- Joined
- Sep 5, 2022
- Messages
- 1
- Reaction score
- 1
My ex-wife’s 2003 Subaru Forester with 168,000 miles developed a problem this summer – it would crank but not start when it got hot after driving. Specifically, she would start it at home, drive and do some errands and after a stop or two, it would crank but not start after sitting for about 5-15 minutes. She let it sit for another 20-30 minutes and then it would start.
Long story short: I replaced both the camshaft and crankshaft position sensors and it now starts reliably even on the hottest days. The camshaft position sensor cost $59 and the crankshaft position sensor cost $36 at the local auto parts store. Took about 20 minutes to replace the camshaft sensor and about 40 minutes to replace the crankshaft sensor using two excellent YouTube videos for guidance.
Camshaft Position Sensor Replacement:
Even though this is for an outback, it also worked for a 2003 Forester. I also removed the battery for more access to the sensor.
Crankshaft Position Sensor:
I didn’t replace the belts as they had been replaced last year. But you might check yours before starting as this would be a good time to replace belts that need it.
BACKGROUND:
She lives in Grand Junction, Colorado where it gets over 100F in the summer time. This summer, when she would go out on errands, she would stop the car, do the errand, and come out and the car would crank but not start. She could wait 20-30 minutes and it would start and she could continue her journey. If more stops, it often had the same starting problem. She described as like “vapor lock” on old cars where it would crank but not get gas to the cylinders. She couldn’t smell any gas either.
She took it to a reputable repair shop and after troubleshooting for a day, they told her they couldn’t replicate the problem. The ECU didn’t give them any error codes, so a sensor must be failing but they couldn’t help until the sensor failed leaving an error message for them to use as it could be any of several sensors or even the ECU.
Well, that was just unacceptable as it could stall out at any time (like the middle of a busy intersection) and would create just a mess to deal with. So she called her handy dandy ex-husband for help!
After reviewing the internet forums and Subaru Manuals, it became pretty obvious that Subaru’s engine error code reporting system in 2003 wasn’t real robust and it wasn’t going to give an engine error code unless the engine actually was running when it died as it doesn’t store error codes unless the engine is running (I guess there are many codes in transition states during start-up). So, it had to come down to analytical troubleshooting just like the old days in order to fix the problem!
It sounded like something was getting too hot after the engine shutdown with no glycol circulating and no airflow across the engine. When it cooled down, it started, indicating there was some kind of sensor problem or fuel pump or injector problem. Since it ran OK when running, the fuel injector problem sounded less likely than a sensor.
It turns out both the camshaft position sensor and the crankshaft position sensor send signals to the ECU which in turn controls fuel injection and spark. These are both magnetic position sensors. Several internet sources indicated these kind of sensors fail with time, especially if overheated. The ex-husband worked in the oil patch for 30 years and was familiar with these types of sensors in other applications and was aware of how they fail – either they fail “open” causing an open circuit, the resistance of the sensor increases with heat and results in low voltage signals the ECU can’t read, the magnet in the sensor cracks also resulting in low voltage sensor signals or the sensor signal gets “noisy” when hot and thus the ECU can’t read the camshaft or crankshaft position. The other sensors (throttle position, airflow, temperatures) aren’t typically a “start permissive” like a camshaft or crankshaft sensor.
Checked out the price of the sensors and they were cheap enough it was worth a try replacing them to prevent this problem from reoccurring at inconvenient times. YouTube had some excellent videos on how to change both sensors (links above) and it wasn’t hard using typical home auto repair tools.
After changing the camshaft sensor, we went out for a drive and then stopped and let it sit like on an errand. It did start but with some difficulty and when running it was pretty obvious the engine was misfiring because the timing was wrong. Turned it off and let it sit and cool off and then it restarted OK. We went home and changed the crankshaft sensor and the engine started up real nice and ran smoother than it had before. It also had more “get-up-and-go” when driving. We drove around and tried to repeat the problem but it restarted OK every time. After 4 weeks there has been no repeat of the starting problem and the problem appears fixed.
As an aside, I checked the magnets on both the old and new sensors and they both hold about the same number of paper clips – LOTS! I also checked out the resistance of both the old and new sensors at room temperature and they were similar (~2100 ohms). I put the old sensors in the oven at home up to 300F and the resistance increased to about 2250 ohms but that seemed reasonable based on my references. So, I assume one or both of the sensors must have had a noisy signal to the ECU that caused the engine to not start when hot, probably not injecting gas or sparking the plugs. Since the problem has gone away and appears solved, I quit my investigation and thought I would share the experience with you in case you have a similar problem. For about $95.00 in parts and about an hour’s time that any home mechanic can do, it was worth the risk to us to solve the problem and it appears to have worked this time. Hope this helps.
