SUBARU SPEED DENSITY 2.

0L GUIDE
Version 1.0

1/08/2014
Contents
Overview ..............................................................................................................................................................................................5
Uses......................................................................................................................................................................................................5
Supported Vehicles List ........................................................................................................................................................................5
Glossary of Acronyms ..........................................................................................................................................................................6
Features ...............................................................................................................................................................................................7
Hardware Requirements ......................................................................................................................................................................8
Warnings ..............................................................................................................................................................................................9
OFF-ROADUSEONLY .........................................................................................................................................................................9
READALLDOCUMENTATIONBEFORETUNING ...................................................................................................................................9
SD IS NOT FOR INEXPERIENCED SUBARU TUNERS.........................................................................................................................10
MANIFOLD PRESURE SENSOR CHECK ENGINE LIGHT ERATIC LOAD CALCULATION ......................................................................10
SD Installation Steps...........................................................................................................................................................................10
What is Speed Density?......................................................................................................................................................................11
What is Volumetric Efficiency? ..........................................................................................................................................................11
Volumetric Efficiency Table................................................................................................................................................................12
SD Modes ...........................................................................................................................................................................................12
Tuning SD – Mechanical Configuration ..............................................................................................................................................13
Tuning SD – Getting Started ...............................................................................................................................................................13
Tuning SD – Initial Map Configuration ...............................................................................................................................................13
Manifold Pressure Sensor Diagnostic Trouble Codes ................................................................................................................15
Mass Airflow Sensor Diagnostic Trouble Codes .........................................................................................................................15
Load and MAF Compensation ....................................................................................................................................................16
Conservative Fuel and Timing Maps ..........................................................................................................................................16
Fuel Injector Scale and Latency..................................................................................................................................................17
Engine Displacement .................................................................................................................................................................17
VE Table Axis Scaling ..................................................................................................................................................................17
Reflash Changes .........................................................................................................................................................................17
Tuning SD – Starting Values for the VE Table .....................................................................................................................................17
 Car with Working and Accurate MAF Sensor .............................................................................................................................18
Car without Viable MAF Sensor Configuration ..........................................................................................................................19
Tuning SD – The Tables ......................................................................................................................................................................20
“SD Volumetric Efficiency” Table ...................................................................................................................................................20
“SD Load Compensation (Intake Temp)” Table .............................................................................................................................21
“SD Load Compensation (Coolant Temp)” Table ...........................................................................................................................23
“SD Load Compensation (Barometric)” Table................................................................................................................................24
Tuning SD – Miscellaneous Tables .................................................................................................................................................24
Load Filtering in MAF and SD Mode...........................................................................................................................................24
Map Determination Averaging Window ....................................................................................................................................25
SD Feature Activation ................................................................................................................................................................25
Tuning SD – Post-Tune Recommendations ........................................................................................................................................26
Tuning SD – Additional Topics ............................................................................................................................................................26
Recommended Calibrations for Aftermarket IAT Sensors .........................................................................................................26
Recommended Calibrations for Aftermarket MAP Sensors ......................................................................................................27
Forcing Open Loop Fueling ........................................................................................................................................................27
Long-Term Fuel Trims ................................................................................................................................................................28
Tuning MAF Mode..............................................................................................................................................................................29
Tuning Hybrid Mode ..........................................................................................................................................................................29
Overview ....................................................................................................................................................................................29
Hybrid Mode - Uses ...................................................................................................................................................................30
Hybrid Lower/Upper Mode........................................................................................................................................................30
Hybrid Threshold Switching Behavior ........................................................................................................................................30
Hybrid Thresholds Overview ......................................................................................................................................................31
Hybrid Threshold Individual Deactivation..................................................................................................................................31
Hybrid Transition Blending ........................................................................................................................................................31
Tuning Hybrid Mode – Post-Tune Recommendations .......................................................................................................................32
Tuning Miscellaneous New Features .................................................................................................................................................33
Gear Position Determination Expanded to Cover 6-Speed Transmissions and New Gear Speed Monitor...............................33
Per Gear Boost Targets and Wastegate Compensation.............................................................................................................34
Injector Latency Adder ...............................................................................................................................................................34
Neutral Position Switch Invert ...................................................................................................................................................34
 Boost Targets Limit Raised .........................................................................................................................................................35
Airflow and Load Limits Raised ..................................................................................................................................................35
More Flexible MAP Calibration Offset .......................................................................................................................................35
Tunable IAT Calibration Axis and Latency Axis ...........................................................................................................................35
How to Monitor the Tune ..................................................................................................................................................................36
SD Real-Time Tuning ..........................................................................................................................................................................39
Row/Column Table Data Removed from Real-Time ..................................................................................................................39
Tables Added to Real-Time ........................................................................................................................................................39
Real-Time Tunable SD Tables .....................................................................................................................................................39
SD Load Math .....................................................................................................................................................................................40
Ideal Gas Law – Introduction .....................................................................................................................................................40
Ideal Gas Law – Real-World Inputs ............................................................................................................................................40
Ideal Gas Law – Volumetric Efficiency .......................................................................................................................................41
Ideal Gas Law – Mass Airflow.....................................................................................................................................................41
Ideal Gas Law – Engine Load ......................................................................................................................................................42
Ideal Gas Law – SD Reference Load ...........................................................................................................................................42
SD Final Load ..............................................................................................................................................................................43
Estimated VE Calculation ...........................................................................................................................................................44
Appendix – Aftermarket IAT Sensor Install ........................................................................................................................................45
 INTRODUCTION
Note: This guide covers COBB Speed Density for the 02-05 USDM 2.0L WRX. For information pertaining to the
2.5L USDM Subarus, please see our 2.5L Speed Density guide.

Overview
The COBB Speed Density feature is a powerful yet easy-to-use solution that integrates Speed Density tuning into
the Subaru engine control unit (ECU) and can be used to entirely replace or work in conjunction with the existing
factory mass airflow (MAF) sensor. It is highly customizable and features such as real-time tuning aid in a speedy
and efficient tuning process.

Uses
COBB Speed Density (SD) has a number of potential uses that can improve the tuning capability for particular
set-ups:

• SD can eliminate the noisy airflow calculation sometimes seen when using a MAF sensor-based configu-
ration with heavily modified cars.
• SD allows for the MAF sensor to be removed, eliminating a potential restriction in the intake tract and al-
lowing for more freedom in intake piping design.
• In a special hybrid mode, SD can be used on the high end to overcome a maxed out MAF sensor, while
still retaining MAF sensor operation on the low end. Or SD can be used on the low end to improve
idle/cruise characteristics when a big MAF is used, while still retaining MAF sensor operation on the high
end.

Supported Vehicles List

This guide covers the following vehicles designed for and sold in North America:

• 2002-2005 Subaru Impreza WRX (Manual Transmission

Glossary of Acronyms
• ECT = Engine Coolant Temperature

• ECU = Engine Control Unit

• IAT = Intake Air Temperature

• IC = Intercooler

• MAF = Mass Airflow

• MAP = Manifold Absolute Pressure

• RPM = Revolutions Per Minute (referring to engine speed)

• SD = Speed Density

• VE = Volumetric Efficiency

• WBO2 Sensor = Wideband Oxygen Sensor

Features
The following is a list of the key features of COBB Speed Density (SD):

• SD works by calculating a new SD-based engine load to replace the factory MAF sensor-based engine
load when SD is active. This keeps much of the existing factory logic in place, allowing for the safe and re-
liable operation of the factory ECU while reducing the learning curve associated with tuning SD.

• Tuning SD is achieved through manipulating a real-time tunable volumetric efficiency (VE) table. Because
the VE table units represent actual VE, properly calibrated SD tunes can be used to compare VE across
different cars with different mods, even if the SD load is drastically different.

• An intake air temperature (IAT) compensation table allows the tuner to tweak the SD charge temperature
correction. This table can be tweaked according to manifold pressure, allowing for changes that may be
needed based on IAT sensor placement. SD load compensations for engine coolant temp (ECT) and baro-
metric pressure are also available.

• Tuners can select from three different real-time tunable modes of operation. MAF mode mimics the fac-
tory logic in which load is determined by the MAF sensor, “MAF Calibration” table, and RPM. This mode
allows you to get a starting VE table up and running before actually running SD or allows you to tune the
MAF sensor calibration as you would with the factory ECU. SD mode is full-time SD operation where the
ECU uses the SD-based load calculation. Hybrid mode allows for switching between MAF and SD mode
(and vice versa) based on thresholds of throttle, RPM, MAP and MAF voltage. During the transition, the
MAF sensor and SD-based load calculations will be blended over a short period of time to allow for a
smooth transition between modes. The speed of this transition and how the switching takes place can also
be configured.

• The SD and MAF sensor-based calculations are made regardless of mode. This means that the MAF sen-
sor-based load can be compared to SD load regardless of which value is currently being used.

• For cars with a properly calibrated and installed MAF sensor, an available estimated VE monitor can be
used to more quickly get the initial tune for the VE table up and running. This can even be done in MAF
mode, where the VE table (and other SD elements) can still be tuned before actually running SD.
 • Because the response characteristics of the MAF sensor (used in MAF mode) and the MAP sensor (used
in SD mode) are different, a tunable load smoothing factor is available for each mode. This determines
how load is filtered by the ECU and allows for better idle and tip-in/tip-out behavior in SD mode, while re-
taining the correct filtering in MAF mode.

• Load and airflow limits inherent to the factory 2.0L ECU have been significantly raised. No longer is there
a need to trick the ECU by manipulating airflow/load references and the fuel injector scale in order to over-
come the limits in the tune. Simply tune with actual real-world values up to 10.0 g/rev of load (SD or MAF
mode) and up to 655 g/s of airflow (MAF mode). Additionally, SD load is directly calculated and is subject
only to the 10.0 g/rev load limit (regardless of airflow).

• Special new features add new functionality or overcome limitations of the factory 2.0L ECU. These include
new real-time tunable per gear boost and wastegate compensation tables, a new real-time tunable low in-
jector pulse width compensation table, MAP sensor calibration with positive or negative offset, tunable IAT
sensor calibration and fuel injector latency axes, and maximum boost targets extended to 49 psig.

• “Tip-in Enrichment” and “Fuel Injector Latency” tables have been added to real-time for the SD 2.0L ECU
allowing for even faster and more efficient tuning with aftermarket injectors.

Hardware Requirements
The following minimum hardware requirements must be met in order to use the SD feature:

• Manual Transmission Only – COBB SD for the 02-05 WRX supports manual transmission cars only. Al-
though the AccessPORT will still allow you to install COBB SD on a 2.0L WRX with an automatic transmis-
sion, it has not been tested on this combination and the vehicle may not run or may have severe driveabil-
ity issues.

