I have the P0420 code on my 1998 Honda Civic DX (catalyst system efficiency below threshold bank 1). I want to be able to properly diagnose the problem myself without just throwing new parts at it. A new catalyst is fairly expensive. I used a scan tool to log some data from the O2 sensors as I drove around. My question is how do I interpret this data and can I determined that it is the catalyst and not the O2 sensors that have failed?

First I let the engine idle for 15 minutes to reach operating temperature. This is the graph of the O2 sensors at idle:

Graph of O2 sensor voltage at idle

From my research I can say that the upstream O2 sensor (red) look okay. It is fluctuating between lean and rich conditions. The downstream sensor (green) is stuck at around 0.9V. I know that the sensor is supposed to be a flat line around 0.5V so this is a problem. It is not clear to me whether this is due to a failed catalyst of a failed sensor.

Here is a graph of normal driving:

Graph of O2 sensor voltage under normal driving

Here is a graph of hard acceleration:

Graph of O2 sensor voltage under hard acceleration

Finally, here is a graph of coasting down a hill. You can see where the computer shuts off the fuel injectors.

Graph of O2 sensor voltage while coasting

These graphs make me think that the downstream O2 sensor (green) is not faulty. Especially the graph from coasting. When the computer shuts off the fuel injectors, the intake becomes extremely lean and both sensors respond immediately. If the downstream sensor was faulty, it should stay stuck at a high voltage. Is this interpretation correct? Can I conclude that I need to replace the catalyst?


Excellent graphs. The after catalyst sensor shows normal mixture readings. The PCM can and will operate the system lean or rich depending on operating conditions. The voltage will rarely be held at any intermediate voltage, such as 0.5 volt. This is an urban legend not based on engineering fact. The rear sensor voltage should remain steady at any steady rpm. It can be rich or lean but should be a flat line. The graphs here indicate rapid mixture change variation which strongly suggests catalyst failure.

Several other test need to be completed to make sure that problems in other areas are not present. A more detailed process of catalyst testing follows.

1) Verify proper fuel control. This is best done by calculating Lambda from tailpipe gas analyzer readings. Since this is not commonly available most will substitute fuel trim, make sure that the trims are less than 10% when long term is added to the average of short term.

2) Verify that there are no exhaust system leaks anywhere near the mixture (oxygen) sensors or the catalyst. Leaks will cause false readings.

3)Verify that both sensors are in good condition. That they will move quickly from rich to lean when the mixture is changed. Usually this is done with a snap throttle. They should both go below 0.1 volts and above 0.9 volts.

4) Monitor the pre and post catalyst sensors at 2500 rpm with the engine and catalyst fully warmed and held at steady rpm. The rear sensor should stay smooth. It can be at any voltage and can change voltage but should do so very often, more that 30 seconds.

  • So the fact that the post catalyst sensor voltage changes at all indicates that the catalyst has failed? – Robert Stiffler Mar 15 '16 at 1:32
  • In general yes, see my edits of the answer. – Fred Wilson Mar 15 '16 at 4:29
  • Actually, I'm pretty sure I've read that the downstream sensor readings shouldn't necessarily be a flatline. The point is that they should be a heavily filtered copy of the upstream sensor readings, I'd say at idle it looks great, not so with the normal driving one. Are you absolutely sure about what you just wrote? Also, unrelated, but the graphs seem to indicate that the engine is running a little lean - there is a slight shift towards 0V. – I have no idea what I'm doing Mar 15 '16 at 8:45
  • @IhavenoideawhatI'mdoing A high quality catalyst should clean up all but the largest fuel mixture excursions. I have studied hundreds of pre and post sensors graphs in my 30+ years as an automotive diagnostician. I have an Mechanical Engineering degree and teach on this subject at training conferences. That said, my comments here are a simplification. This subject is complicated and has lots of folk lore and myths. Lean or rich is OK, it just an indication of what the engineers wanted. Honda liked to vary mixture rich to lean about once a minute in cars of that era. – Fred Wilson Mar 15 '16 at 15:59
  • 1
    Fair enough. So if the graph is a little wavy, does that mean the cat is on it's way out? – I have no idea what I'm doing Mar 15 '16 at 16:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.