Long story short: I replaced both the camshaft and crankshaft position sensors and it now starts reliably even on the hottest days. The camshaft position sensor cost $59 and the crankshaft position sensor cost $36 at the local auto parts store. Took about 20 minutes to replace the camshaft sensor and about 40 minutes to replace the crankshaft sensor using two excellent YouTube videos for guidance.
Camshaft Position Sensor Replacement:
Crankshaft Position Sensor:
BACKGROUND:
She lives in Grand Junction, Colorado where it gets over 100F in the summer time. This summer, when she would go out on errands, she would stop the car, do the errand, and come out and the car would crank but not start. She could wait 20-30 minutes and it would start and she could continue her journey. If more stops, it often had the same starting problem. She described as like “vapor lock” on old cars where it would crank but not get gas to the cylinders. She couldn’t smell any gas either.
She took it to a reputable repair shop and after troubleshooting for a day, they told her they couldn’t replicate the problem. The ECU didn’t give them any error codes, so a sensor must be failing but they couldn’t help until the sensor failed leaving an error message for them to use as it could be any of several sensors or even the ECU.
Well, that was just unacceptable as it could stall out at any time (like the middle of a busy intersection) and would create just a mess to deal with. So she called her handy dandy ex-husband for help!
After reviewing the internet forums and Subaru Manuals, it became pretty obvious that Subaru’s engine error code reporting system in 2003 wasn’t real robust and it wasn’t going to give an engine error code unless the engine actually was running when it died as it doesn’t store error codes unless the engine is running (I guess there are many codes in transition states during start-up). So, it had to come down to analytical troubleshooting just like the old days in order to fix the problem!
It sounded like something was getting too hot after the engine shutdown with no glycol circulating and no airflow across the engine. When it cooled down, it started, indicating there was some kind of sensor problem or fuel pump or injector problem. Since it ran OK when running, the fuel injector problem sounded less likely than a sensor.
It turns out both the camshaft position sensor and the crankshaft position sensor send signals to the ECU which in turn controls fuel injection and spark. These are both magnetic position sensors. Several internet sources indicated these kind of sensors fail with time, especially if overheated. The ex-husband worked in the oil patch for 30 years and was familiar with these types of sensors in other applications and was aware of how they fail – either they fail “open” causing an open circuit, the resistance of the sensor increases with heat and results in low voltage signals the ECU can’t read, the magnet in the sensor cracks also resulting in low voltage sensor signals or the sensor signal gets “noisy” when hot and thus the ECU can’t read the camshaft or crankshaft position. The other sensors (throttle position, airflow, temperatures) aren’t typically a “start permissive” like a camshaft or crankshaft sensor.
Checked out the price of the sensors and they were cheap enough it was worth a try replacing them to prevent this problem from reoccurring at inconvenient times. YouTube had some excellent videos on how to change both sensors (links above) and it wasn’t hard using typical home auto repair tools.
After changing the camshaft sensor, we went out for a drive and then stopped and let it sit like on an errand. It did start but with some difficulty and when running it was pretty obvious the engine was misfiring because the timing was wrong. Turned it off and let it sit and cool off and then it restarted OK. We went home and changed the crankshaft sensor and the engine started up real nice and ran smoother than it had before. It also had more “get-up-and-go” when driving. We drove around and tried to repeat the problem but it restarted OK every time. After 4 weeks there has been no repeat of the starting problem and the problem appears fixed.
As an aside, I checked the magnets on both the old and new sensors and they both hold about the same number of paper clips – LOTS! I also checked out the resistance of both the old and new sensors at room temperature and they were similar (~2100 ohms). I put the old sensors in the oven at home up to 300F and the resistance increased to about 2250 ohms but that seemed reasonable based on my references. So, I assume one or both of the sensors must have had a noisy signal to the ECU that caused the engine to not start when hot, probably not injecting gas or sparking the plugs. Since the problem has gone away and appears solved, I quit my investigation and thought I would share the experience with you in case you have a similar problem. For about $95.00 in parts and about an hour’s time that any home mechanic can do, it was worth the risk to us to solve the problem and it appears to have worked this time. Hope this helps.