• MAP Sensor - The manifold absolute pressure (MAP) sensor installed in the car must be accurate, reli-
able, and capable of reading boost greater than the car can ever achieve. A typical car that would necessi-
tate an SD tune would likely max out the factory MAP sensor and require an aftermarket MAP sensor to
be installed. Any aftermarket MAP sensor must be properly scaled in the map (via “MAP Sensor Calibra-
tion (Offset)” and “MAP Sensor Calibration (Multiplier)” tables) and its accuracy should be verified by an
external boost gauge before tuning SD. Warning: A MAP sensor related diagnostic trouble code (P0068,
P0107, P0108) can cause a failsafe load calculation to come into play, potentially causing issues when SD
mode is active. It is not enough to merely uncheck the DTC toggles in the “DTCs (Base)” menu. Rather,
you must also uncheck the related fault detection toggles in the “Toggles (Base)” Menu (see “Tuning SD –
Initial Map Configuration” section for more details).

• IAT Sensor - Vehicle must have a working, accurate and properly calibrated intake air temperature (IAT)
sensor. The IAT sensor is crucial to the SD load calculation. The SD load calculation relies on a theoreti-
cal input of the cylinder charge temperature. The closer the IAT reading is to the actual cylinder charge
temperature, the more accurate (and consistent) the SD load calculation will be and the easier SD will be
to tune. The factory IAT sensor is in the MAF sensor assembly in a draw-through configuration (i.e. pre-
 turbo). While it is possible to tune SD with an IAT sensor in this location, it is more difficult and will require
more tweaking of the SD IAT compensation table (because this placement does not take into considera-
tion turbo and intercooler efficiency). We recommend that the IAT sensor be placed post-intercooler, if
possible. This could be using the factory MAF/IAT sensor assembly moved to a blow-through configuration
(i.e. post-front mount intercooler) or an aftermarket IAT sensor placed in the intake piping in the same lo-
cation (see appendix for details on how to install an aftermarket IAT sensor). Keep in mind that an after-
market IAT sensor will require a different calibration that the factory IAT sensor.

• Other Sensors – Any sensor or component necessary for the operation of the factory ECU must be work-
ing, installed, and properly calibrated. The only exception is the MAF sensor assembly if the tune is set to
run only in full-time SD mode (see “MAF Sensor Removal” below).

• MAF Sensor – The MAF sensor assembly must be installed and properly calibrated if the MAF mode, hy-
brid mode, or estimated VE monitor is to be used. It must also be installed if the factory IAT sensor (which
is part of the MAF sensor assembly) is to be used for SD (rather than an aftermarket IAT sensor).

• MAF Sensor Removal – If the MAF sensor is removed, disconnected, or otherwise non-functional, you
will only be able to run the car in full-time SD mode and the car will not start or run when MAF mode is ac-
tive. You will also need to uncheck the special P0101, P0102, and P0103 MAF sensor fault detection tog-
gles. Warning: It is not enough to merely uncheck the DTC toggles in the “DTCs (Base)” menu. Rather,
you must also uncheck the related fault detection toggles in the “Toggles (Base)” Menu (see “Tuning SD –
Initial Map Configuration” section for more details). While the ECU’s primitive failsafe load calculation has
been disabled in the COBB SD ECU, disabling the MAF sensor DTC fault detection toggle is still neces-
sary because other fail-safe logic can come into play. Also, removal of the factory MAF sensor assembly
also removes the factory IAT sensor requiring you to install an aftermarket IAT sensor.

• Wideband Oxygen Sensor – A properly functioning and installed wideband o2 (wbo2) sensor is neces-
sary to tune SD. It is also highly recommended that a permanently installed wbo2 sensor with an interior
gauge is used so that the driver can monitor fueling after the tune is complete.

• Mechanical Issues - Any mechanical problem with the car needs to be addressed before attempting to
tune SD.

Warnings

OFF-ROADUSEONLY
Some or all of the features and modifications discussed in this guide may not be legal to use outside of off-road
racing applications. Always consult local, state and federal laws to determine what is legal for your particular situ-
ation.

READ ALLDOCUMENTATIONBEFORETUNING
COBB SD for Subarus has been designed with the purpose of coming up with the best implementation for the
unique attributes of the Subaru ECU, while maintaining the same general characteristics found in current non-SD
tuning. COBB SD is not like other SD systems that you may be familiar with, including even COBB imple-
mentations for other platforms. As such, it is critical that you read through this guide and understand
how COBB SD for 2.0L Subarus works before attempting to tune. If you have any questions, we are always
willing to help.

SD IS NOT FOR INEXPERIENCED SUBARU TUNERS

There are many unique qualities to Subaru ECU logic that can make it challenging for someone new to the plat-
form. If you are new to Subaru tuning, it is recommended that you first become proficient at tuning MAF sen-
sor-only set-ups before tackling SD tunes. MAF sensor-only tuning can be much more forgiving to mistakes than
SD tuning.

MANIFOLD PRESURE SENSOR CHECK ENGINE LIGHT ERATIC LOAD CALCULATION

Any diagnostic trouble code (DTC) related to the manifold pressure sensor will cause the Subaru ECU to estimate
manifold absolute pressure (MAP) based on the current calculated load (as a failsafe). This can result in an erratic
load calculation in SD mode because actual MAP (a key input for SD) would no longer be determined correctly. If
this occurs when the vehicle is accelerating, a lean condition and incorrect timing can result. It will also likely
cause the engine to eventually stall. If there is the possibility that any of the MAP sensor related DTCs (P0068,
P0107, or P0108) could be triggered, it is critical that special toggles for these DTCs are unchecked in the tune.
Please note that it is not enough to merely uncheck the DTC toggles in the “DTCs (Base)” menu. Rather, you
must also uncheck the related fault detection toggles in the “Toggles (Base)” menu. The installation of an after-
market MAP sensor (required for SD if the factory MAP sensor is not sufficient) will make it more likely for these
DTCs to be triggered, even though there may be nothing wrong with the sensor itself. Please see “Tuning SD –
Initial Map Configuration” section for more details.

ENGINEHARDWARECHANGESMAY REQUIREARE-TUNEFORSD
It is important to understand that after the SD tune is complete for a given car, any further changes to engine
hardware that impacts airflow efficiency in or out of the engine can potentially require tweaking or re-tuning of the
VE table to avoid fueling/timing issues (due to incorrectly calculated SD load). Additionally, mechanical issues,
such as intake/exhaust leaks, and issues related to the aging of the motor, such as combustion deposits and loss
of compression, can also impact actual VE. It is highly recommended that a permanent wideband o2 sensor and
gauge is installed in the vehicle and that the driver understands how to read the gauge and determine what is nor-
mal for their tune.

POTENTIALRISKSFORSDWITHAHEAT SOAKEDIAT SENSOR

Any SD calculation, including COBB SD, requires an input for cylinder charge temperature, which is critical to the
determination of accurate load via SD. The estimation of cylinder change temperature is accomplished for COBB
SD via the IAT sensor input. Generally, when the IAT sensor is in the recommended location (post-IC), the vehicle
is moving and the driver is on the throttle, the IAT input can be a fairly reliable representation of actual cylinder
charge temp. However, when the vehicle is sitting still (or at low speeds) and the driver is off the throttle (or low
throttle), or the vehicle has been sitting with the engine off and a hot engine bay for a period of time, there is the
potential for the IAT sensor to become heat soaked. That is, the sensor now reads higher than the actual intake
air temp. When SD is active, this would cause the calculated SD load to be lower than it should be, causing the
car to run lean (and with generally more timing advance). This effect may subside after the vehicle gets moving
and throttle (as well as MAP) increases, but it will generally not be an instantaneous improvement. As such, it is
critical that the owner/driver of the car understands the specific scenarios in which a heat soaked IAT sensor can
potentially occur and to avoid putting the car under high load when these scenarios are present (and for a period
shortly after). This is another reason why a wideband o2 sensor and gauge should be installed in the car and that
the driver instructed on how to determine when fueling is incorrect.
 INSTALLATION
SD Installation Steps
Before tuning with SD, you’ll first need to update your AccessTUNER software and AccessPORT firmware to ver-
sions compatible with the SD feature as follows. Always make sure to periodically check for future updates.

1. If your current AccessTUNER version has the auto updater feature, you can update the software to the SD
version via the internet (“Help” -> “Check for Updates”). If your version does not have auto-updater, you
will need to download or submit a request for a new version:

For AccessTUNER Pro, you can find the latest version here:
http://www.accessecu.com/support/TunerPro_Suba_Setup.exe

For AccessTUNER Race, you can submit a request for the latest version here:
http://www.cobbtuning.com/accesstuner-race-request-s/70716.htm

2. Update your AccessPORT firmware to the latest version via the AccessPORT Manager software. The fol-
lowing video shows you how to update your firmware:

http://www.cobbtuning.com/AccessPORT-Support-s/40200.htm#/v/accessport/support-tutorials-
firmware

Now that your software and firmware is updated, you will need to reflash an SD-capable map to the car you will
be tuning:

1. Run the AccessTUNER software and when prompted with the ECU selection dialog, make sure the
“Speed Density” checkbox is checked. For AccessTUNER Pro, you will also need to select the ECU of the
car you wish to tune.

2. Make any initial changes to this map that you wish to start from for SD (see “Tuning SD – Initial Map Con-
figuration” later in this document).

Copyright © 2014 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 10
 3. Save your map. This map will now have the SD feature.

4. Transfer the map to the AccessPORT with the AccessPORT Manager software.

5. Reflash the new SD map to the car via the AccessPORT.

6. The car is now ready to be tuned via SD.

SD BASICS

What is Speed Density?

The Subaru ECU, as well as any engine management solution, needs to determine the mass of air entering the
engine in order to determine the correct amount of fuel to inject for a given desired fueling target. With the Subaru
ECU, it is represented in terms of mass airflow (grams per second), which, along with engine speed (RPM), is
used to determine load (grams per crankshaft revolution). For the Subaru ECU, load is not only used to determine
the proper injector pulse width, but also the desired fueling/timing targets. The modern factory Subaru determines
mass airflow via the mass airflow (MAF) sensor. The MAF sensor-based system attempts to directly measure the
actual mass airflow (given the relationship between MAF voltage and airflow for a specific intake and sensor).
Load is then calculated as a function of engine speed (RPM) and the measured mass airflow.

Speed Density, on the other hand, attempts to estimate load via other inputs. The basis for this calculation is
given by the ideal gas law. The ideal gas law is a relationship in physics between various inputs that allows for an
estimation of the mass of an ideal gas. In our case, the ideal gas is the air entering the combustion chamber of
the motor. The variables involved in the calculation (for COBB SD) include manifold absolute pressure (from the
MAP sensor), cylinder charge temperature (i.e. approximated by our IAT sensor input), volumetric efficiency (from
our VE calibration table), and engine displacement (tunable parameter). From this, an estimation of load can be
made (see “SD Math” section for a detailed explanation of the math involved).

What is Volumetric Efficiency?

