An Engineering Analysis of the P0420 / P0430 Catalytic Efficiency code

Multiple times in the past I've tried to jump in and offer some insight into how to handle P0420 / P0430 Catalytic Efficiency codes, both on this board and on a few others. There is a lot of confusion and misinformation out there. I thought I'd consolidate my understanding in one place, and maybe others with specific insight will share as well. Forgive me in advance if I don't have it all letter perfect, but it's still got to be way better than what we currently have. And while many folks may attempt to give you some practical help, I believe that unless you understand the “WHY”, you won’t appreciate the diagnostic or potential fix information you are being offered. I’ve attempted to write this from a systems engineering perspective, and a look at how Toyota and others have responded to the mandates set forth by the EPA and other government agencies.


In the beginning:
What happens when you start your car from cold? For the first 300 seconds or so, the system is in “open loop” operation. That means very little is being monitored, and the engine air-fuel ratio is determined from a ‘look-up’ table of factory stored and some learned-modified values. It’s better than running with a carburetor, but along those lines when it comes to clean exhaust. During this time, small heaters are running within the front wideband Air-Fuel Ratio (fancy oxygen sensor) sensor and the rear O[SUB]2[/SUB] sensor, because much below 550’F, these gizmo’s don’t work all that well. That’s the ‘H’ you often see in the more technical description of these components (as in HO[SUB]2[/SUB]S – heated oxygen sensor). The cat itself is being heated by the hot exhaust gases, but that heat wouldn’t transmit to these externally mounted sensors fast enough for them to work in time for the switchover to “closed loop” operation.

Interestingly enough, the EPA’s draft OBD-II docs actually called for electric blankets around the cats to speed their heating to enable a 100 or 200 second changeover period, but that would have required a monster alternator. I’m not really sure if there are any heated cats in production today. The EPA did hold out some nice labels (like ‘super low emission vehicle’) to those that had faster warm-up cycles. Maybe if we ever move to 42V electrical systems this would be practical. There are some P0xxx codes for heated catalyst failure, but I doubt you’ll ever see them called out.

[FONT=&quot]Almost warm:[/FONT]
At the 5 minute mark, the cooling system should be reporting at least 145’ F, the heaters on the sensors should have done their job, etc., and if all sensors report seemingly valid data, the system switches over to “closed loop” operation. Now there is a steady flow of feedback data on throttle position, RPM, atm pressure & air temp for calculating mass and thus actual oxygen content of the intake air, front AF sensor’s look at exhaust purity, knock sensor, etc., and from this we calculate loading and injector duration needed to keep the exhaust cycling between slightly rich and slightly lean. We need that cycling in order for the 3-way catalytic converter to do a proper job of dealing with all three of the chief pollutants the EPA wants to control: Carbon monoxide, complex unburned hydrocarbons, and various oxides of nitrogen.

Why cycle? Surely with today’s management systems we could keep it dead nuts at an even 14.7. The problem is that oxidation and reduction are somewhat inverse reactions, and we need periods of lean to gather up and store that ‘excess’ oxygen that we will need when we go rich. But we also know that if we stray too far from stoichiometric ratio, bad things can happen (more on that later). So could we do better than live with this intentional cycling? The short answer is yes, but with more complex hardware that we may see with OBD-III and the true ‘clean air’ engine.

If you were to look at some of the live output data from the OBD monitor port, you’ll see things like Long Term Fuel Trim and Short Term Fuel Trim. This is part of the immediate self adjusting and longer range table building – the Learning Mode, that the system does to improve the precise control around the magic 14.7 air-fuel ratio. And this is exactly the data that you kill off when you attempt to clear a code by disconnecting the battery! You torture your converter every time you do it this way. And on newer systems that don’t use a separate IAC channel and instead rely on a learned value for the main throttle plate, you may have days of severe idle instability and misfires as a result. Buy a reader and clear codes without wiping out these essential stored values.

[FONT=&quot]Sensors & OBD-II:[/FONT]
Let’s talk about sensors themselves for a minute. How reliable is the prognostication in a thrown code? In general, it’s pretty good but far from perfect. A big part of this is the balance between cost and complexity. In an aircraft, you might have 3 sensors in each spot, and the computer uses balloting to determine if there is a real fault, or just bad data. In your car? Other than the drive-by-wire throttle (which contains 2), everything is single. So if an AF sensor is out of spec, it will throw off the mixture. The system really has no way of knowing if it’s a sensor problem or an actual injector, computer, temperature, MAF, etc., problem. By polling all other sensors and using a lot of deductive reasoning to determine that SOMETHING in all the reports just makes no sense, it throws a code. It may or may not be the right response, but something is probably wrong! OBD operational software is up to the manufacturer to write, and some simply have better programmers than others.

