There are three issues to be clarified:

1. This is not an oxygen sensor. Any time anyone says it is, or implies that that is how it works, or that this is even its purpose is causing misinformation to be disseminated. And, what is worse, these exact sensors are used as real oxygen sensors in other applications - so it gets tricky as there is already information out there that can be used in conflict to their application as lambda sensors. In the presence of oxygen, they produce no voltage. In the absence of oxygen, they produce no voltage. It is only in the presence of combustibles and not enough oxygen to complete oxidation of them that they produce voltage.

2. In order to meet the conditions above, in a complex gas stream containing both combustibles and oxygen (and carbon monoxide and hydrogen count as combustibles - it is not just the fuel vapor) - the combustibles have to be a low enough concentration that the sensor's natural catalytic action can complete combustion enough to use up all the free oxygen in the gas mix. That is, the combustion has to run to completion in order for the sensor to create voltage in a rich mixture. (It only creates voltage in a rich mixture - oxygen has nothing to do with it.)

3. If the second condition above is not met, the sensor cannot accurately read the lambda state of the gas mix - and will respond in error. In the case of a bad misfire, the catalytic action of the sensor is not sufficient to complete the reactions, and no voltage is created. The ECU dumps more fuel in there until the catalytic action can use up the free oxygen is depleted enough for the combustibles to be sensed.

The idea that nothing will cause the voltage to go low except for oxygen is an odd way to look at a sensor. Sensors are only sensors when they produce a response. (An open wire, for example, also produces no voltage.) Voltage is created by the sensor only when oxygen ions flow through the sensor, and that only occurs when there is something pulling them.

My point is that stressing the absence of response is simply not the way to look at the function of a sensor - and will naturally lead to problems - as there are a multitude of conditions that will cause no response - but only one (for a good sensor) that will cause a response.

This is intended to be aa lambda sensor , and for a lambda sensor, there are only two important gas conditions:

1. Rich - excess combustibles. In this case, (as long as the chemical reactions can be run to completion) the sensor produces a voltage.

2. Lean - excess oxygen. In this case, (again, as long as the chemical reactions can be run to completion) - the sensor does not respond.

(There are those who insist that there is a third - that where the mixture is neither rich nor lean - the balance point of perfect lambda. While this is the theoretical ideal, in actual practice, it both never happens nor is it measureable. If the gas mix is perfectly in balance, and the chemical reactions can be run to completion, the sensor theoretically would produce no output.)

My points are two:

1. This is not an oxygen sensor, nor is it intended to be. (Even though this is a simple way to teach - it does not explain what is really going on - so you are bound to run into conditions where it does not behave as anticipated.) It is intended to be used to tell when the air/fuel mixture is rich.

2. The catalytic action is as important as the intrinsic mechanism of the sensor itself - as it only responds to a deficiency (a need for) oxygen in an equilibrium (chemical reactions completed) condition.

I know that there is a lot of discussion about this sensor - to the point of being absurd. However, it actually has a simple purpose and function, as is pretty easily understood - but not until the misconceptions are cleared away.

That is why I jump on posts that maintain the misconceptions.

Remember that Humberto's original question was a query about the Brettschneider formula - and the fact that he got high CO in (what he calculated) was a lean mixture.

This is an interesting case where both the lambda sensor (due to incomplete catalytic reaction) and the Brettschneider formula (due to the high level of HC concentration) fell apart - and both erred substantially in the lean direction - as the true mixture was quite rich (about 15% rich), but had really terrible combustion efficiency (about 55%, as I recall, as opposed to 95% normally).

So - the point is that even under pretty off-the-wall conditions, the theory behind both the Brettschneider formula and the lambda sensor hold up.

Enough said on my side, I think.

In essence, we all pretty much know how these things work - and are simply approaching the situation from our own perspectives.

Being in the gas analysis environment, though, I look at these topics in a global (any condition goes) sense, rather than just looking at one specific application - so my focus is pretty broad.

Best Regards,

Robert from California