Simply put, it is the amount of air inducted into the engine relative to the engine’s displacement. An engine is es-
sentially an air pump and volumetric efficiency (VE) defines how efficient that process is. A volumetric efficiency of
100% would indicate that the amount of air inducted is the same as the engine displacement (at standard condi-
tions) for a given engine cycle. If all engines always operated at 100% VE, there would be no need to account for
VE in determining the SD load. But, this is far from the case.

Numerous factors impact how VE varies with conditions for a given engine. As far as engine hardware, this can
include, but not limited to, the entire intake tract (from air filter to turbo to intercooler to throttle body), all exhaust
components, intake manifold design, cylinder head design, intake/exhaust valve design, compression ratio,
camshaft timing and lift, and so on.

Mechanical/aging issues, depending on their severity, can also potentially impact VE. Examples would include
combustion deposits, intake/exhaust leaks, and loss of engine compression.

Copyright © 2014 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 11
Among engine operating inputs, VE is most likely to change according to manifold absolute pressure (MAP) and
engine speed (RPM). This is why the VE table uses MAP and RPM axes (as is typical with most SD solutions).

Volumetric Efficiency Table

The VE table is the primary means by which an SD tune is accomplished for a given car. Once tuned, changes to
engine hardware and/or mechanical/aging issues that arise (as described in the section above) may require re-
tuning of the VE table. What areas of the table that need to re-tuned or tweaked is going to depend on the change
itself and how it impacts actual VE across a range of MAP and RPM.

Tuning the VE table is simply determining the actual VE so that the SD load is as close to the actual load as pos-
sible. This can be accomplished by comparing the ECU’s fueling target with actual fueling (via wideband o2 sen-
sor). All things being equal, increasing VE in the VE table will result in an increase in SD load (for the MAP/RPM
area impacted) and fueling will become richer. On the other hand, decreasing VE in the VE table will result in a
decrease in SD load with fueling becoming leaner. Keep in mind that because load is also used as an input to ig-
nition timing and the primary fuel tables (among other tables), the change in load will also impact the desired tar-
gets for these tables.

Generally speaking, the peak VE for a given column of the VE table will generally occur at the RPM of the motor’s
peak torque and VE will progressively drop on either side of that RPM point. Also, VE generally tends to increase
as MAP increases. Keep in mind that these are not hard and fast rules. You may find that the VE tune necessary
for a given car does not follow these guidelines completely.

SD Modes
COBB SD gives you the option of running several different modes that determine how load will be calculated. The
modes are determined by the following “SD Mode” table:

MAFmode – A value of 0 in this table will enable full-time MAF mode. This results in airflow and load being deter-
mined exactly as the factory ECU logic dictates. That is, airflow is determined based on MAF sensor input and the
“MAF Calibration” table and load is determined as function of this airflow and RPM. This requires that the MAF
sensor is still installed, functional and properly calibrated. In this mode, the SD load will still be calculated even
though it is never used. This will allow you to compare the SD load to the MAF sensor-based load via the “SD
Load (Post-Comp)” and “Calculated Load (MAF Sensor Based)” monitors. In addition, you can also tune the SD
tables in this mode. The “SD VE Estimated (MAF)” monitor (which is available in any mode) will estimate VE
based on the MAF sensor-based load (and other inputs) which you can use as reference to get an initial tune of
the VE table started. In MAF mode, this will allow you to make changes to the VE table while still running the MAF
sensor-based tune.

Copyright © 2014 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 12
SDmode – A value of 1 in this table will enable full-time SD mode. That is, the load normally determined by the
“MAF Calibration” table and RPM in the factory ECU will always be replaced by the SD load calculation deter-
mined by your SD tune. Even though the MAF sensor-based load will not be used by the ECU in this mode, you
can still monitor/log this value via the “Calculated Load (MAF Sensor Based)” monitor. If the MAF sensor is still in-
stalled, functional and properly calibrated, you can use this value as a comparison to the SD load determined by
your SD tune.

Hybridmode – A value of 2 in this table will enable hybrid mode. This allows you to switch to either MAF or SD
mode based on a series of tunable thresholds for throttle, RPM, MAP and MAF voltage. That is, you can decide
when the ECU will use the MAF sensor-based load and when it will use the SD calculated load. When transition-
ing between MAF and SD mode (or vice versa), the current load will be a blend of the two mode calculations to al-
low for a smooth transition. The speed of this transition can be tuned.

SD TUNING

Tuning SD – Mechanical Configuration

Before jumping in and starting tuning, make sure the car meets the minimum hardware requirements outlined in
the “Hardware Requirements” section found earlier in this document. Failure to do so can result in an inconsistent
tune as well as potential engine damage.

Tuning SD – Getting Started

Installing SD – Make sure an SD map has been reflashed to the car’s ECU and you have the latest SD-
capable software and firmware updates as outlined in the “SD Installation” section earlier in this document.

Opening Map – Run the AccessTUNER software, select the “Speed Density” checkbox, and select the
ECU of the car you are tuning. Open the SD map you reflashed to the vehicle.

Tuning SD – Initial Map Configuration

Airflow and Load Limits

The COBB SD ECU significantly raises the airflow and load limits inherent to the 02-05 USDM WRX ECU. There
is no need to deal with hacks that, for example, halve the airflow/load references and double the fuel injector
scale. Instead, you simply tune with the real-world actual airflow, load and fuel injector scale values. The COBB
SD ECU allows you to tune up to 10.0 g/rev of load for MAF and SD operations. And up to 655 g/s of airflow for
MAF operations. SD load is directly calculated so there is no need to worry about an airflow limit for SD operation.

Copying Existing Tune

Unlike COBB SD for the 2.5L USDM Subarus, the 2.0L USDM implementation does not allow for opening existing

Copyright © 2014 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 13
non-SD maps when an SD ECU is selected. This was necessary due to the internal changes required to raise the
airflow and load limits inherent to the factory 2.0L ECU. However, you can copy and paste table values from one
map to another with two instances of the AccessTUNER software open. There are two caveats to keep in mind
when attempting this procedure:
• Some of the OEM table sizes have changed between the non-SD and SD ECUs. This is especially the
case for the 02/03 ECU. Additionally, the Boost Targets and Dynamic Advance tables are smaller in size
for all SD ECUs. If there is a difference in size for a given table you are copying, you will have to modify
the new table in the SD tune appropriately.
• If your old non-SD tune used a hack to overcome the factory ECU's load/airflow limits (by manipulating air-
flow, load, and fuel injector scale references), you will want to modify these in your new SD tune to repre-
sent the actual real-world values. For example, a typical method for the existing hack would be to use the
“halving” method which involves halving the airflow and load values in every table and doubling the fuel in-
jector scale. In this example, when you transplant your tune to the SD ECU, you will want to double the
airflow and load values and halve the fuel injector scale in your old tune. Keep in mind, however, that there
are other ratios that could be used in the original hack and not necessarily the “halving” method. You need
to be aware of what was actually used in the original hack before copying values. Failure to account for
this in the new tune may cause engine damage.
The following is the procedure to copy and paste your existing table values from your old non-SD map to your
new SD map:
• Run AccessTUNER and select the vehicle's ECU (with speed density box unchecked) and open your ex-
isting non-SD map.
• Run another instance of AccessTUNER, selecting the vehicle's ECU (with speed density box checked)
and opening the newly created SD map that you reflashed to the vehicle.
• Double check that the size of the table you wish to copy hasn't changed between the old map and the new
SD map. If it has, then you will need to manually modify parts or all of the table in the new map appropri-
ately.
• In the non-SD map, select the cells of a table you wish to copy with the mouse and hit Ctrl-C (or select
“Copy Selected Cell Values” from the Edit menu).
• In the SD map, select the first cell in the table (corresponding to the area you wish to copy) and hit Ctrl-V
(or select “Paste Copied Cell Values” from the Edit menu). If you also wish to copy an axis value, select
the first corresponding axis cell in the SD map and paste as well. For 3D tables, you will need to do this for
each of the two axes.
• If your old map has hacked airflow/load/fuel injector scale values (in order to overcome the factory
airflow/load limits), you will want to modify these values in the SD tune if they are present in the table you
are copying (as described earlier in this section).
• When you are finished, go back through the tune to make sure everything looks reasonable and that noth-
ing was missed before reflashing the new map to the vehicle.

Copyright © 2014 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 14
Manifold Pressure Sensor Diagnostic Trouble Codes
If a diagnostic trouble code is set related to the manifold absolute pressure (MAP) sensor (with or without a check
engine light), the ECU will switch to an alternate failsafe calculation for manifold pressure based on load. Because
MAP is a key input to the load calculation in SD mode, this results in a feedback loop in which actual MAP no
longer plays a role in determining SD load. This erratic load calculation could cause a dangerous lean condition
(with incorrect timing), if actual MAP is increasing at the time (for example, if the vehicle were accelerating). It will
also likely cause the car to eventually stall.

One or more of the MAP sensor diagnostic trouble codes (DTCs) can even be set when there is no actual failure
of the MAP sensor. This can occur in the following circumstances:

• High level of boost (i.e. MAP voltage) is seen by the factory or aftermarket MAP sensor that exceeds the
factory DTC limits.

• Lower level of MAP voltage due to aftermarket MAP sensor’s expanded relative range.

The following are the MAP sensor related DTCs in question:

• P0068 - Manifold Absolute Pressure Circuit Range/Performance Problem

• P0107 - Manifold Absolute Pressure Circuit Low Input

• P0108 - Manifold Absolute Pressure Circuit High Input

It is critical that if the possibility exists that any of these DTCs may be set, the fault detection should be disabled
for the SD ECU. This can be accomplished by unchecking the following toggles in the “Edit” -> “Advanced Param-
eters” -> “Toggles (Base)” menu, saving the map, and then reflashing the map to the ECU (Note: this is different
from the typical procedure to disable DTCs) :

• MAP Sensor P0068 Fault Detection

• MAP Sensor P0107 Fault Detection

• MAP Sensor P0108 Fault Detection

For the P0107 and P0108 codes, the DTC voltage limits can be configured in the AccessTUNER software via the
“MAP Sensor Voltage DTC Limit...” tables (found in the “Boost Control Tables” -> “Boost Limiters” group). Addi-
tionally, the “MAP Sensor Voltage DTC Delay...” tables determines how long the voltage limit must be exceeded
before the DTC is set. Both of these tables can be used in place of disabling the fault detection of the two DTCs if
a proper range can be determined. Keep in mind that the P0068 DTC limit is not configurable. This DTC is set
when MAP voltage is not above a specified threshold given specific conditions related to RPM, throttle, and load.

Mass Airflow Sensor Diagnostic Trouble Codes

The COBB SD ECU disables the ECU’s primitive MAP-based load calculation from coming into play when a MAF
sensor related DTC is active (regardless of whether the check engine light is on or not). As such, if the MAF sen-
sor is removed, disconnected, or otherwise non-functional, you will only be able to run the car in full-time SD

Copyright © 2014 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 15
mode and the car will not start or run when MAF mode is active. If this is the case, you will need to take additional
steps in order to avoid other MAF failure related ECU logic from coming into play.

