Hunting Idle and AFR

Hey Guys, I have a 05 USDM STi that I just cant figure out. Just finished a complete rebuild and am currently breaking in the engine. Since everything was new I decided to start a fresh tune beginning with a Cobb stage 2 93 OTS map working on Access Tuner Race. I installed a new front A/F sensor and Wideband sensor for my AEM UEGO to guarantee accurate data. MAF has been swapped with known good unit. I have removed the rear o2 and zeroed all of the closed loop adders associated with it (now wideband location). The car has ELH, Divorced catless DP, New billet wheel VF39, Walbro 255, AEM cold air, Spearco TMIC, 8mm phenolic spacers, evap delete, IAG AOS, 3 port EBCS, NGK one step colder plugs, and Lightweight crank pulley. The idle hunts up and down along with the AFR swinging a full point at times. A quick tap on the throttle at idle results in a drop in rpm, afr goes rich then lean and engine speed comes back up. The car drives great, nice and smooth. Afr does not stay constant (swings about +/-.5) while light load cruising. For short periods of time the AFR will stabilize. I also noticed the front A/F sensor and Wideband do not register the same readings. Before doing any further tuning, I need to figure this out. Im almost leaning towards a failure in the DBW throttle body ( I have a backup and may swap to rule out). I have a quick log at idle but will be happy to provide any other requested data. Any help or direction is greatly appreciated!

Your AFR swings between 13.4 - 16.2. Idle ignition and throttle valve actuator seem to be doing their work just fine.

Can you re-enable closed loop in idle and run another log file here?

Absolutely, same parameters as the first log?

Yes, same parameters plus make sure aircon is off for the log. Eventually battery voltage as well. I assume the steering wheel is not being turned too (hydraulic load on the pump. I know it is obvious so please forgive). Speed at zero, vehicle stationary.

You got it sir! I appreciate the help

Here is the log. I have noticed there is random roughness registering but am not sure if its part of the cause of my issue or an effect. Any thought as to why front A/F and UEGO do not match and almost seem to move opposite directions?

mmhh.. I see in rows 604 - 617 roughness index 1 registered at cyl. 1,3,4. In some places cyl roughness goes up to value 2. I wonder whether this is an indication that could guide you into the right direction ? It also seems that the idle ignition advance has very little effect on the actual engine rpm.. ? The AF seems much more consistent now but apart from all that.. I apologize of not being of to much help here. Often it is just necessary to hear, feel, smell.. the engine running. Pinpointing only trough data is often not possible. Maybe others have some input here for you?

I also question the same thing. I appreciate you taking the time to help. Whether the roughness is the root of the issue or an effect of my root issue. It seems that the idle is stable and Afr is rock solid when in open loop. This makes me think that I have a closed loop adder somewhere that I may have overlooked. Although I've been through every table there, multiple times. Someone please educate me in what I'm doing wrong

Anyone?

What ECU are you using? Could you post a log of it in open loop and possibly a copy of the map?

Your OE Lambda and UEGO sensors are reading the same values, just approx a second apart, the ECU is compensating using the OE sensor

It's the factory ECU. The map is basically an OTS Cobb Stage 2 slightly manipulated for the mods. It won't allow me to attach the map to this post. Is there an email you'd like me to send it to? The Afr and idle stabilize for a short period and as if some one flipped a switch, idle dips, then afr starts sweeping roughly a point. Then it will stabilize again. For the open loop log, would you like that in a cruise or open loop at idle? Thanks for taking the time to help

I don't have software that'd allow me to see COBB maps, it was more so if someone else who did then they could view it.

If the AFR's are only oscillating at idle then open loop at idle would be enough, also if you could record the coolant temperature and idle control valve duty in closed loop as another log?

I would be happy to share the map for review by anyone willing/capable. I will run another two logs as described other than this car is DBW so there is no idle air control valve

What happens when you set the software back to 100% factory stock ? Do the symptoms disappear then ?

I haven't tried reflashing the stock map due to the mods on the car. I think I forgot to mention that after double checking everything, I settled on the theory of injector imbalance. I changed the injectors to a new set of dw1000 side feeds (I don't prefer anything DW but I had them new, on the shelf. I installed them, rescaled and figured battery offset (latentcy) and that's when you see the afr in the logs are way more stable. The AFR is still drifting for no apparent reason, just not as wildly. The engine will idle a steady afr for a few seconds then suddenly go into the sweep. After a short time sweeping, it will suddenly stop and idle and afr are stable again. I have a set of ID 1000 top feeds on the way that should be here Tues. Maybe the DW1000 also have an imbalance although it doesnt make a lot of sense how the swinging afr comes and goes so suddenly. A thought that's way out in left field but what it the phenolic spacers are having a negative effect on engine grounding?

It is my opinion that setting it back to stock temporarily is inevitable. Although I doubht the software causes this random error, this approach will enable you to judge whether or not that is the root cause.

I apologize for not providing the requested log yet, but I have found that with the MAF disconnected the afr and idle are solid and the random roughness stops registering all together leading me to believe MAF cal is off. I used the Cobb MAF calculator based on diameter of both OEM and AEM intake and generated a global correction factor. I applied this to the whole table to start. This resulted in the behavior I'm trying to resolve. I have since re started the process but with the factory values. Seems to be helping but not the entire solution. I may install a different intake and see if that could be part of the issue.