should be getting back there.

Not exactly. The voltage starts out as 13-13.5 but usually ends up 1-1.5v lower at the pump. That's the whole point of a rewire- To get full voltage at the pump.

A previous owner has spliced into the MPI pump relay pin 7 (ECU pin 56) with a wire that disappears into the harness. They have also spliced together some ground wires (I think) with bare wires, twisted all of them together and bound them with electrical tape before jamming them back into the harness.

From what I can see, that's factory work. The O2 sensor and knock sensor will have wire shield around them, which is then bundled and clamped to a sensor ground input. Then just wrapped in grey/tan tape, right? That's just the fine workmanship of Chrysler. So I'd leave those two clamps alone.

As for pin 56, that goes to Pin 7 on the MPI relay which is the ground side of the fuel pump relay. It's how the ECU turns the pump on/off. I'm not sure about the wire spliced into it or not. Maybe someone with 1g experience can chime in.

 

[steve]

As for pin 56, that goes to Pin 7 on the MPI relay which is the ground side of the fuel pump relay. It's how the ECU turns the pump on/off. I'm not sure about the wire spliced into it or not. Maybe someone with 1g experience can chime in.

The ECU switches that pin to ground to activate the relay. You might tap into it to get the signal for the fuel pump rewire or to add a switch to control the pump.

 

[Mokkan]

From what I can see, that's factory work. The O2 sensor and knock sensor will have wire shield around them, which is then bundled and clamped to a sensor ground input. Then just wrapped in grey/tan tape, right? That's just the fine workmanship of Chrysler. So I'd leave those two clamps alone.

Well, the electrical tape was the generic black sort, not tan or grey, but I can bundle it back up and put it back where it was.

As for pin 56, that goes to Pin 7 on the MPI relay which is the ground side of the fuel pump relay. It's how the ECU turns the pump on/off. I'm not sure about the wire spliced into it or not. Maybe someone with 1g experience can chime in.

The ECU switches that pin to ground to activate the relay. You might tap into it to get the signal for the fuel pump rewire or to add a switch to control the pump.

Right, the wire is grounded to make the pump run. It just seems like a really weird wire to tap into. There are perfectly good actual ground wires right next to it. Oh well, I guess I'll find out when I dig into it after I get home from work.

 

[steve]

Right, the wire is grounded to make the pump run. It just seems like a really weird wire to tap into. There are perfectly good actual ground wires right next to it. Oh well, I guess I'll find out when I dig into it after I get home from work.

Where does the other end go?

I doubt anyone was using that to obtain a ground so that there are real ground wires nearby is misdirection.

If the other end goes to a switch then it's intended to bypass the ECU operation and activate the fuel pump with said switch.
If the other end goes to a relay it's to allow the ECU to activate that relay like it did the Fuel Pump side of the MPI relay. Usually for the FP rewire you pick up the original power output from the MPI relay and used it to drive the high side of the relay coil with the low side grounded all the time.

 

[Mokkan]

OK, I was looking at the wrong plug

I was able to find the MPI relay output and I rewired it to the pump with 10 gage wire. The pump is now getting a solid 12v! The bad news is that now the check engine light is on, the windows no longer work, the car still doesn't pull power above 4000rpm and it overheats.

I'm off to get an analog voltmeter to see what the check engine light is all about.

Update: I checked the OBD port and it gave me 4 long pulses and 2 short, which indicates a fuel pump issue. I'm assuming I bypassed the voltage sensor when I ran my own power line to the pump so it thinks the pump has no power. Will the ECU react to this or will it behave normally and just show the CEL?

 

[Mokkan]

Well, I've been messing around with the car and I may have solved the idle issue. The ground connection between the throttle body and intake manifold was not attached (the little useless looking bracket that goes with the bolt through the elbow). I assume that the idle switch was not completing its circuit, so the car had no idea it was supposed to idle.

The power loss at 4k rpm has not been fixed, though. I got a new TPS, however, its resistance readings are nearly identical to the current one. Oh well, I guess I'm out shipping to return that one.

Out of ideas, I opened up the ECU I just purchased and there is some hardened gunk on a few components. Has anyone seen this stuff before and does it indicate an issue? Its totally hardened - not sticky or pliable at all.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

 

[steve]

Looks like white Silicone but it has no business being there.
Is there anything on the bottom side of the top cover since you have it off?

 

[Mokkan]

Yes, there are some white circular splotches. I can post a pic after I get home from work today.

Also, the coil pack did not ohm out correctly, 1.3 and 2.6 ohms, rather than 0.7-0.86. Also, also, the pins on the power transistor are pretty corroded. I found a good deal on a new transistor and I'll pick up a new coil pack today and see if it measures any better - if not, I'll just return it.

Here's a pic of the cover & board:

You must be logged in to view this image or video.

I also removed the spark plugs and #2 plug is different from all the rest. This pic shows #1 and #2. #3 and #4 look just like #1.

You must be logged in to view this image or video.

I also touched the idle position switch and it broke . Can these be purchased separately or am I in for a new throttle body?

You must be logged in to view this image or video.

 

[Mokkan]

Woohoo! The car's finally running. I found an idle position sensor in a 1998 Hyundai Accent, adjusted the throttle position sensor, and replaced the power transistor. The car now pulls well through first and slowly works its way through 3rd. Now, I just need a transmission that will let me use all the gears...