The following are the MAF sensor related DTCs in question:

• P0101 - Mass Air Flow (MAF) Sensor Circuit Range/Performance Problem

• P0102 - Mass Air Flow (MAF) Sensor Circuit Low Voltage

• P0103 - Mass Air Flow (MAF) Sensor Circuit High Voltage

It is critical that if the possibility exists that any of these DTCs may be set, the fault detection should be disabled
for the SD ECU. This can be accomplished by unchecking the following toggles in the “Edit” -> “Advanced Param-
eters” -> “Toggles (Base)” menu, saving the map, and then reflashing the map to the ECU (Note: this is different
from the typical procedure to disable DTCs) :

• MAF Sensor P0101 Fault Detection

• MAF Sensor P0102 Fault Detection

• MAF Sensor P0103 Fault Detection

For the P0103 code, the DTC voltage limit can be configured in the AccessTUNER software via the “MAF Sensor
Voltage DTC Limit (High Input)” table (found in the “Miscellaneous Limiters” group).

Load and MAF Compensation

The factory ECU includes airflow and load compensation tables that are still applied to the MAF sensor-based air-
flow/load calculations in the COBB SD ECU. However, these compensations are NOT applied to the SD calcula-
tions. It is important to look at and understand these compensations as they will ultimately determine the final air-
flow and load that is used in MAF mode:

• MAF Compensation - The airflow compensation table is called “MAF Compensation (Intake Temp)” and
can be found under the “Sensor Calibrations” group. This determines the compensation to final airflow
based on IAT and current airflow in MAF mode.

• Load Compensation - The load compensation table is called “Load Compensation (Manifold Pressure)”.
This determines a compensation to final load based on MAP and RPM in MAF mode.

Conservative Fuel and Timing Maps

It is important to consider using conservative fuel and timing maps during the initial SD tuning phase. When start-
ing your tune, the VE table will not be perfect until you’ve had a chance to dial it in. While you are tuning the table,
any error from actual VE will not only result in the incorrect fueling, but also the incorrect load values. If the VE for
a given cell in your table is less than actual VE, this will result in load that is less than actual load. Because the
timing tables (“Primary Ignition” and “Dynamic Advance”) use load as an input, the timing advance will generally
be higher than intended in this case. The primary fuel table also use load as an input and the desired fueling tar-
get will generally be leaner than intended in this example. This issue would also impact any other table that uses
Copyright © 2014 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 16
load as an input. If the VE for a given cell in your VE table is greater than actual VE, the opposite will occur,
where load will be greater than actual. This is why you generally want to bias your VE values to a higher estima-
tion of VE when starting out, rather than a lower one.

Fuel Injector Scale and Latency

Tuning for new injectors for COBB SD is no different than the factory MAF sensor-only process. As such, you will
need to make sure your fuel injector scale and latency values are tuned correctly prior to tuning with SD. If you
are changing injectors at the same time as changing over to SD, you’ll need to at least make sure that you have
reasonable starting values for the new injectors.

Engine Displacement

COBB SD uses engine displacement as one of the inputs to the SD load calculation. As such, this requires that
you input the correct displacement of the motor you are tuning via the “SD Engine Displacement” table (under the
“Speed Density” table group). This should be the closest value (in liters) to the car’s actual engine displacement.
The default value is 1.994 liters, which is the closest actual displacement to the factory 2.0L USDM turbo motor.

VE Table Axis Scaling

The default values for the MAP axis of the VE table is in 2.5 psia steps up to a maximum of 45 psia and the de-
fault values for the RPM axis is in 350 RPM steps up a max of 7700 RPM. You need to make sure that the maxi-
mum values are enough for the max MAP and RPM that the motor will likely see. When either RPM or MAP ex-
ceeds their corresponding max axis values, the ECU will continue to use the VE values in the last row (or
column). If you need to make a change, simply re-scale the axis values and reflash those changes to the ECU.
Keep in mind that the MAP axis is in units of manifold absolute pressure, not relative pressure. If you wish to de-
termine the corresponding relative pressure values (for reference) simply subtract your barometric pressure from
MAP. The barometric pressure can be read via the “Barometric Pressure” monitor. If you are at/near sea level
(barometric pressure around 14.7 psi) and you wish to quickly determine the relative pressure in your head, sim-
ply subtract 15 psi from the MAP axis value you are looking at (for example, 30 psia – 15 psi = 15 psig). This will
give you a quick approximation when you wish to think in terms of relative pressure.

Reflash Changes
Make sure you save this map with the initial changes and reflash to the car before continuing with the tune.

Tuning SD – Starting Values for the VE Table

The default values for the VE table are set to 100% across the entire table. This is not meant to be a starting point
Copyright © 2014 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 17
to run the vehicle in SD mode and tune. Instead, you want to consider coming up with initial values in the VE table
that are reasonable enough to run the car in SD mode. The following gives suggestions based on whether or not
there is a functional MAF sensor installed in the car.

Car with Working and Accurate MAF Sensor

If the car has a properly functioning and calibrated MAF sensor installed, you can use the “SD VE Estimated
(MAF)” monitor to aid in determining starting values for your VE table. If there are issues with the accuracy of the
MAF sensor (and resulting load calculation) or problems with the IAT sensor, MAP sensor, and/or engine dis-
placement inputs, this will impact the accuracy of the estimated VE monitor. Also, because MAF sensor-based
load tends to be overestimated on throttle tip-in and underestimated on throttle tip-out, the estimated VE monitor
will also be inaccurate during these periods.

The idea here is that you will be tuning your starting VE table while remaining entirely in MAF mode. That is, if the
car you are tuning runs well in MAF mode, you can get a decent initial set-up for the VE table before actually run-
ning the car in SD mode. In addition to the estimated VE monitor, you can also view/log the “SD VE (Com-
manded)” monitor, which is the current VE table look-up. This is VE based on the VE table that would be used in
the SD load calculation if SD mode was active. By comparing the estimated VE and commanded VE, you can
tweak the VE table appropriately. Keep in mind that the estimated VE monitor is not a substitution for actually tun-
ing the VE table in SD mode based on actual fueling.

There are a handful of methods you can use to take advantage of the estimated VE monitor in order to come up
your starting VE tune in MAF mode. You may find some or all of these are useful, depending on what tools you
have available and the time you have to spend with the vehicle:

• Data logging – One way to come up with reasonable initial VE values for specific MAP and RPM areas of
the VE table is via data logging. You’ll need to at least log the “SD VE Estimated (MAF)”, “Manifold Abs.
Pressure”, “Engine Speed”, and “Throttle Position” monitors. Logging throttle position will allow you to filter
out the inaccurate estimated VE values when throttle is rapidly changing (ex. during tip-in and tip-out).
With data logging, the more data points you have for a given MAP/RPM range, the more accurate your es-
timation of VE will be as it will allow you to throw out more extreme values. But, you should keep your
starting VE values towards the higher side to reduce the chances of a VE error in the starting map result-
ing in a lean condition when you start to tune in SD mode. Keep in mind that if the IAT sensor becomes
heat soaked, which may occur with extended idling and stop and go driving, the estimated VE monitor will
overestimate VE compared to actual. You may want to also log “Intake Temperature” and “Vehicle Speed”
so you can see if there’s any relationship between a higher estimated VE and the conditions that may re-
sult in a heat soaked IAT sensor. As you accumulate data, modify the VE table appropriately and then
add the “SD VE (Commanded)” monitor for additional logging. You can then compare your table VE to es-
timated VE so you can tweak VE further.

• Steady-state tuning on dyno – If you are tuning on a load-bearing dyno, you can also hold cells of the VE
table and come up with an initial VE for a given cell based on the estimated VE monitor. This would be
accomplished by enabling live tracing on the table and taking advantage of real-time tuning. As you view
the estimated VE monitor, it will show a range of values as the inputs to this calculation rapidly change.
Try to use the highest reasonable value shown. It is not realistic to attempt to do this for every cell of the
table just for an initial VE pre-tune. But, you could focus on, for example, the RPM row that is likely to be
Copyright © 2014 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 18
 closest to peak torque and, therefore, where VE is likely to be highest (for a given MAP column).

• Filling in the blanks – Depending on the time available for the tune, it may not be feasible to try to esti-
mate the majority of the cells in the VE table using the methods described above. Instead, you can focus
on more narrow portions of the map representing low, mid and high MAP areas and then “fill in the blanks”
between these areas. Generally speaking, as MAP increases, VE also increases. Also, as it relates to
RPM, VE will tend to follow the torque curve of the motor, generally peaking at the same RPM as the peak
torque of the motor. These are not hard and fast rules, but they are sufficient enough to allow you to inter-
polate between the areas of the map you have worked on for the starting VE tune. Keep in mind that the
WOT area of the VE table (i.e. high MAP) is going to be the most crucial area of your pre-tune. This is be-
cause load errors at high MAP have the most potential to cause engine damage. You do not want your
first WOT pull in SD mode to result in an overly lean condition with more timing advance than anticipated.
This is also a reason to go with more conservative fuel and timing maps until you get the VE table di-
aled-in after switching to SD mode.

Once you have a reasonable starting VE table, you may want to consider bumping up the values in the entire ta-
ble just to make sure your pre-tune is more likely to overestimate VE than to underestimate it.

Car without Viable MAF Sensor Configuration

If the MAF sensor has been removed from the car or is otherwise inaccurate, inoperable or not properly cali-
brated, then there is not a means to directly estimate VE for a given car. Over time, however, you will see specific
patterns in the VE table for cars you’ve tuned (for a given set of mods) and setting up a reasonable starting VE ta-
ble for similar vehicles will not be difficult. If you have no frame of reference, however, you can start with some
general guidelines. The default values in the VE table will be 100% for all cells. This will be generally too high in
the lift-throttle, idle, and cruise areas (i.e. much too rich). Generally speaking, you may see more like 50%-60% in
lift-throttle/idle areas and 60%-80% in cruise areas. Moderate boost may be in the 80%-90% range. Higher boost
may be in the 90% to over 100% range. To stay on the conservative side of things with your starting VE table, you
want to bias your estimate towards higher values rather than lower. This will reduce the chance of a lean condi-
tion when the table VE is less than actual VE. This is especially critical at higher boost where a lean condition
(and greater timing advance) due to lower than actual load would be more of a problem.

Copyright © 2014 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 19
Tuning SD – The Tables

“SD Volumetric Efficiency” Table

The “SD Volumetric Efficiency” table is the primary means by which the SD tune is accomplished. Once you have
your starting values configured for this table (as described in the previous section), you can get to the actual tune.
You’ll need to make sure you are in SD mode (1 in the SD mode table) and that the car is at operating tempera-
ture. As you hit the MAP/RPM area corresponding to a given cell, if your wideband o2 (wbo2) sensor reads richer
than what the ECU is targeting, you reduce VE for that cell. If your wbo2 sensor reads leaner than what the ECU
is targeting, you increase VE for that cell. You continue this process until you’ve dialed in as much of the VE table
as you can.

