- This topic has 35 replies, 5 voices, and was last updated 10 years, 8 months ago by
.
-
Topic
-
When the engine warms up, it loses ignition advance! It seems that the signal from the cam position sensor is not getting to the computer, so, it defaults to no advance. It doesn’t give a “check engine” light; reading the pre-OBDII codes doesn’t show any faults. I’ve tried a new cam position sensor – no change. I’ve also groomed all the connectors related to the computer & sensors. I suspect a wiring harness issue, although the resistance readings (cold) agree with the factory manual.
Anybody know of favorite places to look for wiring issues? Or have a better diagnosis?
RoyL
-
Replies
-
A check with a timing light showed that the ignition timing is retarded far more than the manual calls for at idle. Since the ignition scope showed reasonable secondary waveforms, I didn’t see any benefit in trying to look at primary waveforms. It’s hard to conceive of something that would give normal secondaries from faulty primaries.
I did a casual check for vacuum leaks, mostly a visual inspection of hoses. Tomorrow is supposed to be warm enough to do a little work outside – it’s easy enough to throw a gauge on it. I’ve also got a nifty mechanic’s stethoscope with a bell fitting that’s great for finding vacuum leaks.
I thought of cam timing, so I popped the covers off to make sure the various timing marks were still aligned. The timing belt is only about a year old. I changed it as part of repairing the bolt hole in the end of the crankshaft. (The threads in the crank had stripped, letting the accessory drive pulley shear its key & spin freely. There’s enough room to re-tap to M14x1.5 without having to Helicoil the crank.)
The more I think about it, the more I think there is either a vacuum leak or something going on with the idle air control valve that’s forcing the computer to retard timing to bring idle speed down. That also fits with it being more driveable at higher rpm/larger throttle opening. If I was somewhere other than Virginia Beach, I’d try going uphill with it 😉
RoyL
Like I’d said several posts ago, this car doesn’t have a knock sensor. I was surprised to find that there isn’t one on this engine.
RoyL
Progress(?) update:
I’ve checked the temp sensor & wiring – they’re OK. Changed the PCV valve (a new one is cheaper than the amount of solvent it takes to clean an old one!) I’ve gone over the whole intake system with a stethoscope without hearing a leak. The one area that sounded odd was the idle air control valve; now that that’s clean it doesn’t sound odd, seemed to help a tiny bit. (It took a lot of little strips of index card & solvent to get all the crud out.) Also cleaned the throttle valve & intake throat; not positive I don’t need to take the whole assembly off so I can actually see what I’m doing without a mirror.
It’s still running very retarded timing at idle to get idle speed down to spec – once it gets above 2000 RPM it begins to act normally. The span between idle & 2K is pitiful – when I test drove it around the neighborhood I almost got rear-ended by a jogger. The deeper I get into this, the more it seems there’s something letting unmetered air into the manifold. That makes me wonder if I need to clean the air mass meter – what are the symptoms of a dirty air mass meter? If it’s dirty, what’s the best way to clean it?
RoyL
There is a part listed for the knock sensor on your vehicle.
It may be difficult to find but it’s in the same location as in the videos I posted for you. It’s also a known issue on Subaru’s.
For the 3rd time; this is a ’93 Impreza – there’s no knock sensor! It’s not that it’s hard to find or reach; that engine doesn’t have one. The genuine Subaru shop manual wiring diagram for the ECU confirms that there isn’t one as does the text describing all the engine control & construction features.
RoyL
My wife has a 2004 Impreza WRX we bought used. Its been a nice little car but not a lot of power. The other day it needed power steering hoses which I installed. In the process I discovered a clamp loose on the intake. I put it back together and now it flies.
So, this car limped around for the 38K we’ve had it never setting a CEL or doing anything unusual until I accidentally fixed it. So, sorry you are having so much trouble but crazy Subarus.
Replaced the O2 sensor, no improvement. It was due for replacement; the car has over 100K mi on it. At this point, about the only thing I haven’t replaced/groomed is the throttle position sensor! I still think there’s something basically simple that we’ve all missed & will make us slap our foreheads after the solution is found.
Recap: when thoroughly warmed up, car is very reluctant to accelerate from idle to about 2K rpm; over 2K, it’s fine. When cold, it performs fine; when the temp gauge first settles at normal operating temp it’s still OK, starting to act sluggish. After a little more time, it acts up. When it’s warm, the ignition timing is severely retarded (by the ECU) which is what’s making it run so poorly after it’s warmed up. I’m pretty sure there’s not a vacuum leak – did a thorough check with a stethoscope; when that didn’t show anything, used an ultrasonic leak detector (which I don’t usually use because it finds leaks that even a bubble test says aren’t there) followed by a water spray test. There still isn’t a knock sensor on the engine, unless the factory manual is wrong in the description of engine controls.
