Background:
I have a 2005 Mazda3 2.0l.
Over the past year I had been intermittently getting a MIL caused by P0138 “HO2S RR Circuit High Input”.
To pass inspection I brought the vehicle to a local garage who recommended replacing the O2 sensor. They replaced the sensor, but the problem was not fixed – in fact it seemed to get worse. After resetting the DTC the MIL reappeared after about 70 miles driving. I think I read somewhere that this is a kind of ‘DTC detection cycle distance’ (but I could be wrong). Before I could return to the garage the car was damaged so I took it to (a different) garage for repair. This shop tried to reset the DTC (as a courtesy), but the MIL did not turn off this time.

The ‘original’ local garage recommended first re-flashing and then (if that didn’t work) replacing the PCM (ECU). I figured that a Mazda service shop would be more qualified to handle this so I took the vehicle to them.

Mazda said that the re-flashed the PCM, checked the wiring harness, and ran through the troubleshooting procedure. The Mazda service shop diagnosed the problem as a bad (replacement) O2 sensor. By the way, supposedly the replacement was an OEM part. Upon hearing this I returned to the original local garage because their work was under warranty. They replaced the O2 sensor again and found this not to fix the problem. Again the local garage recommended replacing the PCM (which is very costly).

The original local garage says that the PCM is affecting the O2 sensor output so that it is out of range. They say the signal goes ‘negative’ when it shouldn’t.

I have reviewed the troubleshooting information and I feel that I understand it well – I design/troubleshoot electronics at work.

Some things seem suspicious to me. First, the DTC description is:
“The P0138 code is set when the Engine Control Module (ECM) detects that the rear O2 sensor signal voltage remains excessively high for an extended period of time.”
This doesn’t seem to correlate with the information from the garage. Second, it seems strange to me (but not impossible) that the analog input circuit to the PCM would be affecting the signal in the way the garage describes.

At this point I am considering doing my own testing before jumping on the ‘replace PCM’ boat. I would appreciate some feedback to these questions.

1) Has anyone heard of the PCM input affecting the O2 output signal in the way described by the garage?

2) Does it make sense to disconnect the signal to the PCM (while leaving the other wires connected) to measure the O2 sensor output voltage on an oscilloscope? I plan on comparing the waveform to that with the PCM connected to see if there is a difference. From the sensor description I think the sensor would (ideally) provide the same voltage regardless of PCM connection. Can anyone verify this?

3) The mechanics have not recommended looking a fuel pressure, fuel injectors, Intake air leaks, Exhaust gas leaks although I see these as possible P0138 causes. If any of these were in fact the cause should I expect other DTCs? Or do any of these possible causes throw a red flag for 2007 Mazda 3?


I copied this from another page covering P0138:
Possible causes
- Rear Heated Oxygen Sensor Bank 1 harness is open or shorted
- Rear Heated Oxygen Sensor Bank 1 circuit poor electrical connection
- Faulty Rear Heated Oxygen Sensor Bank 1
- Inappropriate fuel pressure
- Faulty fuel injectors
- Intake air leaks may be faulty
- Exhaust gas leaks

Tech notes
Replacing the rear O2 Sensor bank 1 usually takes care of the problem.

Thanks in advance,
-Toll

 

1) If the PCM was bad, it could cause a bad reading, and it wouldn't be fixed with a new sensor since the circuitry has a problem. Looking at graphs of the O2 sensor output should be able to show something odd happening, but it depends what failed in the circuit.

2) It is a good idea to detach the sensor to look at the raw signal, though I think the O2 sensor needs to be driven with a specific voltage to be able to be read. I have looked at other gas chemical sensors and it's very sensitive circuits to get them to operate properly, so I don't think this is a real feasable option.

3) I would check the other items in that "possible causes" list. Though things like improper fuel pressure or faulty injectors would probably make it run rough / lost power etc...


Something to check is if the wiring on the O2 sensor looks correct. I replaced one on my 2001 forester and the replacment needed the wires switched around to get a good signal.
As for that, the comment you had of "They say the signal goes ‘negative’ when it shouldn’t"... well the signal is positive only, they probably meant to say it drops instead of goes negative. But if it isn't operating properly i'd look at the wiring like I mentioned above as a first step. It would have been good to get readings from the original sensor before it was replaced just to compare.

The other things on that list have poor connections as a cause. This would have changed with replacing the sensor since that connector got moved. But wherever the other end of that harness is going to the PCM should be removed and reseated just incase. There could be corrosion or some other issue on that side of things.

Good luck.
I can get you some "good" reference graphs of the O2 sensor response from a PC program i've got for the car if that would help.

 

Hello Toll.
I wonder if you have found a solution to your PCM problem. I own Mazda 3 2.0L 2007 and have been suffering from the same problem for the last 2+ years.
Multiple O2 sensor and EGR replacements, catalytic converter replacement, wiring harness inspection, diagnostics have not solved my problem.
The last recommendation - replace ECU/PCM. Costing as much as 2000$ it is not a cheap item to just give it a try.
So, any practical feedback from you would be helpful. Thanks!

 

Hi, I would like to know how the shop tested the voltage output, Did they use their scan tool to see the "negative" reading output from the O2 sensor? If so, they could be right.
If I were you, I would backprobe the O2 sensor to see the voltage output, and then using a scan tool (I doubt you will have one, but the garage might be happy to show you to verify the problem) read the O2 sensor output according to the PCM, the voltage output of the O2 sensor should be between 0-1 Volt, and it should cycle up and down fairly constantly.

 

Hi!
Finally the entire engine wiring harness was replaced (piece from the engine to ECU) and it works fine since.
Years of diagnostics and tons of money for various replacements and...

 

i got here too late. keep in mind that the O2 sensors do not get a signal from the ECU; they send a signal (voltage) reacting to the environment where they are. when you see a code stating a high input, it could be a bad sensor or an open in the wiring to the ecu. when you see a low input, it is a shorted sensor or a short in the wiring to the ecu. how to check? car on, fully warmed, unplug the sensor connector and measure voltage at the ground and signal pins. you should have a voltage close to 1v or a little less. the sensor at close to the exhaust manifold is more active than the one after the cat converter. if you get a heater circuit code, then the same applies, but measure the resistance at the sensor heater pins. that heater fails more often than the sensor voltage.

 

Hi,
I have CEL on in my 2007 mazda 3. Got the code PO 138 from PARTS SOURCE for free. I saw in some articles that O2 sensor replacement shouls be the final step. So I did a Fuel system flush and no light for next 3 months...One fine morning it came again. This time I tried SEAFORM Fuel system treatment 2 times and changed my oil into Synthetic. It doesn't make any difference. Then I noticed that when I start up the car at morning the light comes up after 1-2 minutes ( while I am waiting for the car warm up) and It goes off afternoon..!!!( I am a sales representative, I need to stop the car at frequent intervals). I didn't notice any difference in performance. Is this the problem with the sensor?? If YES the how there will be no lights afternoon?? Before changing the sensor should I do something else?

Thanks.