The final ECU fueling target can be determined via the “Commanded Fuel Final” monitor. This may be different at
times than your primary open loop fueling map due to additional fueling compensations that come into play, such
as post-start/warm-up enrichment, long-term fuel trims, and others. Also, in closed loop, the ECU is not always
targeting 1.0 lambda (i.e. 14.7:1 AFR gas). The “Commanded Fuel Final” will account for all these changes as it is
the final fueling target used to ultimately determine injector pulse width.

You should be aware of when the closed to open loop fueling transition occurs for your tune. When the transition
to open loop happens, you may see a lean or rich spike when the short-term fuel trims are no longer applied if

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 20
that area of VE table needs work. You can determine the closed/open loop status via the “Closed/Open Loop
Switch” monitor.

If the car has a functional and accurate front o2 sensor, you can also tune VE in closed loop via the short and
long-term fuel trims. Simply add together the “A/F Correction #1” and “A/F Learning #1” monitors and try to get
this sum as close to zero as feasible by manipulating VE. If the sum is positive, the ECU is adding fuel because of
a lean condition and you will need to increase VE. If the sum is negative, the ECU is removing fuel because of a
rich condition and you will need to decrease VE.

You may be tempted to simply manipulate the VE table to hit the fueling target you desire for a given MAP/RPM
area regardless of the commanded fuel final (and the primary open loop fueling table). While this is certainly pos-
sible, it does have a few downsides. First, tuning this way would result in VE values that are not representative of
actual. That is, you could not easily compare them to other tunes. It can be quite useful to compare VE tables
across tunes as it gives you an idea of VE changes with different mods. It can also help you use those values as
a starting point for new tunes for cars with similar mods. Second, making changes to fueling after completing the
tune is a little easier if the VE table represents actual VE. You would simply modify the primary open loop fueling
table to your new desired open loop targets.

When you’ve dialed in the VE tune, use the AccessTUNER graphing to help to smooth the table. Generally
speaking, actual VE should not drastically change with small changes in RPM and MAP and, therefore, you
should not see drastic changes in VE between cells in the VE table. Smoothing the table will result in a more con-
sistent load calculation and therefore, more consistent fueling and timing.

In addition, you will also need to estimate the areas of the VE table you were not able to hit when tuning. The val-
ues should be reasonable given the adjacent cells that were directly tuned. Also, keep in mind when estimating
VE, that VE will generally increase with an increase in MAP and for RPM, it will generally follow the torque curve
of the motor (with the peak VE occurring at the RPM in which peak torque occurs and decreasing on both sides of
that peak).

“SD Load Compensation (Intake Temp)” Table

Like the VE table, the “SD Load Compensation (Intake Temp)” table is another critical component of the SD load
calculation. In order for our SD load to be accurate, we not only need to know MAP, VE, and engine displace-
ment, but we also need to know the temperature of the cylinder charge. Without this input and its corresponding
correction, any given cell in our VE table would only be valid at the charge temperature in which it was tuned.
Without this correction, if the cylinder charge temp dropped from our tuned charge temp, we would go lean and if
it went up, we would go rich.
Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 21
We have no means of directly measuring cylinder charge temperature so we have to rely on measuring intake
temperature at some point as a means of estimating charge temp. Generally, the closer to the combustion cham-
ber we measure the intake temp, the closer we’ll be to the charge temp. The Subarus we are tuning do not have
intake manifold temperature sensors, which would be the typical implementation for SD from the factory. Instead,
Subarus have the IAT sensor housed in the MAF sensor assembly which is located pre-turbo in the intake tract.
This is the only input that the factory ECU has for intake temp. In the case of a car with a front-mount intercooler,
the MAF/IAT assembly could be relocated post-intercooler (or an aftermarket IAT sensor installed in the same lo-
cation) which would give you a more accurate representation of charge temperature. Regardless, having the abil-
ity to tweak the charge temp correction is an important component to the SD tune.

Normally, if applying the ideal gas law to estimate engine load, the non-linear correction factor for charge temp
would be wrapped up in the ideal gas law equation. However, in order to allow the ability to tune this correction, it
has to be externalized. For COBB SD, this is accomplished by first calculating SD load at a specific reference
temperature of 86 degrees Fahrenheit. Then the “SD Load Compensation (Intake Temp)” is applied (along with
the other SD compensations) to determine the final SD load. As you can see from the default values in the table,
the 86 deg. F column has 0% correction. The default values for the rest of the columns are scaled with the ideal
gas law and our reference temperature in mind. As such, this allows for a tunable charge temp correction.

What this all means is that there may be scenarios where your intake temp reading may be notably different from
the actual charge temp. As such, the default ideal gas law based values in this table may need to be tweaked. As
you can see, the table also has a MAP axis. This allows you to come up with different sets of corrections based
on MAP. This is important because MAP is one of the primary factors which can impact necessary tweaks to the
table. The default values for the MAP axis are mostly in vacuum, but these can be changed as needed (axis value
changes require a reflash).

One scenario in which the IAT reading may be notably different than the charge temp is when using an IAT sen-
sor that is placed before the turbo (known as a draw-through location). This is the case with the factory placed
MAF housing, which also contains the IAT sensor. When the IAT sensor is pre-turbo/intercooler, the impact of the
turbo and intercooler on charge temperature is not reflected in the IAT reading. Generally speaking, at very low
MAP, the effect is negligible. However, as MAP increases, you may find the need to tweak the IAT correction dif-
ferently based on the input temperature and MAP. Keep in mind that heat soak of the IAT sensor (described be-
low) may also be another factor that you would have to account for in addition to the pre-turbo IAT reading. This is
why, if you are running a front-mount intercooler, we recommend that your IAT sensor be placed in a blow-
through configuration (i.e. post-IC), whether than involves relocating the factory MAF/IAT assembly or wiring-in
an aftermarket IAT sensor to the factory IAT input.

The other scenario in which the IAT reading may differ markedly from actual charge temp is when the IAT sensor
becomes heat-soaked. What happens is that the sensor itself absorbs heat from the surrounding air and intake
piping and reads a temperature higher than the actual temperature of the airflow. Generally, this effect is progres-
sively more pronounced at lower MAP/airflow and therefore you may be more likely to see it in conditions such as
stop and go driving, extended idling, or a hot restart. Because the IAT reading is higher than the charge temp, this
would cause the load calculation to be less than actual and you would end up going lean (as well as causing po-
tentially more timing advance). When MAP/airflow does increase, the heat soak effect of the sensor does not in-
stantaneously diminish, so the heat soaked reading may continue at higher MAP for a period of time. This may
present a problem if, for example, the car is subject to high loads (such as wide open throttle) soon after heat-
soak of the sensor sets in. This is one reason why a wideband o2 sensor and gauge in the car running SD is im-
portant, as well as the driver understanding how to recognize a problem.
Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 22
There are some steps that can help partially mitigate specific heat soak scenarios. First is that the use of a short-
ram intake in a draw-through IAT configuration appears to be the combo where heat soak potential is the great-
est. The post-intercooler IAT sensor placement appears to have the least heat soak potential (but can still defi-
nitely occur). Second is some tweaking of the IAT comp table may help in certain cases. Generally speaking, the
higher the IAT reading, the more likely heat soak of the sensor can occur. This may not always be the case, but it
is something you can attempt to mitigate by increasing the load compensation at higher IATs (i.e. higher correc-
tion means higher load and more fuel). For example, you may find that the car becomes progressively leaner as
IATs increase above 140F. In that case, tapering the negative correction progressively (from the default values) at
IATs of 140F and greater could help. Also, the effect is likely to be more pronounced at lower MAP, so plan your
changes accordingly. You may want to, however, keep some of your changes at higher MAP as the heat soak ef-
fect may linger for a period in going from low MAP to high MAP. Keep in mind, though, that the heat soaked IAT
reading can potentially occur even at colder temps. For example, the charge temp may be 40 deg. F and the IAT
reading is 60 deg. F. This would still be a heat soak scenario (although one which would be difficult to account for)
even though the temperatures are relatively mild.

“SD Load Compensation (Coolant Temp)” Table

It is not necessary to tune the “SD Load Compensation (Coolant Temp)” table as far as the ideal gas law and our SD
load calculation is concerned. As such, the default values are set to 0% across the entire table. However, you
may find some circumstances where it may be useful to tweak this table for your SD tune. One example would be
more extreme coolant temp readings (on the high end). You may want to increase correction at this extreme as
this may mean a more likely heat soak scenario for the IAT sensor (high ECT likely means higher radiant engine
heat). As described in the previous section, a heat soaked IAT sensor (relative to actual charge temp) will result in
a lean condition with SD.

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 23
“SD Load Compensation (Barometric)” Table

As you go up in altitude, the barometric pressure drops. The SD load calculation for COBB SD accounts for this
change because MAP is a part of the ideal gas law calculation and therefore, as barometric pressure decreases,
MAP also decreases (all else equal) and SD load will also decrease. However, exhaust gas backpressure also
decreases as barometric pressure decreases, which can impact VE. As such, you may need to tune the “SD Load
Compensation (Barometric)” table if the car is going to see notable changes in altitude. In addition to the baromet-
ric pressure axis, this table also has a MAP axis. This will allow you to tune the compensation against MAP where
the effect of exhaust gas pressure on VE may vary.

Tuning SD – Miscellaneous Tables

The following are some miscellaneous SD-related tables that generally would not need to be tweaked for most SD
tunes:

Load Filtering in MAF and SD Mode

The factory ECU calculates a smoothed version of load for the final load calculation. Smoothing is a means of fil-
tering out “noise” in a value by considering the change in the current value versus the previous value. That is, it
will tend to dampen larger short-term changes. The dampening effect is dependent on the smoothing factor used

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 24
in filtered calculation. The factory ECU’s load smoothing factor is set up for the response characteristics of the
MAF sensor which is notably different than the MAP sensor, especially during transients. As such, the load
smoothing is not optimal when calculating load based on SD inputs (i.e. when the primary input is the MAP sen-
sor) and would potentially cause tip-in/tip-out and idle issues (among others) if not accounted for.

COBB SD solves this problem by allowing for separate tunable load smoothing factors for SD and MAF mode.
The smoothing factor is a value that ranges from above zero to 1.0. A value of 1.0 means that there is no smooth-
ing involved and the load calculation is not manipulated. With values below 1.0, the smaller the smoothing factor,
the more dampened the load will become. The “Load Determination Smoothing Factor (MAF Mode)” table, found
under the “Speed Density” -> “Miscellaneous” group, has a default value that is the same as the factory map
(0.125). Generally, this should not be changed. The “Load Determination Smoothing Factor (SD Mode)” has a de-
fault value of 1.0. We have found that this works best in most scenarios for SD tuning (i.e. no smoothing for load
in SD mode). If, however, you are seeing load changes that are too erratic (or cause a richer than expected tip-in)
and cannot be solved by further tuning/smoothing of the VE table, you may find it useful to lower the SD mode
load smoothing factor. Generally, though, it should still be much higher than the default MAF mode smoothing
value.

