Open

Open

Introduction

This document outlines some of the unique engine control unit (ECU) logic and tuning strategies relevant to later model Subaru 2.4L DIT as compared to earlier 2.4L DIT. It is not the intention of this document to outline every difference but rather the most notable differences (known to date) that can impact your ability to tune effectively.

Associated Information

Please refer to the other DIT tuning supplement guides for additional information not described here:

Subaru Accesstuner 2.4L DIT Tuning SupplementPreview (This has the core logic differences for the 2.4L DIT as compared to 2.0L DIT)
Subaru Accesstuner 2.0L DIT Tuning SupplementPreview (General DIT tuning info - some of this still applies to the 2.4L)

Applicable Vehicles

The following Subaru models are applicable to this document:

• 2022+ Subaru WRX

• 2023+ Subaru Legacy / Outback

• 2023+ Subaru Ascent

Table Descriptions

Besides this document, an additional resource for ECU logic details are the table descriptions shown in the Accesstuner software. These will show up in the lower left corner of the software for the current table that is selected. If not visible, make sure the "Table Description" check box is selected in the View menu. Additionally, the Help menu in the software contains links to documents that list all of these table descriptions as well as all of the data monitor descriptions.

ECU Changes

When compared to early 2.4L DIT vehicles, the later 2.4L models have some differences in operation:

Limits

• Boost Reading (MAP Sensor Values)

  • Currently limited to a max of about 61.5 psig at sea level, not including the hardware limit for the factory MAP sensor (which likely is limited to about 24 psig at sea level). We have included the MAP calibration tables to account for the installation of an aftermarket MAP sensor. Keep in mind if you choose to replace it, the factory MAP sensor includes an IAT sensor to read air temperature in the manifold.

• Boost Targets

  • Boost targets can go to a maximum of ~61.5 psig at sea level when the Boost Targets Max. Limit table is raised from its 24 psi limit.

  • The Boost Targets Compensation tables apply compensation both to the Boost Target and Boost Limits, unlike Boost Targets only on early 2.4L DIT.

• Load Limit

  • Limited to just under 4.0 g/rev. Gen3 ECU types have a Load Extend feature that increases this to about 7.7 g/rev on key tables.

• Injector Pulse Width (IPW)

  • At very high loads, a hard-coded IPW limit can be hit in the factory ECU. The Gen3 ECU type removes this limit.

• Requested Torque

  • Internal values in the ECU cap the requested torque value at 350 N-m.

Throttle Mapping

The 2.4L WRX has some unique changes related to requested torque:

• Requested Torque

  • (WRX only) For the 6MT, tables are split by the estimated (calculated) gear position (1st through 6th). This ECU utilizes a new monitor (Gear Position ESTIMATED Req Torque) which tracks the unique gear position monitor used for table switching.

  • (WRX only) For the CVT, tables are split by the “gear” position (reported by TCM) and TC Lockup status for each SI-DRIVE category. This ECU utilizes a new monitor (Gear Position ESTIMATED Req Torque) which tracks the unique gear position monitor used for table switching as well as the CVT Lockup Status monitor to track the lockup states.

  • (ALL) Due to internal ECU limits, the maximum Requested Torque value is capped at 350 N-m.

• Requested Torque (Gear Transition)…

  • (WRX only) This table is used when the car is transitioning between gears.
    
    
    

Tuning Strategies

Light Load Knock

During testing, we observed that certain 2022+ WRX vehicles can demonstrate very significant knock corrections under very “light duty” conditions. This typically occurs when:

• The driver reduces throttle input after an acceleration event, but does not lift off of the accelerator entirely, such as when reaching desired speed after leaving a stop light or entering the freeway.

• Engine Speed is between 2000 and 3000 RPM.

• Calculated Load is below 0.5 g/rev.

When this behavior begins:

• Feedback Knock Correction (FBKC) can rapidly increment negative into large absolute values

• It will often continue until the driver inputs change, such as increasing or decreasing accelerator position, or until the FBKC value reaches the maximum negative value allowed by the calibration.

• If you keep an eye on the Knock Sensor Noise Level Cylinder 3 monitor, you will typically see large spikes in the value.

• If you compare these spikes against the Knock Sensor Level Threshold Cylinder 3 monitor, you’ll see that any time the noise level exceeds the threshold, there will be an increment in the FBKC.
  
  

Our testing indicated that:

• The behavior is not consistent across cars but will be easily observed on the stock (or Stage0) calibration for those cars that are most heavily impacted.

• This issue is not specifically worsened or alleviated by changes in octane; it was observed on 91 octane fuel up to 96 octane fuel.

• It is unlikely that the noise is being generated by legitimate knock events and that it is mechanical/external noise (or another anomaly).

• Many FA24 vehicles, including Ascent, also exhibit the noise spikes on Cylinder 3, but often to a reduced maximum, effectively reducing or entirely negating the severity of the issue.

Mitigating the Light Load Knock

Our Stage1 maps have a series of small changes included that help reduce the frequency of these light load knock events, though larger changes may be needed depending on the vehicle and the specific behaviors observed.

Below you can see a visual representation of the issue occurring on an in-house test vehicle during early R&D efforts. This is on the stock calibration and a vehicle running 92 octane located in Portland, OR. The 3rd graph shows the noise spikes (red) exceeding the threshold (white), the 2nd graph shows FBKC incrementing downwards when those spikes occur, and the 1st graph shows the benign conditions under which it is occurring (Calculated Load below 0.5 g/rev, APP below 15%, etc.). While not shown in this graph, Cylinders 1, 2 and 4 do not demonstrate the noise spikes at all during this event.

Open