I've got an s4 rtek 2 with the latest update and it's controlling a mildly ported motor. I'm running cleaned stock primaries and 750 secondaries.

When it's running in closed loop for fuel correction, it runs way too lean and lacks power. I've got a wide band and it shows the AFR oscillating between 16.0 and 14.5, never lower. I'm also running the narrow band sensor input from the simulator output of my wide band. I've verified that it runs correctly and switches at the right afr. At low speeds in first or second gear, I get awful jerking from this.

Last night I installed a switch to cut the input for the narrow band to the ecu. The car runs so smooth now. It's like a whole new machine. I did have the same jerking before I installed the wide band.

What afr and correction percentage is the ecu trying to achieve in the closed loop and is it possible for me to fix/change it?

 

bad narrowband? unshielded wires?

 

When i 1st installed my wide band i noticed similar way to lean for my liking numbers during cruise also but never the jerking. Even when id set the SAFC way rich it would still do the exact same thing. I always assumed it was because the cars were such gas hogs that during cruising they ran them super lean like that on purpose since my car was bone stock but RB turboback and perfect running shape when i bought it.

The air fuel numbers with the narrowband 02 disconnected from ECU were about perfect 14.7ish anyway so i never really thought or cared much about how to run the narrow band without it so lean and this was before i had the 2.1 to even more fine tune it. It seemed like there was lag when switching between the cruise/reading the O2 and main fuel map when id 1st get on it where it would dip lean as i was 1st boosting that freaked me out more than a tad lean when cruising but again i had no noticeable runability problems when it was hooked up.

i would think it would be possible to run some tiny resistor on the narrowband line to trick the cruise to be a tad richer but never saw much point in it.

 

Sounds to me like your fuel settings are set too lean resulting in the afr/stumble your seeing during closed loop. Settings are too lean for those vacuum readings during crusing/steady driving. Like add a full five percent across the board and see how it runs.

IF you lived at a higher altitude I'd suggest moving your ATP slider to zero feet and see how things now look. But you live in Fla. so that's not a viable solution.

 

Perhaps the narrowband simulator from the wideband output doesn't read like the stock narrowband sensor. I drove 800 miles to KY last week and I tweaked the maps as I went.

They were probably too high, but I was running about 14.5-14.7 across the -12 to -4 vacuum map. I was getting about 21mpg running 80mph. The car had plenty of power going up the hills unlike last trip when I had to dip into boost to have any sort of hill climbing power.

The maps were stock before I started adjusting them this trip so I'm not sure how it could be too lean that way. I was under the assumption that is had an arbitrary lean/rich count per second ratio for closed loop. Does it just calculate based on the maps or pre-programmed ratio?

I've also noticed the lag when switching from too and from closed loop.

 

It's just looking at the 02 sensors voltage signal and comparing it with a voltage that represents stoic. The ECU'll either add fuel to bring the 02 sensors voltage up to the voltage that represents stoic or it'll reduce fuel to bring the 02 sensors voltage down to match the voltage that represents stoic.

The closer the fuel reading is to stoic without the 02 sensor in the loop, the easier it's going to be for the ECU keep stoic when the 02 is in the loop. Sorta like a BAC. The closer the engine idle is to 750rpm without the BAC in the circuit, the easier its for the BAC to do its job when it's in the circuit and a load is put on the engine.

 

to me it looks like you are jumping between cells of the VE table due to the resolution and the oscillation of the MAP signal.

Try this as a short term test: double your resolution. cut your rpm into 200 rpm increments (max: something like 4000rpm) and evenly space all your pressure axis in 0.5psi increments. Set a rev limiter at 4200rpm and boost limiter at 1psi or whatever it ends up being. At that point you can easily interpolate the new scaling (find the mid point between each cell).

With that enhanced resolution, try to fine tune the VE in the area of the oscillation. If you can improve the oscillation through VE tuning on a high resolution table, you know your VE is the root cause and you just need to come up with a compromise VE table scaling that meets your needs.