RoyL
Despite your insistence that you don’t have a knock sensor, I’m still include to disagree. Mainly because I can find several part listings for it. Engines of that vintage use knock sensors. It’s like standard equipment on a fuel injected engine, especially on a 93. I’ll grant you that it is difficult to find and spot. I think looking for a knock sensor is less of a waste of time than replacing a TPS that you don’t have any evidence that it’s bad. How about posting a pick of what’s under the throttle body in the location of where the knock sensor should be? As I stated, knock sensors are known problems on Subaru’s of that vintage and when they go bad, they can cause hesitations because when they’re faulty, they adjust the ignition timing inappropriately.
From what I can find, the US version of the ’93 Impreza doesn’t have a knock sensor, versions sold in other countries apparently do have one. Checking common parts suppliers (Auto Zone, Advance Auto Parts, Pep Boys & NAPA) show that there isn’t one; oddly, O’Reilly Auto Parts claims there is one! Like I mentioned, I’ve got the genuine, official Subaru service manual for the car; it confirms that there’s no knock sensor.
Where would the knock sensor be if there was one? It’s entirely possible that the manual I’ve got has an uncorrected error. The only reason I’m inclined to believe it is the lack of a knock sensor on major auto parts sites. I usually use Auto Zone & Advance, mostly because they’re the closest to my house – I’ve never run into an accuracy issue with Auto Zone, I’ve had some problems with Advance’s accuracy. I haven’t used NAPA or Pep Boys enough to have a feel for their accuracy; O’Reilly just opened, I haven’t even been in the store yet.
The car’s behavior fits with a knock sensor issue. If there was one, I think the dealer would have caught it when I took the car in to have the pre-OBDII codes read. Once it’s warmed up, the timing is severely retarded, which makes it undriveable.
RoyL
This video shows the knock sensor location, it’s connector, and how to check it.
Additional information in this video.
I spent a fair amount of time rechecking the engine (& manual) – there is no knock sensor on it. That fits with most of the US parts suppliers not listing one; it’s not on US spec cars. Now that it’s been checked multiple times, let’s acknowledge that this engine doesn’t have a knock sensor.
The basic problem persists – when the engine is at normal operating temp, the timing is severely retarded at idle. Because of the retarded timing, the car is pretty much undriveable. If engine RPM can be coaxed over 2K, the car behaves almost normally. I’ve tried different cam & crank position sensors, coils, ignition modules & ECUs with no change. I’ve also cleaned & groomed all the relevant electrical connectors, verified that fuel pressure is normal & that there are no vacuum leaks.
Still looking for other possible causes of the problem.
RoyL
After checking everything in the engine control electronics & finding no problems, I think what’s really wrong is leaking valve lifters. The engine is OK when it’s cold, has problems once the oil is warm. That fits with the check valves inside the lifters leaking; it also fits with them working less poorly at higher RPM. Before I replace them, I’m giving some snake oil a try – I replaced 1 qt of oil with Marvel Mystery Oil to see if that’ll fool the lifters into working properly. I suspect what I’m really doing is delaying replacing them by a few days.
RoyL
After a few delays, replaced the lifters – no change in symptoms. Since I still have no power below about 2200 RPM + occasional “coughing” thru the intake at low rpm, I suspect the basic cam timing is off. About a year ago, I had to re-tap the crank after the pulley retaining bolt hole threads stripped out. With the bolt loose, the key sheared & chewed up both the cam & accessory drive pulleys – I replaced both of them & tapped the crank to M14x1.5 as a repair to the stripped M13x1.5 original opening.
Anybody got any thoughts that might save me from going down another dead end route?
One of the things that slowed the lifter replacement was breaking one of the rocker shaft retaining bolts when my (cheap) torque wrench didn’t indicate it was anywhere near properly tight. Murphy being in control, the bolt that broke was the lower left rear one – the one where there isn’t room for an angle drill to fit between the fender & the engine. By adding a piece of hex stock to a drill bit, I was able to use an angle screwdriver to drill the broken bolt deeply enough to get it out with an extractor. The torque wrench was replaced under warranty, at the same time, I picked up a really nifty “torque adapter” – actually, a digital read out torque indicator.
http://www.harborfreight.com/digital-torue-adapter-68283.html
I’m impressed with it! Despite what the specs say, it actually reads much lower torque values with no problem. It’s kind of blorky to actually use when tightening something; it’s great for setting a torque wrench that’s of questionable accuracy before using it. I’m apparently death to torque wrenches – half of my Craftsman ones have failed after the warranty expired, with repair costs exceeding the price of new ones! This is for things I consider precision tools – they’re kept in their cases & only taken out for specific uses.
Roy
- You must be logged in to reply to this topic.