Suicidal Subaru

Maintenance/Repairs
,
1

In the good old days, Japanese pilots would join their ancestors by diving their planes into our aircraft carriers. I think one of these pilots’ spirits is now dwelling in the engine computer in my 2003 Subaru Baja.

When I start it in the morning, it idles normally at about 650 rpm. However, when I’ve been on the highway for a few minutes, the idle rises to about 2000-3000 rpm, which is normally not a big problem except I live in the mountains of Colorado. On the steep downhill dive going into Denver, its like a suicidal dive into the Capital building downtown.

The car has otherwise been fine for its 180,000 miles of loyal service. Being rather cheap (my wife has noted this several times over the years), I have been working on diagnosing the problem myself. I have cleaned the idle air control (IAC) several times, and did resistance checks on the IAC, the engine temperature sensor, the transmission neutral switch, the TPS, and the air pressure sensor in the throttle body. All were great. I have checked everywhere for vacuum leaks using propane and WD40. Couldnt find any leaks, but I currently run the motor with every vacuum port capped to avoid even a chance that a leaky device or hose could affect its idling. The throttle cable has plenty of slack, and I have disconnected the cable going to the cruise control to be sure its not infuencing anything.

I broke down and bought a new IAC thinking it would help but it had no effect, everything was the same… Note that I bought an aftermarket IAC for a lot less that the dealer wanted. It was exact fit part, but maybe you get what you pay for?

The ‘check engine’ light has not turned on, either (except when it should, when you first start it). A friend lent me his sophisticated car scanner and it didnt turm up any trouble codes, but it did show one weird thing- when the car is allowed to warm up in the garage, the scanner shows the idle throttle position stays at 0%, and its idle stays at 650 rpm, all day. When you take it on the road, though, the high idle will manifest itself, and the scanner says the idle throttle position has crept up to 2%, and idle rpm is about 2000. Oddly, though, if you shut off the engine and restart it immediately, it usually goes back to 0% and 650 rpm…

I have verified continuity and voltages on all 6 circuits between the IAC and engine computer. So the IAC seems to be doing what the engine computer is telling it to do.

I can only conclude that the spirit of a long deceased Japanese pilot crept into the engine’s computer sometime when my car was being assembled in Japan, and he is trying to do our country harm by crashing me and it into the Capital building.

Could I be correct in this? what should I do next? A new car is not in the trade space - we are living on my pension, and I am far too cheap, anyway.

2

There’s the throttle positioner as well as the PCM to wonder about. You may not be able to test the PCM problem. Nobody wants to guess on that one but if it could be done for free we would replace PCMs just for fun.

3

Get a can of freeze spray, or a can of “dust off” held upside down so the liquid propellant sprays out will do.
Go out and drive until it acts up. Then spray the throttle position sensor to cool it off.
If the idle drops back down you know where the problem is.

4

My first two suspects would be the TPS is faulty in some intermittent way (like a light switch that sometimes works, and sometime doesn’t) , or that the throttle body is gunked up and not allowing the throttle plate to move freely. Suggest to remove both from the engine and bench-inspect under good light.

5

Question: When you’re doing downhill for that portion of your drive, what gear are you in? Are you using the engine to brake the vehicle for the descent?

6

These are great suggestions! One option I’m looking at is a throttle body assembly from a junkyard… Prices for these on Ebay arent too bad, and you get the entire assembly with IAC and all sensors… Does anyone know if the manual trans throttle body is same as one on an automatic?

pleasedodgevan2 - know any way to get a loan of a good PCM?

circuitsmith -I will chill the TPS with liquid propane and see if it responds. And pull it off and see if resistance vs position is goofy.

GeorgeSanJose - I gave the throttle body and passages a bath with carb cleaner. Looks better inside but no improvement. After I get the TPS off I will check throttle body in good light

Oblivion - when I go downhill, I sometimes do use the engine as a brake by dropping into 3rd gear…What are you thinking of?

7

Before trying to diagnose this any further, reconnect all the vacuum lines.

8

Remove and clean the throttle body, it may be sticking. TPS has a low failure rate on Subaru.

9

@"El Cheapo Diablo"‌

I’m sorry that I don’t have anything to contribute as far as a solution, but I did want to say that the level of detail in your original post is much higher than most people who post about their automotive issues.

