UPDATE 2: The misfiring part has NOT been solved. The expensive repairs were not necessary. The real reason the car died the first time and the second time (yes there was a second time after the repair was completed) was because I had run out of gas. I went back to the shop and was advised to put some gas in the tank. I also found out that the whole extent of the "diagnostics" consisted in putting in some gas ($5 worth, I was told) to start it. I specifically asked what error codes the ECU had returned, to help me diagnose the low idle issue and it transpired that they hadn't even attempted to plug in a diagnostic computer. ("the car's too old for OBD2"). When I found out the problem was running out of gas, I calculated that I'm getting 4 miles per gallon. Since the gas gauge was intermittent, and I was used to a range of at least 350 miles, I used that as a guide.

Since then I have replaced the dashboard and now the fuel and temperature gauges work, but this hasn't solved the low idle problem or the low mileage. The manual says to check the timing for poor gas mileage, but since this requires that the engine idles normally when warm, I need to first find the cause of the low idle.

Except for very minor misfires (worst when cold) and the low idle that results in stalling when the engine is warm, there is good power and acceleration is smooth.

I will dispute the charges for most of the large bill, but how can I find out the cause of the low idle? The indisputable symptoms right now are:

* horrible gas mileage (4 mpg), probably resulting from a super-rich mixture that in turn is causing the misfires,

* the fact that on cold startup the engine used to rev up to 2000 RPM before it dropped down to about 800 when warm and now it starts at about 800 and drops to a barely stable 400 or so RPM when warm,

The obvious things I have already replaced, as described above, including the idle speed controller. Now that the fuel and temperature gauges work, it also means that the temperature sensor also works. I still think that some sensor is telling the ECU incorrect data, but which one? Any ideas?

Thanks.

 

UPDATE 3: Throttle Position Sensor. The manual says (page 14-30) that the connector has 8 pins. This is not true. The vertical connector has 4 pins and I measured as follows:

Numbering from top to bottom.

1 ------------------------------\
2 --\
3 4 ohm to 4.5 K ohm 4.5 K ohm
4 --/---------------------------/

The manual says it goes from 0.5 K ohm to the value between 1 and 4. When at idle, it appears to be shorted (4 ohms and then as the throttle opens up, there is a jump with a gap. Based on this, I ordered a new one. But another measurement a few days later showed no low resistance. When at idle the resistance between pins 2 and 4 was close to 0.5 K Ohm and varied slightly when the sensor was rotated by loosening the mounting bolts.

Most importantly, unplugging the sensor with the engine running did not cause any change in RPMs or general engine sound. Is this normal?

There is 5V on one of the pins that goes to the sensor and 4.6 on another.

Just in case I replaced it with a new one, thinking that even if I had not messed up the original measurement and it meant some intermittent or transient problem, better be safe. It didn't make any difference. The misfires are still there and it stalls when warm. Turning the Throttle Position Sensor to adjust the idle RPMs has no effect that I can see.

Still getting around 4 mpg. 1/4 of a tank and only 30 miles.

I did try the timing light and can't make sense of the white marks in the little window near the crankshaft pulley, but the fact that I can see some marks seems to imply that the timing must not be off by a lot, if anything at all. Now I'm wondering if the timing belt might have skipped. Since I had it replaced about 7000 miles ago, along with a complete valve job then (it broke) it would be extremely aggravating if this were the cause of all these troubles.

Any ideas out there?

 

UPDATE 4:

The manual appears to say (page 14-76) that there is an "INSPECTION PROCEDURE - (USING VOLTMETER)" by testing certain pins of the "Diagnosis connector". The picture displayed is like this: (the formatting is messing up all the spacing, so instead of the arrows I had originally, the "m" pin is the "MPI-Diagnosis" and the "g" pin is the "Ground" pin.)