Map Determination Averaging Window

Because the MAP sensor is one of the crucial inputs to the SD load calculation, a collection of factory tables that
manipulate the final MAP value have been exposed in the software and can be found under the “Sensor Calibra-
tions” group (see table list below):

When the MAP delta (current MAP – previous MAP) is within the averaging window of the MAP delta threshold,
the ECU will use an average of the last two MAP values as follows:

(current MAP + previous MAP )

MAP =
2

There are two separate averaging windows which are used based on an RPM threshold (see table list above).

Generally, these do not need to be modified for SD tuning. However, if you find that MAP is too erratic (and not
representative of actual changes in MAP), you may find that expanding the MAP delta range in which the averag-
ing is applied may help. On the other side of the coin, if you find that MAP is not reacting fast enough to actual
changes in MAP, you may find it useful to narrow the MAP delta window.

SD Feature Activation
The SD feature set can be completely deactivated by setting the following table to zero:

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 25
Disabling the SD feature will cause the ECU to revert back to the OEM logic for determining airflow/load (based on
MAF sensor input). However, this is different from MAF mode in that the SD load will no longer be calculated and
the opera- tional mode cannot be changed to SD or Hybrid via a real-time map. SD feature deactivation should
not normally be used unless you experience a bug that impacts the safe operation of the SD ECU (although it is
recommended that you switch to the non-SD ECU in this case).

Tuning SD – Post-Tune Recommendations

When you feel your SD tune is complete, there are several things you should consider in order to maximize the
long-term reliability of the SD tune:

• It is important that your reflashed map has the same “SD mode” table value as your real-time tune. For ex-
ample, if you switch to an SD mode of 1 (i.e. SD load) via real-time and tune the car, but do not change
the SD mode from the default of 0 in the reflashed map (i.e. MAF mode), then the ECU will end up switch-
ing to MAF mode the next time the car’s battery is disconnected or the ECU is reset. In addition to the SD
mode, it is also important to reflash you final tune to make sure the final tune is always part of the “base”
map.

• If the SD tune was completed on a dyno, it is important to drive the car under conditions that it is likely to
see during normal operation and verify that the tune is safe and that there are no driveability concerns.

• It is highly recommended that a wideband o2 sensor and in-car gauge is installed in the car running SD
and that the owner/driver of the car is instructed on how to read the gauge and how to determine what is a
normal reading for the tune and given operating conditions.

• The owner/driver of the car needs to be instructed about the potential for a heat soaked IAT sensor and
under what conditions this is likely to occur (extended idling, hot soaked re-start, etc). It is important that
the owner/driver understands that putting the car immediately under heavy load when a heat soaked IAT
sensor may be a possibility should be avoided when running SD. Instead, under these conditions, allow
the car to get moving without getting heavily into boost for at least a few minutes before putting the car un-
der heavy load. A wideband o2 sensor with in-car gauge can help here as a heat-soaked IAT sensor will
generally cause the fuel reading to be leaner than expected.

• The owner of the car also needs to understand that practically any engine mod that impacts airflow in any
way may require a re-tune or tweaking of the VE table for cars running SD. This is important as even
seemingly minor mods that a MAF sensor-based tune would have no problem accounting for, might cause
a significant enough change in VE that there could be fueling and load issues for SD.

Tuning SD – Additional Topics

Recommended Calibrations for Aftermarket IAT Sensors

When an aftermarket IAT sensor is installed, you will need to tune the IAT sensor calibration which will be differ-
ent with an aftermarket sensor as compared to the factory IAT sensor. This is done via the “Intake Temp. Sensor

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 26
Calibration” table in the “Sensor Calibrations” group. Unlike the factory 2.0L ECU, the COBB SD 2.0L ECU allows
you to edit each individual IAT voltage cell in this table.

The following shows the recommended IAT scaling for the GM and AEM aftermarket IAT sensors (each screen-
shot shows one half of the table). Keep in mind that both the axis values (top row) and data values (bottom row)
change in this case. Note:Thesecalibrationsareprovidedforyourconvenienceonlyanddonotrepresentanen-
dorsementfororagainstanyparticularproduct.Manufacturer’sspecificationscanchangeatanytime.Itisimpor- tantthat
youverifythesensorcalibrationyouaregoingtousebeforetuningSD.

Recommended Calibrations for Aftermarket MAP Sensors

When an aftermarket MAP sensor is installed, you will need to set up the MAP sensor calibration appropriately.
This is done via the “MAP Sensor Calibration (Multiplier)” and “MAP Sensor Calibration (Offset)” tables in the
“Sensor Calibrations” group. Unlike the factory 2.0L ECU, the COBB SD ECU has been modified to allow for a
positive or negative offset value (rather than negative only). After the new calibration is complete, the boost read-
ing (via AccessTUNER software or AccessPORT) should be compared to an external boost gauge to verify accu-
racy before tuning SD. The recommended calibrations for some of the most popular aftermarket MAP sensors are
shown below (note: values are shown in psi). Note:Thesecalibrationsareprovidedforyourconvenienceonly anddo
notrepresentanendorsementfororagainstanyparticularproduct.Manufacturer’sspecificationscan changeatanytime.It
isimportantthatyouverifythesensorcalibrationyouaregoingtousebeforetuningSD.

1. AEM 3.5 bar (Part # 30-2130-50) –> MULTIPLIER = 12.499 psi , OFFSET =-6.246 psi

2. AEM 5 bar (Part # 30-2130-75) -> MULTIPLIER = 18.749 psi, OFFSET = -9.378 psi

3. GM 3 bar (Part #12223861) -> MULTIPLIER = 8.941 psi, OFFSET = 0.155 psi

4. OmniPower 3 bar (Part #MAP-STI-3BR) -> MULTIPLIER = 9.123 psi, OFFSET = 0.155 psi

5. OmniPower 4 bar (Part #MAP-STI-4BR) -> MULTIPLIER = 12.085 psi, OFFSET = 0.174 psi

Forcing Open Loop Fueling

In some cases, you may find it more straightforward to temporarily force the ECU into full-time open loop fueling
when tuning the VE table. This can be accomplished by the following procedure:
Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 27
 1. Change the “Primary Open Loop Fueling Min. Activation” table (found under the “Fuel Tables” -> “Open
Loop (Primary)” group) to 14.7:1 AFR (or 1.0 lambda for non-standard units).

2. In the “Primary Open Loop Fueling” table, change the 14.7:1 AFR cells (or 1.0 lambda cells) to the next
richest value. For example, change 14.7:1 to 14.59:1 AFR (or 1.0 lambda to 0.99 lambda).

3. Change the “Closed to Open Loop Delays” table (found under the “Fuel Tables” -> “Closed/Open Loop
Transition” -> “Delay” group) to zero for every cell.

4. Save map and reflash map to car.

5. Verify that the ECU remains in open loop full-time by viewing or logging the “Closed/Open Loop Switch”
monitor in the AccessTUNER software.

It is not recommended that you run full-time open loop operation after the tune is complete. This is because cer-
tain scenarios with SD that would result in fueling errors, such as a heat soaked IAT sensor or changes in VE due
to mechanical/aging issues, can be partially mitigated by short-term fueling corrections (and potentially long-term
fuel trims) when closed loop operation is active.

Long-Term Fuel Trims

You may find it necessary (or more optimal) to manipulate how the long-term fuel trims are determined when run-
ning SD (and in some cases, for MAF mode as well). The long-term fuel trims are determined based on patterns
of short-term correction. They are calculated and applied across four airflow ranges. These ranges are deter-
mined by the “A/F Learning #1” table (under the “Fuel Tables” -> “A/F Learning” group). The following is an exam-
ple:

As you can see, there are three values to this table. The airflow ranges are determined from these values as fol-
lows (given the example):

Range A: 0-<5.60 g/s

Range B: 5.60-<10.00g/s

Range C: 10.00-<50.00 g/s

Range D: 50.00+

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 28
The “D” range is of particular importance because any correction that is learned in closed loop for this range is not
only applied in closed loop but also open loop as well, since it is applied at any airflow above its threshold (in this
example 50+ g/s). This can be problematic if the correction is extreme and is removing fuel (i.e. correction is neg-
ative). For example, if the “D” range value was -5%, then fueling will be leaner by 5% in closed loop above 50 g/s
(in this example) and 5% leaner in open loop, including wide open throttle. That is, the ECU assumes that the cor-
rection learned in the “D” range is needed in open loop as well, which may not always be the case.

You can disable the “D” long-term fuel trims via the “A/F Learning #1 Modify Airflow Limit (Max)” table in the “Fuel
Tables” -> “A/F Learning” group. This is accomplished by setting this table’s value to less than the “D” range
threshold you are using (in this example, less than 50 g/s):

Additionally, you can also modify the allowable limits for A/F Learning #1 via the “A/F Learning #1 Limits
(Min/Max)” table (also under the “Fuel Tables” -> “A/F Learning” group):

You can view/log long-term fuel trims via the “A/F Learning #1” monitor. This shows the currently applied long-
term fuel trim. You can also view all four ranges via the “A/F Learning #1 Range A”, “…Range B”, “…Range C”,
and “…Range D” monitors or by live connecting to the ECU and viewing the real-time “A/F Learning #1” table.

Tuning MAF Mode

In MAF mode (i.e. a value of 0 in the SD mode table), MAF sensor-based tuning is the same with COBB SD ECU
as it is with the non-SD ECU. That is, tuning for specific intake hardware is accomplished via the “MAF Calibra-
tion” table. The only difference is that the SD ECU will calculate SD load (and other SD-related values) while still
in MAF mode, even though those values are never used in determining load. This allow you to, for example, get a
starting tune ready for the SD VE table while still running the car in MAF mode.

Tuning Hybrid Mode

Overview
In Hybrid mode (i.e. a value of 2 in the SD mode table), you are simply allowing the SD ECU to switch between
MAF mode and SD mode (and vice versa) based on a series of specific thresholds for MAP, RPM, throttle posi-
tion (TPS), and MAF voltage (MAFv). In addition, you can also control the behavior of the switching as well as the
speed of the transition.
Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 29
Hybrid Mode - Uses
There are a handful of reasons why hybrid mode might be a better choice than running pure SD mode. The fol-
lowing are some examples:

• If the car runs well on the MAF sensor-based tune but maxes out the MAF sensor on the high end, hybrid
mode could be used to switch to SD to avoid maxing out the sensor. Running hybrid mode in this case
would reduce the time and effort it would take to get the car to a well-tuned state when compared to a pure
SD tune.