10

This is an update on the suicidal Subaru…

98caddy- Thanks! Time on my hands and poverty are reasons for the detailed diagnostics

circuitsmith - I chilled the TPS with liquid propane while it was idling and verified it got very cold, no change in idle rpm or indicated position on scanner. I pulled TPS off to check it out. It was very clean but I could not check resistance (theres no such check in the manual), but voltage output vs position was normal per the Haynes manual.

GeorgeSanJose and rattlegas - I pulled the throttle body and checked it in good light. There was no sticking or freeplay in the shaft, but the backside of the butterfly plate was very dirty so I cleaned everything with carb cleaner and reinstalled it. It ran about the same (see below).

So, with all above done I started it, got excellent idle at 650 rpm when fully warmed up (radiator fans cycled). The Autoanalysis scanner tool said absolute throttle position was at 0-1% open, ign advance was 8%.

I decided to keep the engine and scanner running and went for a drive. After 1/4 mile of street driving, idle rpm was up to 1000, absolute throttle position was at 0% open, ign advance was up to 22%. Same old problem…

I thought that maybe to PCM had to learn something, so I drove on highway around 30 min. When I was done it was about the same - 1000 rpm idle, ign advance was 22%, but it read 2% absolute throttle position.

I shut off the engine, and I restarted it a few sec later. Now its back to a 680 rpm idle, absolute throttle position was at 0-1% open, but ign advance was 13%!!!

Very puzzled! The variation in idle ign advance makes me worry that the PCM may be bad… But, the variation in idle TPS reading makes me worry about it too…

Muy thanks in advance.

El Cheapo Diablo

11

Is there anything else that’s not working properly? Anything at all, lights, gauges, brakes, anything. Is the tach showing the same as the scanner?

12

Is there anyway to disable the IAC? If the high idle problem were to go away, at least you’d know what was causing the problem.

I had a high idle speed problem on my Corolla caused by the IAC and that’s what I did both to diagnose and fix it, just disable the IAC. Been working fine since. No more high idle speed problem. Another problem was introduced of course, the engine idles at too low of speed when cold. My solution, I just hold my foot down a little on the pedal for 30 seconds when I do a cold-start is all, by then it is running good enough I can take my foot off the gas. I’d need to buy a replacement throttle body to fix it correctly, the IAC is not a replaceable item on Toyota’s 4AFE engine.

13

Wentwest - except the air conditioning, everything is working perfectly, including the tacho. AC doesnt work (freon leaked out of seals) but I havent turned on the AC since August -I’m in the high Rockies and its too cold… FWIW, the scanner shows entire catalyst and O2 sensor system working perfectly, the check engine light hasn’t come on, and there are no trouble codes at all when read by the scanner

GeorgeSanJose - I can electrically unplug the IAC pretty easily and I tried it - result was a extremely low (200 rpm) idle and 4 trouble codes from the PCM about four IAC coil failures.

Cause I have a spare IAC, I tried running it with one IAC physically installed, but the other IAC connected electrically only. Result was way increased idle and 3 trouble codes from the PCM about loss of idle speed control…

Thanks, guys

14

Any chance the AC control is turned on? The compressor won’t run with low freon, but the engine idle might be higher because it’s designed that way to provide enough power to turn the compressor.

15

Idle speed problems are one of the more difficult diagnoses. Especially when they are intermittent. So I feel for you OP.

If you can figure out a way to disable the IAC but keep the car drivable, I think that will get you on the path to the solution of this mystery. Idle speed cannot increase as much as you are reporting unless there’s a path for add’l air into the engine. That air has to be coming from somewhere. Since it is intermittent, some change must be occurring that allows more air into the intake manifold in certain circumstances. So the key question is: How is the extra air getting in?

You might start by listing all the possibilities you can think of, consistent with the observation that whatever it is, it is intermittant:

  • IAC goes haywire either b/c it is a bad part, or it is being incorrectly controlled by the PCM.
  • Leak in brake booster diaphragm
  • Leak in power steering idle speed path
  • Leak in AC idle speed path
  • Leak in headlights/rear window defroster idle speed path
  • Throttle plate sticking

You might need to visit your local library and view their MOTOR manuals or if they offer the All-Data computer database service. You need to find every reason the idle speed is increased w/ this engine design.

