Maybe its a German thing. The CO value was ok but the "lambda value" varied between 1.1 and 1.2 and apparently I've got to get under 1.0. The last time I had it tested it varied between 0.97 and 1.15 and that was good enough.


you have to be more precise. Was it the CO that was high, or what?
There is a page on the lambda sensor on my page, and some links on the internet.
The sensor measures the amount of un-burnt fuel in the exhaust, and adjusts the FI computer accordingly. (this is the closed loop) There are many reasons for the CO to be too high.
On the other hand, before you start troubleshooting:
can you have some other test station do the same measurement? They may be wrong.
I would also make sure that the car is on normal operating temperature, and not in the warming-up cycle. In my believe the computer runs rich during warm up and operates on an static map.


-----Oorspronkelijk bericht-----
Van: Dr. Bob []
Verzonden: dinsdag 15 januari 2002 15:31
Aan: 928
Onderwerp: [928] re: Lambda value

Graham wrote to the list:

> ... I've just took the car for a new inspection <...>
> However they wouldn't give me a new certificate due to a slightly high lambda reading.<...> The lambda sensor is new by the way so what do I need to do next?

And he later added:

> Maybe its a German thing. The CO value was ok but the "lambda value" varied between 1.1 and 1.2 and apparently I've got to get under 1.0. The last time I had it tested it varied between 0.97 and 1.15 and that was good enough.

Theo responded:

> The sensor measures the amount of un-burnt fuel in the exhaust, and adjusts the FI computer accordingly.
> (this is the closed loop) There are many reasons for the CO to be too high.

Max contributed:

> Isn't the term Lamda Sensor the proper name for what most of us call the O2 Sensor?

Some basics--
Lambda is the Greek letter that is typically used to identify the point of optimum combustion. As Max points out, the name is applied to the sensor (but not one that detects O2...) that's installed in the exhaust to measure the combustion efficiency.
The sensor measures carbon monoxide (CO), the amount of which in the exhaust is proportional to the amount of fuel burned. More CO means the mixture is richer, less means leaner. Output voltage higher, more CO. Theo points out that the sensor drives the mixture calculation in the FI computer when the computer is in closed-loop mode. It provides a somewhat continuous adjustment, with an allowance for lag between injection changes and the reading in the header pipe, and also the response time of the sensor itself.

Theo mentions that "There are many reasons for the CO to be too high."
There is only one direct reason, and that is that the mixture is too rich (assuming that you are measuring ACTUAL CO at the header pipe rather than downstream behind a catalyst.) While there may be various possible causes for a rich mixture, it all funnels down to that when you are researching CO levels.

I'm not any kind of expert on German inspection methods. A lambda value of 1.0 indicates that the sensor is working perfectly and the computer is reacting perfectly to the sensor signal and controlling the mixture to maintain that level. One guess is that they read the CO value with their own sensor, and having already determined the ideal CO value and built a
table, do a lookup to see how far off the actual CO value is from the target number, and generate a lambda number based on the error. Worse though might be the possibility that they are reading the sensor already installed in the car, and determining that the car's FI computer is not responding correctly to the signal that it is receiving from its own sensors.

Sensors that you buy come in several flavors. You can get an exact-match sensor at a dealer, or in the US from one of the Big Three vendors. There are also OEM Bosch sensors available that supposedly do the same job but don't have the cost multiplied by 9.28. There are also "universal" senders from various manufacturers, including Bosch, where perhaps the wire leads or connectors need to be spliced together to make them actually fit. On the way to other research, I've noticed that the "quality" of the sensor is related to how well it works in your particular model of car. An "exact fit" Bosch sensor doesn't work well in my Explorer, while the "exact fit" Motorcraft sensor works fine. Hmmmm. So the big question would have to
be-- Which sensor did you buy? After the experience with the Ford, which apparently requires a slower, less sensitive/responsive sensor to run right, I might be able to conclude that the 928, with the FI brain designed for the
faster/more responsive Bosch, might not pass with a less-than-perfect match.

