RomRaider

I have been running in forced open loop for a while now. I only run in open loop at 0.25 engine load column so basically at idle and when engine braking. I have been trying to nail down my idle fueling to steady my idle. I have found that just turning my lights off and on affects my AFR even when running in open loop. It also affects my throttle plate position. The only time turning my headlights on and off didn't affect the AFR was during the radiator fan cycle from 206* down to 194*. Does anyone else have this issue?

My rom is based on Freon's Hacked Speed Density Rom for the 04 STi. I attached the rom and the log below.

it looks like the throttle position is causing your problems. Either a sensor problem or a throttle cable problem (too tight). The idle on my car was out of wack till I flattened my maf scaling around the idle values. Just another thought but it doesnt seem like thats your problem.

-Phil

My throttle is DBW. The engine is idling with my foot off the accelerator pedal. Turning the headlights on and off is what is causing the throttle plate position and AFR to change.

There is an electrical load compensation but I can assume that it is one of the idle speed target tables. Noticed that one of your tables is requesting 850 RPM while the other two are 900 at temp. Also, the 'target throttle plate position' has values in the 800 and 1200rpm area when requested torque is '0'. Any of these will cause AFR and throttle changes.

You may be right about the idle speed target tables being electrical load compensation. I took all three stock idle speed target tables and increased them the same amount. I always wondered what each table was for.

The target throttle plate position table was intentionally set that way to help my idle stumbles. I just switched to e85 and 1200cc injectors so I am still trying to get my fueling right and have had trouble with my idle stability. It used to take an hour or so for my car to relearn the closed loop idle at varying loads after reflashing. When I set the minimum throttle opening at idle rpm to around 3%, it seems to learn the closed loop idle throttle control very quickly. I think I have read somewhere that Cobb uses this trick too on some of their maps.

I logged battery voltage because I wanted to see if the injector latency was causing the issue. The battery voltage remains almost constant and doesn't seem to follow the same trend of the light switch so maybe the electrical load compensation isn't based on battery voltage.?? The AFR is jumping two full points when the switch goes from open to closed, except when the radiator fan cycles from 207 to 194.

You know how data logging can be started and stopped via the defogger switch? I suspect that the whole reason the ECU was wired into the defogger switch in the first place is so that it can compensate for the extra electrical demand. I wouldn't be surprised if the ECU knows whether the headlights are on or off for the same reason.

I also set the leftmost column of the target throttle plate angle table to 3%, so the ECU doesn't have to re-learn idle after a reflash. Someone at legacygt.com suggested it, and I wish I could remember who so I could give them credit. Works great.

All electrical ties in to the charging system so if there is a change in demand, the ECU increases the alternator duty and rpm to compensate.
I have never thought of increasing the target throttle angle, just spent a lot of time dialing in the bottom of the MAF scale instead.

You know, with the latest version of RR Logger and logger.xml def you can see exactly which switches are supported by your ECU. There are a few which relate to electrical load. Even moving the steering wheel will increase the idle target.

The whole point of CL is to compensate for the varying demands on the engine and keep the AFR consistent. I think you'll have a hard time nailing down OL AFR at idle and very light load conditions, pick a happy medium. Learning A can easily swing + or - 5% or more under various conditions (loads and environmental). I kinda gave up on tuning the low end of my MAF scale for these reasons. Some days mine is - other days it's +.
Get your setup close and then let CL do it's job.

I understand the point of CL; however, I was trying to nail down my AFR to less than + or -20% swings that I have been seeing on the short term fuel compensations. My long term fuel trims are less than 6%. This all stems from my quest to stabilize my bouncy idle that I get quite often when coming to a stop when running CL fueling at idle. These 1200cc injectors, DBW throttle, and speed density are making it difficult for me. Not to mention I have cosworth cams, ported heads, and 1mm+ valves.

Thanks for the info on the electrical load though. I will adjust my idle target speed tables to the same values and see if it helps and go back to CL and see where I am.

You can't.

The short term compensations are a direct result of the engine operating conditions in the last fraction of a second. i.e.: if you take your foot off the gas while cruising the STFT will go -20% or whatever the max is. This is because it is reacting to the throttle close and the last AFR reading which would be too rich for a closed throttle and no air coming into the engine. It's saying, reduce fuel as there's no air and the mixture is really rich. Then in the next second the STFT will settle down to what's normal for the current conditions. That's why any load on the engine at idle results in widely varying STFT, until the Learning A catches up to what the best comp is for that range over time. I find Learning A will update fairly quickly, within seconds to minutes depending how far off the STFT on average and Learning A are from each other.
The LTFT are the ones you are most interested in.