It’s true an engine with no IAC will be difficult to drive until it warms up, as the idle speed will be too low. Your foot will have to be the IAC until the engine is warm. But if you can reach that situation, and the IAC-less idle works ok when the engine is warm, then you should be on your way to finding the cause. If/when the idle speed increases and you now know for sure it isn’t due to the IAC, then just go through each of the items on the list until you find the one that is causing it. Best of luck.

16

Wentwest- yep, I’m sure the AC is not on. The idle does go up about 200 rpm when I turn it on, tho

GeorgeSanJose - good analysis. I havent checked for brake booster diaphragm vac leak, but will check it tomorrow. The throttle plate wasnt sticking on or off the vehicle. How can I check the idle speed paths you mention above? Are these logic inputs to the PCM?

I have looked in the local library for the MOTOR manuals but none cover Subarus (they do have Haynes and Chilton manuals for Subes but I suspect your talking about something more detailed…). I did find a distant library with “Motor imported engine performance & driveability manual” for 2003, but its several counties away, and they will not loan it to my library.

As far as a vacuum leak goes, the idle vacuum is about 17 inHg (which actually sounds real good to me when you consider that I’m in Colorado at 8000 feet altitude) and steady as a rock, even when the idle is elevated by this problem. The vacuum response to blipping the throttle is normal, too…

Will try driving with disabled IAC tomorrow, too.

Many thanks
El Cheapo Diablo

17

I wish I knew more to help you with this. The ignition advance is still the disturbing issue, because it seems wrong to be at 22 degrees with an idle of 1000. And why is it so advanced? I think the advance is set by the ICM, but it gets info from a number of sensors. You don’t mention any error codes that would explain it, but something is telling the computer to advance the spark, and whatever that is, it resets when you shut down the engine and restart it. I asked about the tach reading because I was wondering about the crankshaft position sensor. But I’m way over my head at that point, and I’m a person who harbors an irrational lack of faith in Subarus generally and in “boxer” type engines, so I’ve got no idea where you should look.

18

GeorgeSanJose- I verified no vacuum leak in the brake booster - warm idle didnt change when I disconnected and capped its vac port on manifold, stayed about 1050 rpm

I then unplugged the IAC. Warm idle went from 1050 rpm to about 200, then recovered and went to a lumpy 850 rpm idle and stayed there. Went for a highway drive this way and it seemed OK, except for lumpy idle whenever I stopped, which was either at 200 or at 850 rpm. Ran scanner in this condition and it reported 0% absolute throttle position, 10 deg ign advance at idle.

Funny thing was, when I opened throttle slightly to get 950 rpm, it jumped to 20 deg advance. Returned to 10 deg advance / 850 rpm when I let throttle close.

I reconnected the IAC and idle was smooth at 1075 rpm, and ign advance went up to 23 deg. Altho throttle was now closed, scanner said absolute throttle position was now at 3%.

Wentwest- nope, there are no error codes at all (unless I force some by unplugging parts like above). I am mystified too, particularly with the wandering advance. Maybe a bad engine computer? But, I hear these never break…

19

“I chilled the TPS with liquid propane while it was idling and verified it got very cold”

El Cheapo, did you “Go out and drive until it acts up” as I suggested before?
The chilling trick is done when it’s failing, then see if the problem clears up right away.

re. the ignition advance:
when it’s working OK press the throttle to bring rpms to 1000 and hold, what are the advance & throttle position #'s?

20

Circuitsmith - when I chilled the TPS, the car was already acting up and had a high idle. Chilling had no affect on idle.

But, I went to check the values when it was working ok. I restarted the already warm car and the scanner showed a varying rpm, about 610 to 670, with ign advance varying between 8 and 9 deg (all at constant 0% absolute throttle position). This seemed to be the dead band in the
IAC… It was sitting happily idling for about a minute, and then the absolute throttle position shown on the scanner suddenly jumped to 2%, rpm jumped to 1050, and advance to 21 deg. No inputs of any kind from me, no touching the throttle, and no clicking of any relay on the car or anything like that.

I loosened the TPS screws, to rotate the TPS and to see its affect on idle speed. With no change in throttle, idle rpm went up when TPS was rotated to indicate more absolute throttle position up to its limit at about 5%… So, if the TPS is spontaneously shifting absolute throttle position, the system responds with more rpm, probably by having the PCM advance the ignition.

This tells me that the TPS must be bad… Any comments?

Thanks, guys