-----------------------
| | |X|X| | |m|
------------------------
|g| | | | | | |
------------------------

But what I see is this:
--------------------
| | | | | | |g|
---------------------
|m| | |x|x| | |
-----------------------

So I assume the correct pins are as marked above. Yet after connecting the
oscilloscope accordingly, I only see random noise of less than 1 mV. The "DIAGNOSIS CHART (FAULT TREE)" has a table of 14 items with various square wave forms that indicate the failure of various sensors. I see nothing like this, in fact the "check engine light is not on (after the few seconds after ignition on). I don't know what to make of the random noise, item 14 of the chart: "Normal state" is a uniform square wave. All the other outputs have a combination of wider and narrower pulses. Am I using the correct pins? Does anyone know anything about this?

Thanks.

 

UPDATE 5:

Not sure what I did wrong before, but after double-checking the connector (it only has 3 pins) and trying again to read the "MPI-diagnosis" pin with the scope, I got the attached waveform. The manual says that this "Diagnosis item" is the throttle position sensor. I just replaced it. In the "Check item (Remedy)" column it says that these items could be the culprits:

* Harness and connector
* Throttle position sensor
* Idle position switch

On page 14-71, the manual refers to "Throttle position sensor (with idle switch)", so presumably the latter 2 are both in the part I just replaced. On page 14-65 there is a detailed procedure to adjust it (that includes a feeler gauge of 0.65 mm) so I'll do that next, although I have a hard time believing that a tiny adjustment like this will fix the very serious and noticeable idle RPMs and stalling, as well as the 4 mpg gas consumption. In the next page there is also a procedure to measure voltages at the ECU, so presumably that will take care of the "Harness and connector". The idle position switch inspection (on page 14-85) has yet another detailed procedure. On page 14-95 there is more about the throttle position sensor.

Assuming I'm reading all this correctly, the last remaining possible big problem is a broken wire in the harness, which I'll discover when I measure voltages at the ECU, and adjustments that will affect the idle position switch. I'll report back when I've gone through all these procedures.

 

UPDATE 6:

It appears I had messed up the adjustment of the throttle position sensor at some point, but that wasn't the cause of the problem. Now that I have adjusted it so that the idle position switch is conductive at idle, the ECU reports "Normal state" (after erasing the memory by unplugging the ECU for about 10 seconds).

At the socket that plugs into the TPS, there are the following voltages (with ignition on, relative to pin 1/ground): pin 2: 4.66V, pin 3: 0.02V, pin 4: 5V. The numbering of pins is based on the diagram of the TPS pins on page 14-95 of the manual, but there the pins are labeled: 1: Ground, 2: idle position switch, 3: TPS power, 4: TPS output. It seems strange that there is no voltage on pin 3 when the manual says this is the TPS power. Should there be some voltage on pin 3?

I have checked that there is vacuum by putting a vacuum gauge on the hose that goes from what appears to be the "Purge control solenoid valve" to the throttle body. 10 in Hg at idle, about 900 RPMs, engine cold with lots of misfires. The purge control solenoid valve works when I place 12V across it. This has no effect on the idle RPMs or the misfires. When is the purge control solenoid valve activated?

I seem not to have an EGR system (manual transmission and vehicle not sold in California). I only have the "Purge control solenoid valve" and only one vacuum hose coming out of the throttle body. Am I correct?

Yet the symptoms of a bad EGR include "Rough idling", "Poor fuel economy", "Engine stalls while idling" (see https://mechanicalboost.com/egr-valve/) but I don't see "Check engine light" - in fact the ECU claims there is nothing wrong, or "A reduction in engine performance" - acceleration is good, or "Engine knocking". In the past (but not now) I have noticed the "Smell of unburned fuel" so it's possible I have (or had) "Poor Air-Fuel ratio". But if I really don't have an EGR system, it can't be the cause of the problem. Back to square 1 to look for another cause.

Any suggestion as to what to check next?

Thanks.

 

UPDATE 7

Tested that the purge control solenoid valve is activated when the engine reaches the bottom threshold of operating temperature as displayed in the temperature gauge. Tested for other vacuum leaks by spraying carburetor cleaner on all the hoses I could see at the top of the engine, no change in engine RPMs.