• If the car has a big MAF set-up and runs well on the high end, but poorly on the low end (i.e. idle/cruise),
hybrid mode could be used to run SD on the low end and MAF on the high end to mitigate these issues.
This would also reduce the time and effort to finish the car’s tune vs. a pure SD tune.

Hybrid Lower/Upper Mode

The first thing to decide is what mode will be used on the lower end and what mode will be used on the upper
end. This will be either MAF on the lower end and SD on the upper end or MAF on lower and SD on upper. The
MAP, RPM, throttle and/or MAF voltage thresholds determine the switching from the lower to upper mode (and
vice versa). So, for example, if you wanted to switch to SD when one (or more) of these tunable thresholds are
exceeded, then you would want SD on the upper end and MAF on the lower end.

Whether lower/upper is MAF/SD or SD/MAF is configurable via the following table:

If
this table’s value is set to 0 (the default value), then MAF mode is used below the threshold(s) and SD mode is
used above the threshold(s). If it is 1, then SD mode is used below the threshold(s) and MAF mode above.

Hybrid Threshold Switching Behavior

You can also decide how the switching from upper to lower (and vice versa) will work based on the thresholds.
That is, how many of the thresholds must be exceeded to switch to upper mode and how many values must fall
below their corresponding thresholds to switch to the lower mode. This is determined by the following table:

Fo
r example, if this table’s value is 0 (the default value), then the following has to be met to switch modes:

• Switch from lower to upper mode: any single value exceeds its corresponding threshold.

• Switch from upper to lower mode: all values must fall below their corresponding thresholds (less the corre-
sponding hysteresis)

If this table’s value is 1 then the following has to be met:

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 30
 • Switch from lower to upper mode: all values must exceed their corresponding thresholds

• Switch from upper to lower mode: any single value falls below its corresponding threshold (less the hys-
teresis).

Hybrid Thresholds Overview

For the upper mode, a threshold is exceeded if the value in question is greater than the threshold. For the lower
mode, the value is considered to have dropped below the threshold if it is less than or equal to the threshold less
its hysteresis (threshold – hysteresis). The threshold table listing is shown below:

For example, if your MAFv threshold is set to 3.5v and the MAFv hysteresis is 0.25v, then the threshold is ex-
ceeded if MAFv is greater than 3.5v and MAFv has dropped below the threshold when MAFv is less than or equal
to 3.25 v (3.5v – 0.25v). The reason for the hysteresis is to avoid rapid transitions between MAF and SD modes if
the value were to hover around the threshold.

Hybrid Threshold Individual Deactivation

All four thresholds do not have to play a role in determining the MAF and SD mode transitions. You can remove
specific thresholds from the decision process by setting those threshold(s) to unachievable values (which will de-
pend on the threshold switching behavior explained above).

If the switching behavior is 0 (upper switch = single, lower switch = all), then setting any threshold to an extremely
high value that will never be achieved, will remove it from the decision process. For example, you could raise the
RPM threshold to an unachievable value of 10000 RPM and RPM would never play a role in the switching.

If the threshold switching behavior is 1 (upper switch = all, lower switch = single), then setting any threshold to a
value that will always be exceeded, will remove it from the decision process. For example, you could set the RPM
threshold to 0 RPM and RPM would never play a role in the switching.

Hybrid Transition Blending

When the decision to switch from MAF to SD mode (or vice versa) is made, the final load value will not abruptly
switch from the load calculation of one mode to the other. Instead, over a short period of time, the final load will be
calculated as a blend between the on-going load calculation in the old mode and the on-going load calculation in
the new mode, with the bias increasing from old to new as the ramping process proceeds. This allows for a

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 31
smoother load transition between modes (and therefore smoother fueling/timing transition), especially where the
SD-based load and MAF sensor-based load calculations may be somewhat farther apart from one another.

The ramping is determined by the ramping multiplier. This multiplier (RM below), can range from 0 to 1 and deter-
mines the blending of the old and new load values as follows:
Final Load =( SD Load ∗ RM )+ ( MAF ∗(1 −RM ))

In the above equation, if the ramping multiplier is 0, then the final load is entirely based on MAF. If is it 1.0, then
the final load is entirely based on SD load. Any value in-between is a blend of both MAF and SD load with the
bias dependent on the value.

When the decision to move from MAF to SD mode is made, the ramping multiplier, which would begin at 0 in this
case, would be incremented by a specific value (see tables below). The ramping multiplier would continue to be
incremented as long as the decision to switch to SD remained and, over a short period of time, the ramping multi-
plier would reach 1.0 and SD load would entirely be used. The opposite is the case in moving from SD to MAF
mode. The ramping multiplier would begin at 1.0 and would be decremented by a specific value as long as the de-
cision to switch to MAF remained and, over a short period of time, the ramping multiplier would reach 0 and MAF
would entirely be used.

The speed at which the ramping process occurs (i.e. the ramping multiplier adder) is dependent on the following
tables:

As you can see from the tables above, the speed at which the switch occurs can be tuned separately for MAF to
SD and SD to MAF. For MAF to SD, higher values in this table increase the speed of the blended transition, while
lower values decrease the speed. For SD to MAF, the opposite is true.

Tuning Hybrid Mode – Post-Tune Recommendations

If the hybrid tune was completed on a dyno, it is important to drive the car under conditions it is likely to see dur-
ing normal operation to verify the safety of the tune and that the hybrid switching occurs as intended. Additionally,
for the SD part of the tune, the recommendations outlined in the “Tuning SD –Post-Tune Recommendations” sec-
tion earlier in this document should also be considered.

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 32
Tuning Miscellaneous New Features
New logic was added to the COBB SD 2.0L ECU to allow for additional tuning functionality.

Gear Position Determination Expanded to Cover 6-Speed Transmissions and New Gear Speed Monitor The
factory ECU estimates the current manual transmission gear position based on vehicle speed, RPM and the
thresholds dictated by the “Gear Determination Min/Max” table (found in the “Miscellaneous Tables” group). As
long as the clutch is out and the vehicle is moving, the ECU will be able to reasonably estimate the current gear.
However, the factory 2.0L ECU is only set up to handle 5 gears. For the COBB SD 2.0L ECU, this has been modi-
fied to be able to handle 6 gears for possible 6-speed transmission swaps. This swap (or any transmission gear
ratio change) requires that the “Gear Determination Min/Max” table be tuned appropriately so that the correct gear
can be estimated. For the Cobb SD 2.0L ECU, we have simplified this process with the addition of a new “Gear
Speed” monitor. This will allow you to view/log the currently calculated rev/mile value that the ECU uses to com-
pare to the “Gear Determination Min/Max” table thresholds to estimate the current gear. This table's thresholds
are set up as follows:

1st gear estimated if gear speed is greater than or equal to table value #1

2nd gear estimated if gear speed is less than table value #1 and greater than or equal to table value #2

3rd gear estimated if gear speed is less than table value #2 and greater than or equal to table value #3

4th gear estimated if gear speed is less than table value #3 and greater than or equal to table value #4

5th gear estimated if gear speed is less than table value #4 and greater than or equal to table value #5

6th gear estimated if gear speed is less than table value #5

The gear speed value will tend to vary a bit within a certain range for the current gear. As such, you will want to
set up the thresholds in the Gear Determination table so that the average gear speed value you see when logging
for a given gear will be a midpoint between the appropriate two thresholds in the table. For example, if 1st gear
shows a logged gear speed value hovering around 10,000 and 2nd gear shows a logged value around 8,000,
then 9,000 would be reasonable for the first value in the table. This would avoid the incorrect gear being esti-
mated due to noise in the gear speed value. For example, if you set table value #1 too close to the observed gear
speed of 2nd (say 8,200 in the example above), then the ECU would sometimes determine you were in 1st gear
when you were actually in 2nd.

When doing a transmission swap from a USDM 2.5L car, you may decide to copy the values from that ECU's
Gear Determination table into your SD 2.0L tune. While this is a good starting point, we have found that further
tweaking of the table may be necessary to avoid incorrect gear estimation. Always verify your table threshold val-
ues by logging the gear speed monitor for each gear (across a broad RPM range where possible). When you
have finished tweaking the table, verify that the “Gear Position” monitor shows the correct estimated gear (with
Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 33
clutch out and vehicle moving) for every forward gear in the transmission (also across a broad RPM range where
possible). Note: A value of “7” will be shown for gear position when the transmission is in neutral (this is different
than the non-SD ECU which will show a value of “0” in this case).

Per Gear Boost Targets and Wastegate Compensation

New real-time tunable per gear boost targets and wastegate compensation tables have been added to the COBB
SD 2.0L ECU :

“Boost Targets Compensation (Per Gear)(1st,2nd,3rd,4th,5th,6th)” -> Each cell in this table represents a percent-
age compensation to the current boost target (absolute pressure) for a given gear. This will allow you to increase
or decrease the compensation to the boost target individually for each gear.

“Wastegate Duty Cycles Compensation (Per Gear)(1st,2nd,3rd,4th,5th,6th)” -> Each cell in this table represents a
percentage (relative) compensation to the current “Wastegate Duty Cycles (High)” value for a given gear. This will
allow you to increase or decrease the compensation to the maximum wastegate duty individually for each gear.
This can, for example, allow you to have more aggressive boost control (for lower gears) while preventing over-
boost in higher gears.

Injector Latency Adder

A new real-time tunable table named “Fuel Injector Latency Adder (Small IPW)” (found in the “Fuel Tables” -> “In-
jectors” group) has been added to the COBB SD 2.0L ECU. This is a compensation to the current injector latency
based on the current injector pulse width. This can be useful to tune for the non-linearity of larger injectors at low
injector pulse widths to improve idle/low load behavior.

Neutral Position Switch Invert

Some manual transmission swaps require that the neutral position switch determination by the ECU be inverted.
The COBB SD 2.0L allows you to do this via the “Neutral Position Switch Invert (MT)” toggle in the “Edit” -> “Ad-
vanced Parameters” -> “Toggles (Base)” menu. Simply uncheck this toggle, save map, and reflash to car if you
wish the neutral position switch to be inverted. To verify operation, live connect to the car with the AccessTUNER
Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 34
software and bring up the “Neutral Position Switch” monitor in the dashboard. This monitor should show “ON”
when the transmission is in neutral and “OFF” when the transmission is in gear.

Boost Targets Limit Raised

The factory 2.0L ECU has an inherent maximum limit in the boost targets table of 24.75 psig (at sea level). That
is, even though the factory 2.0L ECU can monitor and target effectively any level of boost (given a MAP sensor
with the appropriate range), the boost targets table is still limited to a max of 24.75 psig. The COBB SD 2.0L ECU
raises this limit to 49 psig. To use, simply enter any boost target value up to 49 psig in the Boost Targets table.

