Here is the "Haunted" Sebring.
 

 
 
 
 
 
 
 
Fig. 1 
Here is the special plug required for the scan tool to connect to the PCM. Notice that it uses both and OBD II connector and a Mitsubishi connector.
 

 
 
 
  Fig. 2
This is a close up of the 2 plugs hooked up to the corresponding diagnostic sockets under the driver's side dash.

 
 
 
 
 
 Fig. 3
Hmm...looks like the typical dead sensor stuck at mid range.
 
 
 
  
 
 
 
 
 
 

  
Fig. 4 
What's this? More than a volt on the O2 feedback wire? Alas, our ghost arrives.
 
 
 
 
 
 
 
 
 
 
 
 
 

 
Fig. 5 
Here was my road to nowhere digging inside the main engine harness wiring loom.
 
 

 
] 

 
 
 
 Fig. 6 
Here is the rarely seen O2 sensor ground splice. Unfortunately our ghost was not found here.
 
 
 
 
  
 
 
 
 
 
 
 
 
 
 

Vehicle :       1997 Chrysler Sebring 2dr. Coupe
Powertrain:   2.5 Liter V-6 4Spd. A/T
Mileage:        46,721 
Symptom:      Mil On
 

Problem Overview:

It all started when an employee asked me if I would "take a peek" at his car. Our employee was concerned because his Malfunction Indicator Lamp, (MIL), was on and he wanted to know how serious the problem was. With the growing trend towards intermittent problems these days, I was hoping that this "peek" wouldn't turn into a case of "peek-a-boo". 
 

Vehicle Specifics:

The Chrysler Sebring is sort of a Chrysler-Mitsubishi hybrid product. Since it is a 1997, the vehicle is OBD II compliant and has multipoint fuel injection. Fuel injection is normally carried out once for each cylinder for every two rotations of the crankshaft. Fuel delivery is determined by a speed density system via the MAP sensor. The vehicle utilizes 4 Heated Oxygen sensors. 
 
 
 

Diagnosis:

Starting with the usual under hood inspection, everything easily visible appeared to be in order. No sensors unplugged, no new bodywork, etc. The next logical step with the MIL on was to check for codes and analyze scan data. I feel as though I learn something new everyday and connecting our Snap-On Scan Tool to this car proved that philosophy. On all the prior OBD II cars I have tested, I have only needed an OBD II type connector. This car was unusual in that it requires that you connect both the OBD II and Mitsu-1 connectors in order to communicate with the PCM. They are mounted side by side below the dash between the steering column and center console. (Figs. 1 & 2).  Once the scan-tool was connected I pulled codes. Only one code was logged, P0134 "rear bank up stream H02S stays at center". After a quick check for service bulletins we updated our employee on what we knew so far. He asked us to halt further testing until he could call the local Chrysler dealer. Our employee recalled that about a year ago the dealer replaced one of the O2 sensors under the Chrysler 3 year / 36,000 mile warranty. The question was: Which O2 sensor? The dealer pulled up his vehicle's history and told him that the front upstream O2 sensor was replaced in the past, not the rear, and that the vehicle was out of warranty anyway. Now some of you are probably wondering...3/36 warranty? The new OBD II emission rules mandate 2 years and 24,000 miles new car warranty on all emission related items, right?  (Note that the PCM and catalytic converter are covered to 8/80) Right, but even though the O2 sensor has a minimum warranty of 2/24 the Chrysler 3/36 bumper to bumper warranty overrides. Another warranty tid-bit that causes confusion is the warranty on a replaced part at a dealership and how that effects the 3/36 warranty. Typically, if a part is replaced at the dealership under new car warranty, the part is covered for the remainder of the 3/36 warranty or 12/12 from replacement date, whichever provides greater coverage. I have got a few customers some "free" parts under this loophole. It basically means that if a part is replaced under the 3/36 warranty, then fails after the 3/36 expires, the part still carries a 12/12 warranty from date of installation.  Anyway, back to our O2 sensor problem. A quick read of the datastream revealed that the rear upstream O2 sensor was fixed at .44 volts. 
 
