Air/Fuel Ratio Mystery

[Teuobk]

I noticed an unexpected variation in my AFR on my Vee while racing this past weekend, and after pondering it for a while, I was wondering if anybody can come up with a compelling explanation.

First, some background. I have a 1970 Zink C4 (at least as we call them here in Colorado) that I run in vintage races with RMVR. This past winter, I decided to be very non-vintage and add an AiM MXm datalogger with a bunch of sensors. Most relevant to the following problem, one of the sensors is a throttle position sensor, and another sensor is a lambda sensor, with the Bosch wideband O2 sensor mounted in a bung on the cylinder 3 exhaust pipe.

While reviewing the data from the races we had at High Plains Raceway in Colorado this past weekend, I noticed something odd: during every session on Saturday, the lambda value reported by the O2 sensor appeared to vary according to where I was on the track. In comparison, the lambda value on other days varied only as it normally does, which is to say that it runs a little lean at low RPMS and gets progressively richer at higher RPMs (indicating that I might need a bigger main jet and/or a bigger air-correction jet, but let's set that aside for now).

Here's a view of two laps, one from Sunday (yellow) showing representative normal behavior, and one from Saturday (blue) showing representative unexpected behavior. The top section shows lambda versus track distance at WOT (with a value of 1.0 for non-WOT periods). The section section down shows speed; the third engine RPM, and the fourth time delta. Also note that the turns are labeled across the top (1-15), with blue being lefts and red being rights:



To reiterate, the way the blue trace moves relative to the yellow trace is the weird part. Notice how the engine is running lean into turn 1 (i.e., blue trace higher than yellow), then rich on the straight between turns 3 and 4, then lean on the straight/sweeper between turns 6 and 8, then pretty close to normal for a few turns, then lean on the short straight between 10 and 11, then really rich between 11 and 13, then lean again on the straight coming out of 15 and into 1.

For reference, here's a map of High Plains Raceway, oriented with north to the top and with the turns numbered:



My best guess so far is that it's somehow related to the wind. Temperature, pressure, and humidity were all very similar between the two days. On Sunday (yellow trace), the wind was out of the north at 10 mph during the session shown in the datalogger data. On Saturday (blue trace), the wind was out of the southeast at 10 mph during the respective session. That would have provided a tailwind in the rich sectors and a headwind in the lean sectors on Saturday.

Anybody ever seen anything like this? Any ideas about why this would happen? If it is in fact related to the wind, how would that produce the observed effect? I find it hard to believe that there would be enough of a ram-air effect on my Zink to cause this.

Thanks!

Jeff

[nrider42]

Jeff, I love a good mystery. If your present theory is wind...can you test this with a fan? Run the car and blow some wind up her tailpipe and see if the reading changes? Also any idea what the noise levels are for this sensor? What type of accuracy and repeatability can you expect?

Also, how far up the #3 exhaust pipe? How far from the collector? I have been pondering installing one of these and was wondering where the best place to put it was. Maybe you have the answer!

Cheers,

Nate

[ProstWest]

How does the air get to the inlets in your car?

[Teuobk]

Jeff, I love a good mystery. If your present theory is wind...can you test this with a fan? Run the car and blow some wind up her tailpipe and see if the reading changes? Also any idea what the noise levels are for this sensor? What type of accuracy and repeatability can you expect?

Also, how far up the #3 exhaust pipe? How far from the collector? I have been pondering installing one of these and was wondering where the best place to put it was. Maybe you have the answer!

Great questions. Although I don't have a fan that would be powerful enough, I do have a leaf blower. I will give that a try and see what happens.

It's a Bosch LSU 4.9 wideband O2 sensor, so the noise levels should be very low and the repeatability should be quite good. Specified accuracy is +/- 0.01 lambda at lambda 0.80

I have mine mounted about half way between the collector and the head. My understanding is that the really important thing was to put it before the collector, not after, as the collector isn't sealed well enough to prevent external air from being drawn in.

Jeff

[Teuobk]

How does the air get to the inlets in your car?

On this car, the air filter is just sitting on top of the carb, and everything is loosely covered by the engine cowling. There aren't any funnels or ram-air-type devices intentionally present, though I can't help but wonder if the engine cowling might have been having an effect like that.

Jeff

[Hardingfv32]

1) You tune a FV engine on a dyno and baseline O2 and EGT at that time. O2 and EGT are for reference only at the track.

2) Unleaded gas? How much oil do you burn? O2 sensors are easy to upset.

3) Knowing how glitchy O2 sensors are on a FV, I would not not expect be able to answer your question with any confidence.

4) This is the most important note: The individual cylinder O2 readings are all different. Issues caused by the siamese ports and intake manifold. NUMBER 3 IS THE FARTHEST OFF THE AVERAGE! Yes, an O2 sensor in the collector can be inaccurate.

Brian