The problem persists: I still get low idle (900 RPMs when cold) with plenty of misfires and horrendous gas mileage. This is really puzzling as the ECU claims "Normal state", which means that it doesn't see any problem with any of the error codes listed in the "DIAGNOSIS CHART": 1 - ECU, 2 - Oxygen sensor, 3 - Air flow sensor, 4 - Intake air temperature sensor, 5 - Throttle position sensor, 6 - Engine coolant temperature sensor, 7 - Crank angle sensor, 8 - Top dead center sensor, 9 - Vehicle speed sensor, 10 - Barometric pressure sensor, 11 - Injector, 12 - Fuel pump, or 13 - EGR, which apparently I don't have.

Any ideas out there?

Thanks.

 

This may not have anything to do with your problem but some of your symptoms are similar to my recent issue.

1) Low RPM around 1K when warming up.
2) ~500 RPM when warmed up. (Rough Idle)
3) Started as intermittent but got progressively more frequent.
4) When it was stumbling it felt like misfire and my scanner agreed. P0300 (I believe)

When I finally got serious, I unplugged each of the coils and one was "not" contributing.
Upon visual inspection of the Ignition Power Transistor / Ignition Control Module I noticed the 3 wire connector had broken at one corner.
When I wiggled then pushed toward the firewall it ran smooth. When I would let go it would be ok for a bit and later would start to stumble again.
I replaced the connector and it ran great at first but then the intermittent cycle hit a couple times. I swapped in an old ICM I had kept for a problem with the RPM gauge.
She's been running great for a couple days now. (No RPMs on the gauge) New ICM shipped yesterday.

So the fix it would seem was the ICM and it appeared the connector played a role as well.
I read a lot of this thread and there's a lot to read but thought I would just throw this out there.

 

UPDATE 8

Thanks for the suggestion. Given that the power transistor was the only part left in the ignition chain that I hadn't tested or replaced, I was already running some tests and I just ran a dynamic test. See the scope traces below.

I only have one coil and I had already tested with a new one. No difference. The connector is solid.

The static test, as described in the manual: Put a 1.5 V battery between pins 1 and 2 (+ on pin 1, - on pin 2/ground) and check resistance between pins 2 and 3 (+ of meter at 2, - of meter at 3). What I should see is "continuity" when the voltage is applied and "non-continuity" when not. Presumably "continuity" means I should see at most a few Ohms resistance, and "non-continuity" infinite resistance or at least M Ohms. What I get is: "non-continuity" - 2.7 M Ohm, "continuity" - -5.9 M Ohm. This doesn't look right. Leaky semiconductor?

The dynamic test: Pin 1 receives the pulse from the ECU, pin 2 is ground, pin 3 goes to the coil. With ignition on, pin 1 has 0 V, pin 3 has 12 V. When running I get "out of range" even when setting the probe at 10x, more than 50v. Is this normal? Trace A is the input from the ECU, trace B is the coil output. There is variation in the width of the pulses. This might account for the weak or intermittent spark I had measured weeks go, in relation to the length of the spark plug wires. The shortest one ran perfectly and as they got longer there were more and more missed sparks, as described above in this thread. Does anyone know if the shape of those pulses is normal? I'll report back when I install the new power transistor already on order.

A note about prices on the web. Many vendors have the original Mitsubishi part (MD 304018) for $250. This seems completely outlandish for a simple transistor. Others on Ebay have (probably) generic parts for $30-$50. Still way too expensive for a very basic semiconductor. I ended up finding one for $22.99 on amazon, probably from China. We'll see.

 

UPDATE 9

Figured out how to zoom on the scope traces. The output to the coil looks quite irregular. Definitely not right.

 

UPDATE 10

Replaced the power transistor/Ignition control module and the misfires are better. They now occur mostly between 3000-3500 RPMs, when cold. But when cold the engine does not rev up to 2K RPMs (as it did before the problem developed) and when it gets warm the idle drops to the point of stalling, so the problem persists. Replaced the coil too but, as before, any improvement due to the coil (if any) is marginal.

Incidentally, I ordered the wrong part at first. After weeks of waiting, I got the part pictured below. Then I ordered the correct one.