RomRaider

I forgot about this thread... don't remember if he moved the sensor or sent me a log, but I'll check my email...

Wow. What a great time for you to bump this thread. I recently did some mods that required me to move my O2 sensor to the downpipe. I now have a slightly oscillating idle that I completely attributed to a vacuum leak - which I have been unable to find (hooked up my pressure testers and tested everything that I could).

I have a new O2 sensor on the way. I'll report back if that affects it. It's a shot in the dark but, at the same time, I am overdue to replace this sensor. It was time for a new one anyhow. Mine is coated with MMT from when I use to run Torco a couple years ago.

Another thought is that maybe mine is due to unplugging the rear O2 and running the LC1 in it's place (with the narrowband output routed to the ECU). If the new front O2 doesn't help ... and if this thread doesn't get updated with more advice, then I will try reinstalling my rear O2 just to see if that helps.

I'm also going to do a quick test later this week whereby I cap off every single item except the FPR on the manifold and see if by chance that helps. But, this still makes no sense because I didn't have this issue before I swapped header, turbo, and O2 sensor to the downpipe. No modifications were made to the intake stream.

t

In order to damp AFR oscillations at idle you need to decrease integrating and probably proportional gain in CL control. This is needed due to additional delay for exhaust gases propagation.

The question is how this may be provided for your particular ECU.

Here is a hack that works (not saying this is a great thing to do).

In the open loop fueling table, change the left topmost cell (lowest load and lowest RPM) to a slightly richer value. I used 14.39 and blended the adjacent cells.

I don't understand why this works but it does work for smoothing out the idle. I don't understand how the open loop table can affect closed loop operation

t.

If the open loop enrichment puts your AFR below the open loop threshold, the car stays in open loop.

The open loop threshold is adjustable - I have mine set to 14.65. While I was doing MAF scaling I set the top-left corner of the fuel table to 14.60 to keep the car in open loop.

I can't find an email from that guy I was talking to. If anyone has the O2 sensor in the stock location, please log a minute of idle and a minute of cruise, after the sensor warms up and the ECU goes into closed loop.

Thanks for the input! Its the little things like this that correct my understanding of something .... that makes all the difference in the world.

t

What parameters would you like logged exactly?

Just these two:

AF Sensor #1
Closed loop fueling target

And check learning view before and after each log. Just in case the fuel trims change (I doubt they will).

I think I did this wrong. I logged correction #1 instead, sorry.
But looking at these logs leads me to believe I have a air leak.
This must be as a result of my recent clutch work, I had the intake pre-turbo apart while the car was down.

The 'cruise' log shows what I was looking for: the ECU varies the target AFR from about 14.1 to 14.85 continuously, about one cycle per second. On two or three occasions now I've heard from people who moved their sensor back, saw those oscillations, and decided that moving the sensor to the DP was a mistake. I was pretty sure it happened in the stock location too, thanks for providing proof.

The CL target parameter remains constant during idle, but I suspect that the same thing happens at idle as well. It's just controlled by a difference piece of code, so it isn't visible with the same CL target logging parameter.

"the ECU varies the target AFR from about 14.1 to 14.85 continuously"

This is needed for ecology - this is the way ThreeWayCatalist lives.
TWC periodically stores oxygen and burn CO&CH when mixture is lean and restores NOx when mixture is rich.

BUT. Moving AFR sensor far from exhaust you may get those oscillation at idle due to improper CL calibrations - lesser CL gain(s) are required badly due to AFR sensor signal delayed.

So you are saying we have not yet defined the code that allows us to change the CL operating parameters?
Or, we haven't tuned the parameters we already know of?

Sasha has an interesting point. Dschultz, if you could do some logging of idle (just sit for 30 seconds, after driving for a few minutes to get the engine warm), that would be great. I'm very curious if it fluctuates as much. My car behaves similarly at both idle and cruise, but I've moved my O2 sensor to the downpipe.

I got one of the parameter names wrong last time, try to log these two parameters:
* AF Sensor #1
* CL Fueling Target (2-byte)
If "CL Fueling Target" isn't supported for your car, then just log the first one.

Thanks!

I know exactly that it impossible to create any CL system without loopback gain(s) defined.
Fuel CL uses at least proportional and integrating factors (probably load\RPM dependent). They may exist as factors not defined currently or "hard coded".

Dunno about your ROM, but the definition for mine does not include parameters that would affect the CL fueling feedback loop, or the frequency or amplitude of the target AFR fluctuations. Most people using RomRaider are primarily interested in power, which happens in open loop, so none of that stuff matters to most of us.

But, you can disable it completely if you set the OL threshold to 14.65, and make the highest target AFR in your fuel table 14.60. Getting rid of the feedback loop makes MAF and injector scaling a little easier, but my car drives the same either way, so I went back to normal when I was done.