 

A Note Of Warning:

Many scan tools will incorrectly display the names of the various O2 sensors. OBD II terms specify that the heated upstream O2 sensor on the bank with cylinder #1, be called or abbreviated "Heated Oxygen Sensor Bank 1 Sensor 1" Not all scan tools follow this logic. Quite often, the sensors are named left, right, bank 1, bank 2, etc. Worse yet, are the several cars I have seen where Bank 1 and 2 are reversed. Took a bit of hair pulling the first time I got stuck with that one. I wish I could have billed the scan tool company or auto manufacturer for the time I gave away on that job! When is this industry ever going to be organized? And if you think were moving towards industry standardization, wait till you try to call the dealer and order one...Which sensor is considered left on a car with a transversely mounted engine? Be sure to disconnect each sensor to watch for the change in the datastream. This way you can be sure that your diagnosing the right sensor in the first place. 
 

 
Ghost Stories:

The connector for the rear upstream sensor is easily located on the main engine harness support bracket. The wire colors are: BK/DG for sensor ground, WT/BK for O2 signal, BK/RD for heater power and BK for heater ground. Voltmeter testing was fairly cut and dry: 13.5 volts on the heater wire, .03 on the grounds and .44 volts on the signal wire. A shot of carb. cleaner in the fresh air hose did not cause any increase in sensor voltage. Pulling the brake booster vacuum hose did not drive the sensor down either. It seemed like the typical dead sensor. All of a sudden, while watching the .44 fixed voltage value in the datastream, I saw a reading of between 4.21 and 4.98 volts! Wow, almost 5 volts on the sensor signal wire! Then, after a few minutes, the high voltage reading disappeared and it was back to .44 volts. Where did the mystery 5 volt ghost go, and, better yet, where did he come from? I decided not to trust scan data alone and reconnected the voltmeter to the signal wire. Still my ghost was gone, just a dead sensor at .44 volts, (Fig. 3). I shut the car off and went to grab a coffee. Perhaps it was a lack of caffeine, I thought. About a ½ hour later I started the car and began wiggling under hood wire looms while simultaneously watching the voltmeter and scan tool. After wiggling the main engine harness the scan-tool and voltmeter shot up to 4.2 volts, (Fig.4). The ghost was back. I began to theorize that somewhere in the main engine harness the O2 sensor signal wire was chaffed and rubbing against a sensor's 5V reference wire. Sounds possible right? Well, as Fig. 5 will show, I had the entire loom cut open and exposed almost all the way to the PCM. I even found the sensors ground splice but it was fine, (Fig. 6).  Not a single wire was chaffed or damaged in any way. Just out of sheer luck and determination I noticed a peculiar pattern to our ghost. He had a schedule. I started to notice that the 4-5 volt ghost came only once per ignition cycle. A stop watch helped me prove another pattern. The ghost voltage came after exactly 3 minutes of engine run time and only stayed present for 2.5 minutes. Although I found no published information about this, it sort of "rang a bell". In a past Aspire training class I recall my instructor, Bill Cahill, talking about OBD II strategy. When an O2 sensor fault code is stored the PCM can internally apply a reference voltage out the normally input wire to help figure out if the sensor is plugged in or to check the integrity of the circuit. I noticed that if the sensor were unplugged the voltage would still appear on schedule. It would appear that the PCM is intentionally sending voltage down this wire as a circuit test. Had our employee continued to drive the car, the PCM may have set more O2 related codes as it proved out the circuit. 
 
 

Endings And New Beginnings

In the end what I thought was a wiring problem was some little known strategy of the PCM. You might even call it "ghost strategy". I retaped the wiring harness, replaced the O2 sensor and am happy to say the MIL is off. That's enough ghost chasing for one day.