The big telltale of incorrect or worn sensor operation is hunting at idle.
A quick DVM, or better yet an oscilloscope, can be used to examine the voltage emitted by the sensor in closed-loop operation. You can easily see the mixture 'hunting' by watching the actual voltage from the sensor. This method is pretty worthless if the sensor has failed due to old age or contamination, by the way. Ammonia and acetic acids are the common
automotive sensor contaminants/killers, but a lot of raw fuel (from a misfire) will do the job nicely too. Heat is also a popular killer, so excessive ignition timing and exhaust temp are factors to be considered when diagnosing sensor failures.

A side observation-- If I was touring at autobahn speeds and fuel consumption was not a concern, the 1.1 lambda reading under hi-speed load would be my choice compared with a lighter reading. A little more power, a little better cylinder cooling, etc would offset the additional consumption  and load on the catalyst.

Just curious also-- Was the gas sample for the CO measurement taken at the gas sample pipe ahead of the catalyst, or from the tailpipe after the catalyst?

Finally, as another response points out, because you are so close to compliance, I'd look for another test station before doing any mods to the car.

Let us know what happens!
dr bob

thanks for your comprehensive contribution. Very good and thorough al always.
Just wanted to comment...on your comment.
let me give you my top 5 of reasons why the CO is wrong:

1. FI computer defective
2. O2 sensor wrong, defective, disconnected
3. MAF sensor defective, wrongly indicating high flow
4. temp sensor problem indicating cold engine.
5. defective Ignition computer (timing?)

I'm sure there are others. Nevertheless, we agree.
"There are several causes for rich mixture"

'88 928s4 cherry red
The Netherlands


At 11:44 AM 1/15/02, wrote:
>Dr. Bob,
>Nice write-up as always.
>My understanding is the higher the lambda reading to leaner the
>mixture. Am I misinformed? My mechanic adjusted the mixture on our 1992
>MB 300E this weekend after seeing a very high reading. Also, I have yet
>to see a steady lambda reading. I see it go above and below 1.0 at
>idle. My wrench sets the mixture so the lean readings are less often then
>the rich ones.

Air/Fuel ratios by weight - condition - lambda

12.6:1 - max power - lambda of 0.86
14.7:1 - ideal combustion (chemically) - lambda of 1.00
15.4:1 - max economy - lambda of 1.05

Wally Plumley
928 Specialists



Wally pointed out an error in my description of the lamda oxygen sensor in yesterday's post in reply to Graham's problem. Others have also pointed out some inconsistencies.

The sensor detects total oxygen in the exhaust stream, relative to the oxygen available from the surrounding air. Doesn't matter if the molecule is tied up as CO, CO2, NOX, or is free O2. Realistically, oxygen tends to be tied as CO, with a very small percentage as NOX worst case, so it's used to effectively measure CO to determine the mixture.

The description of lambda is backwards, where a higher number (greater than 1) indicates a lean condition rather than the rich condition as I stated. However, the voltage does in fact rise when more fuel is in the mixture.

I promise not to lecture out of my butt on this subject again. :-)



t 07:24 08/12/2002 -0400, Jay wrote:
>Keep in mind that the WOT switch serves to only cut the O2 sensor signal
>out - not really that much impact.

Hummmm, If I might interject: It would be more correct to say the WOT switch causes the fuel control computer to change from a dynamic map to a pre-programmed static one.

Normal operation: O2 sensor is part of a continuous feedback loop constantly trying to hold the fuel mixture at the lowest emission point. (typ. 14.7:1)

WOT: The fuel computer consults a pre-programmed internal map instead.
This map has a richer target setting. (typ. 12.5-12.7:1) which is the max. HP mix. This map assumes a stock, unmodified engine
system, 92 octane gas, etc. (This is the map that most after market "chips" change.)

Why did Bosche build them this way?
Speculation: At WOT and near red-line RPM's, early computer feedback loops could not correct fast enough to prevent sudden lean burn conditions. In a worst case scenario, lean burn holes pistons.

# -Larry Ware