The interesting thing about automotive repair is that there are so many moving parts, and we can’t always assume that common trends or problems will manifest themselves in the same way.
Just take the 2002 Subaru Forester L we recently worked on. The owner of the vehicle, which has logged 103,027 miles, brought it to the shop because the check-engine light came on.
After pulling the diagnostic codes, we saw a P0130 (sensor circuit malfunction, bank 1) and a P0171 (system too lean, bank 1). When I took a step back and looked at the codes, I thought the bank 1 oxygen sensor (P0130) might be going lean, thus causing the lean code (P0171).
However, I have been doing this for way too long and know that we always have to look at the data first.
When a code is set, it is accompanied by freeze-frame data that tell the technician the environmental conditions of the code – for example, speed (mph), temperature, revolutions per minute, position of the throttle, air flow volume into the engine, fuel trims (how much fuel the intake system is using), and so on.
Cracking the code
When I looked at the freeze-frame data, I saw that the P0130 code was triggered after the car was warmed up and hit freeway speeds. I could also see that the car was adding a lot of extra fuel in the fuel trims.
We first checked all of the basics (vacuum leaks, oxygen sensor connector, MAP sensor connector, corroded grounds, control module updates and general engine condition). Everything checked out OK, yet there was an engine control module (ECM) update for the P0130 code. I had to perform the ECM update, but because the freeze-frame conditions were perfect for causing the P0130 and the P0171 codes, I decided to clear them and test-drive the car.
I cleared the codes and moved the car in front of the shop. I was getting ready to take it for the test-drive when I noticed that the idle was really rough. Before I could get it out of the parking lot, the check-engine light went on again. So I pulled the car in and retrieved codes. The only code that came up was P0130. That’s when I knew that I had to do several things – in the correct order.
Nearly all of the 2.5 Subaru motors have a problem with carbon buildup in the throttle body, which can cause a rough idle. I then called the customer and told him that I wanted to clean the throttle body and perform the ECM update for the P0130 first before doing anything else.
The throttle body was dirty, and the cleaning helped the idle a lot. I then test-drove the car after the ECM update, and the P0130 code came back within a mile. I knew that the ECM was updated and the wiring and connector at the bank 1 oxygen sensor were OK.
I called the customer and told him that I needed to replace the bank 1 oxygen sensor. After doing that, I took the car on two extensive test-drives (20 miles each) to set all of the monitors. On the 19th mile of the second test-drive, the check-engine light came on again. I pulled the code: P0171 (lean condition).
MAP-ping out the problem
Because we had already tested the car extensively and knew that there were no vacuum leaks, it was a sure sign the manifold absolute pressure sensor – or MAP sensor, which tests the air pressure inside the intake manifold – had to be replaced.
As I wrote at the start, there are trends in repair procedures, but they don’t always manifest themselves in the same way. It would not be unusual to clean a throttle body, change an oxygen sensor or reprogram an ECM on a Subaru 2.5L, but it would be unusual to replace a MAP sensor. In the past 28 years, I have changed maybe only two MAP sensors on this model engine.
Once we pulled out the MAP sensor, we found that the measuring element was dirty and damaged. We replaced the MAP sensor and drove the car on several test-drives for a total of 60 miles. All monitors passed and the car was fine.
So this car had a dirty throttle body, a bad or worn oxygen sensor, a bad or worn MAP sensor and an ECM update – all of which we discovered by following procedures and performing tests – including test-drives – in the proper order. If I hadn’t spent so much time on the final test-drives, I may not have found the MAP sensor problem on the first visit.