And this is where examining the data stream and a keen, experienced eye come into play. Like it or not, in the end it will come down to a certain amount of experience in reading the tea leaves, looking for other systems on the car that might be causing the failure, performing some parts changing, and re-evaluation. Again, I believe that OBD-III and manufacturer proprietary on-board testing and some redundant sensors will help address this.

[FONT=&quot]Continuous vs Occasional Data Taking:[/FONT]
Much of the engine’s sensors are under continuous monitoring, and so codes for most subsystems could be thrown at any time. Two key subsystems are not – Evaporative Emissions and the infamous Catalytic Converter Efficiency. These are evaluated only once per drive cycle, so the opportunity to pass or fail for these are limited to a few precious minutes. For Evap, this takes place when the system is stone cold, about 5-6 hours after shutdown. The Cat Efficiency test takes place sometime after going over to closed loop operation, but only when certain steady state conditions are met. There is a designated speed range, throttle position, etc., that must be met first, before the testing period begins.

3-way catalytic converter performs two primary reactions, known as oxidation & reduction. In the oxidation reaction, carbon monoxide (CO) and unburned hydrocarbons (H[SUB]x[/SUB]C[SUB]y[/SUB]) are combined with free oxygen and converted to carbon dioxide (CO[SUB]2[/SUB]) and H[SUB]2[/SUB]O. Oxides of nitrogen (NO[SUB]x[/SUB]) are “reduced” to become O[SUB]2 [/SUB]& N[SUB]2 [/SUB](& H[SUB]2[/SUB]O if we also have hydrogen remaining from the oxidation reaction). In a monitored reaction chamber, we would ideally have sensors that report on the efficient conversion and percentages of each constituent. Nice, but that would add at least $5k to the cost of your car! So instead we simply monitor the comings and goings of free oxygen in the exhaust, do some fancy calculations to determine the “Catalytic Efficiency”, and call it a day.

[FONT=&quot]The Once A Cycle Cat Test:[/FONT]
In the Catalytic Efficiency test, the ECU actually momentarily leans out the mixture by up to 30% and evaluates the waveform output response of the rear (post cat) sensor in relationship to that of the front AF sensor. It then returns to the stoichiometric point and watches the rear settle out. It then richens the mixture by up to 30% and compares the waveforms again. So it is monitoring the cat’s ability to store up excess oxygen during a lean period and hold on to it, and then release it on demand to cope with an extended rich mixture condition. How much oxygen should it be able to hold and release? Oxygen storage is a separate oxidation/reduction reaction that occurs using a honeycomb section of the cat coated with Cerium or other similar fast reacting metals. Are the actual precious metal surfaces (Platinum & Palladium mainly for oxidation, Rhodium for reduction) where the exhaust constituents meet up with the freed O[SUB]2[/SUB] properly reactive enough that all this oxygen will be consumed, and the exhaust smell like roses?

Each manufacturer matches a cat capacity with a given engine, and determines a maximum amount of oxygen storage they expect to have available for rich mixture oxidation when components are new and all surfaces are fresh. The failure criteria is typically set at 65% of this max value. Again, it’s not an absolute reading, but a calculation based on the comparison of two waveform. The first time it fails this test, a pending code is logged. If it passes the next time (full cold start cycle), the pending code is erased. Generally, it takes two successive failures to log a hard fail and set the light. Interestingly, this criteria doesn’t seem to be absolutely hard and fast, and reading thru test methodology in Ford, Honda and other manuals I see variations in % and number of successive drive cycles.

So, a number of thoughts come to mind:

1) As it is a comparison, the health of the AF sensor could be just as important as the cat and the rear sensor.

2) Rear sensor issues could provide skewed results.

3) Those that have reported intermittent fails when they clear the codes probably don’t have dead cats – just marginal. Fix minor things, and you might be on the road to recovery. Looking at waveform data would help you assess how bad the cats are.

4) Could you screw the test? Sure. Have a heavy foot in the middle of the testing sequence, and you could make it look a lot worse than it really is. How will you know when the test is running? Tricky…. I’ve sometimes felt a sag on a straight, steady state drive a few moments after the transition to closed loop, and suspected that the ECU was in test mode. This obviously wouldn’t be a smart time to floor the accelerator to pass someone! I would think that if you were running a live monitor, you would probably see a sudden change in fuel trim corresponding to the initiation of the test sequence.

5) Some people mention that their car seems to run fine despite failing this test. Probably true, as this test is simply intended to report the state of affairs. The presence or absence of a working catalytic converter (assuming it isn’t clogged, etc.) will have no impact on the running of the engine. However, if the cat test failure is directly attributable to other factors, such as a bad front sensor, a bad thermostat, or a dozen other issues, then the cat test failure just becomes a symptom of a larger problem that will eventually impact total vehicle performance.


In future sections, I'll addresses some of the factors that can short-life a good catalytic converter, and why so many people still have issues after simply changing them out.