Airflow and Load Limits Raised

Load and airflow limits inherent to the factory 2.0L ECU have been significantly raised. No longer is there a need
to trick the ECU by manipulating airflow/load references and the fuel injector scale in order to overcome the limits
in the tune. Simply tune with actual real-world values up to 10.0 g/rev of load (SD or MAF mode) and up to 655
g/s of airflow (MAF mode). Additionally, SD load is directly calculated and is subject only to the 10.0 g/rev load
limit (regardless of airflow).

More Flexible MAP Calibration Offset

The factory 2.0L ECU only allows for effectively a negative offset for MAP sensor calibration. This presents a
problem when using aftermarket MAP sensors that require a positive offset. The COBB SD 2.0L ECU overcomes
this issue by allowing either a positive or negative offset to be entered. This gives you the ability to tune for any
aftermarket MAP sensor that is compatible with the car.

Tunable IAT Calibration Axis and Latency Axis

The factory 2.0L ECU does not allow for editing the individual cells that make up the axis of the “Intake Temp.
Sensor Calibration” and the “Fuel Injector Latency” tables. This prevents you from modifying the intake temp sen-
sor voltage and battery voltage values (respectively) in these tables. The COBB SD 2.0L ECU overcomes this lim-
itation and allows you to edit any cell in either of these tables.

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 35
 MONITORING

How to Monitor the Tune

In addition to the monitors already included for the non-SD ECUs, the SD ECUs add additional monitors specific
to the COBB SD feature. These will be useful in the tuning process as well as verifying/monitoring the tune once it
is complete.

CalculatedLoad(MAFSensorBased) -> This is the calculated load based on the MAF sensor with all compen-
sations applied. This value will always be calculated regardless of the current mode.

MassAirflow(MAFCalibration) -> This is the mass airflow based on the MAF sensor (the output from the “MAF
Calibration” table with all compensations applied). This value will always be calculated regardless of the current
mode.

SDEngineDisplacement -> This is the engine displacement value in the map that is reflashed to the car.

SDHybridThresh.BitField ->This monitor will indicate which of the hybrid thresholds (RPM, MAP, TPS, MAFv)
have been exceeded and which have not. This is useful to verify that the hybrid thresholds you’ve tuned are
achieving the desired effect. This value is only calculated in hybrid mode. For the AccessTUNER software, this
monitor will be shown as a series of letters indicating which thresholds have been exceeded as follows:

MAF voltage = “V”

TPS = “T”

MAP = “P”

RPM = “R”

For example, if TPS, RPM, and MAP are exceeded, but MAFv is not, then “TPR” will be displayed. If none of the
thresholds are exceeded, then the monitor will be blank.

For the AccessPORT, this monitor is displayed as a bit-encoded byte. To determine the thresholds exceeded,
you’ll want to first convert the decimal value shown to a binary number. The status of each threshold is given as
follows:

(0 = not exceeded, 1 = exceeded)

Bit 0 = Mass Airflow Voltage

Bit 1 = Throttle

Bit 2 = Manifold Absolute Pressure (MAP)

Bit 3 = RPM

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 36
(Bit 0 is the last digit in the binary representation of the number, bit 1 is the second to last digit, and so on)

For example, a logged value of 9 (binary = 1001), would indicate that RPM and MAF voltage thresholds have
been exceeded. A value of 0 means none of the thresholds have been exceeded.

SDLoad(Post-Comp) -> This is the final SD load after the IAT, ECT, and barometric compensations (as de-
scribed in this document) have been applied to the base SD reference load. This value is always calculated re-
gardless of the current mode.

SDLoad(Pre-Comp) -> This is the SD reference load which is calculated based on the ideal gas law using the
reference temperature of 86 deg. F. That is, this is the SD load before the IAT, ECT, and barometric compensa-
tions have been applied. This value is always calculated regardless of the current mode.

SDMode(Load) -> This monitor shows which SD mode is being used to determine the final load. This will tell you
if SD-based or MAF sensor-based load is being used or a blend of both during a hybrid transition. In the Ac-
cessTUNER software, this monitor will display/log one of the following strings (AccessPORT value show in brack-
ets):

• “MAF” [0] – Mode is MAF and MAF sensor-based load is being used.

• “SD” [1] – Mode is SD and SD-based load is being used.

• “MAF-H” [2] – Mode is hybrid and MAF sensor-based load is being used.

• “SD-H” [3] – Mode is hybrid and SD-based load is being used.

• “BLEND” [4] – Mode is hybrid and transition is taking place where SD-based load and MAF sensor-based
load are blended.

• “MAF-E” OR “ERR” [5 or greater] – An error has occurred and the mode has defaulted to MAF. Contact
COBB customer service if this appears.

SDVE(Commanded) -> This is the commanded volumetric efficiency (%) as determined by your “SD Volumetric
Efficiency” table. This value is always calculated regardless of the current mode.

SDVEEstimated(MAF) -> If the vehicle is still equipped with a MAF sensor that is installed and calibrated cor-
rectly, this monitor allows for an estimate of volumetric efficiency based on current MAF sensor-based load, MAP,
IAT and the reflashed engine displacement value. Keep in mind that specific conditions can skew some of the in-
puts, resulting in a VE estimate that is not reliable (see “Tuning SD – Starting Values for VE Table” section for de-
tails). This value is always calculated regardless of mode.

SD/MAF HybridRatio -> This is the ramping multiplier (RM) which determines the blend of MAF sensor-based
and SD-based load when a transition is taking place between modes (when hybrid mode is active). This multiplier
determines the load blend as follows:

Final Load =( SD Load ∗ RM )+ ( MAF ∗(1 −RM ))

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 37
For example, if the RM is 1.0, then only SD load is being used. If it is 0, then only the MAF sensor-based load is
being used. Any value that falls between 0 and 1 means that a blend of both load calculations is being used (as
determined by the equation above). This value is only calculated in hybrid mode.

SD/MAF Load Smoothing Factor ->This is the current load smoothing factor being used by the ECU, which is
dependent on the current load mode. Details about the load smoothing factor can be found in the “Tuning SD –
Miscellaneous Tables” section earlier in this document.

SD/MAF Pre-Final Airflow -> This is the final airflow to be used by the ECU. This could be MAF sensor-based
airflow, SD-based airflow, or a blend of both (when ramping in hybrid mode). In SD mode, this value will be the
back-calculated airflow from the “SD Load (Post-Comp)” value and RPM. In MAF mode, this value will be the
same as the “Mass Airflow (MAF Calibration)” monitor. When transitioning between SD and MAF (or vice versa) in
hybrid mode, this value will show the blended value during that transition.

SD/MAF Pre-FinalLoad -> This is the final load to be used by the ECU. This could be MAF sensor-based load,
SD-based load, or a blend of both (when ramping in hybrid mode). In SD mode, this value will be the same as the
“SD Load (Post-Comp)” monitor. In MAF mode, this value will be the same as the “Calculated Load (MAF Sensor
Based)” monitor. When transitioning between SD and MAF (or vice versa) in hybrid mode, this value will show the
blended value during that transition.

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 38
 SD REAL-TIME

SD Real-Time Tuning
Some changes were made to real-time tuning for the SD ECUs as compared to the non-SD ECUs:

Row/Column Table Data Removed from Real-Time

Because tuning the volumetric efficiency table in real-time is an enormous benefit and because the space avail-
able for real-time is limited, the existing real-time set-up had to be modified to accommodate this large table. This
included removing the row/column table data (where applicable) from real-time. For example, you will not be able
to real-time tune calculated load or RPM for the “Primary Open Loop Fueling” table (i.e. only the fueling targets
are real-time tunable whereas the load and RPM axes have to be changed by reflash). This is also the case with
any new SD tables (where applicable). Additionally, the “Turbo Dynamics Burst” and “Turbo Dynamics Continu-
ous” tables were removed entirely from real-time (these can still be tuned via reflash).

Tables Added to Real-Time

To improve real-time tuning functionality, some new real-time tables were added where possible. These include
some original factory tables and some new tables added due to new features of the COBB SD ECU. The follow-
ing are the new tables added to the real-time assortment:
• “Boost Targets Compensation (Per Gear)(1st,2nd,3rd,4th,5th,6th)”
• “Dynamic Advance Max” (table has been resized)
• “Fuel Injector Latency”
• “Fuel Injector Latency Added (Small IPW)”
• “Tip-in Enrichment”
• “Wastegate Duty Cycles Compensation (Per Gear)(1st,2nd,3rd,4th,5th,6th)”

Real-Time Tunable SD Tables

Most of the new tables added for the SD ECU are real-time tunable. See the “Realtime Tables” group in the Ac-
cessTUNER software for a complete listing of all real-time tables.

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 39
 SD MATH
SD Load Math
This section describes the math behind the COBB SD load calculation which is based on the ideal gas law. It is
not necessary to understand this math in order to tune SD, but some may find it interesting.

Ideal Gas Law – Introduction

The ideal gas law is an equation that governs the relationship between the pressure, volume, amount and temper-
ature of an “ideal” gas:

P = Pressure

V = Volume

n = Number of Moles (i.e. amount)

R = Gas Constant

T = Temperature
PV =nRT

We want to solve for the amount (n) of gas:

PV
n=
RT

Our “ideal” gas is quite simply the air that is being ingested by the engine. For our purposes, we want to know the
mass of air entering the engine. We can determine the mass of air by using a constant that dictates the mass (in
grams) per mole of air (i.e. the molar mass of air, which we’ll call MMA):

Mass of Air ( g )= n(moles )∗MMA( g / mole)

Because n = PV/RT, we can express this as:

PV
Mass of Air ( g )=( )∗MMA
RT

Ideal Gas Law – Real-World Inputs

Given the above equation, we need input data for pressure (P), volume (V), and temperature (T). Given our en-
gine scenario, these inputs would be determined as follows:

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 40
P = Pressure = Manifold absolute pressure (MAP) as determined by the MAP sensor. Units dependent on gas
constant (R) used (see below)

V = Volume = Displacement of the engine (in liters).

T = Temperature = Cylinder charge temperature in Kelvin as estimated by the intake air temperature (IAT) sensor.
Temperature in Kelvin can be determined as follows: IAT Celsius + 273.15

Let’s rename the symbols used for these inputs so they are more appropriate for our engine example:

MAP = P

DISP = V

IAT = T

The constants R and MMA are determined as follows:

R = ideal gas constant. Dependent on the MAP units used -> R = 1.205912 (MAP in psi), R = 0.08314472 (MAP
in bar), R = 8.314472 (MAP in kPa), R = 62.363669 (MAP in mmHg)

MMA = Molar Mass of Air Constant =28.97 g/mole

MAP ∗DISP∗MMA
Estimated Mass Air ( g )=
R∗ IAT

Ideal Gas Law – Volumetric Efficiency

The above equation is only valid for our engine example if volumetric efficiency is always 100%, which is obvi-
ously not the case. We must therefore add VE as a correction factor.

MAP ∗VE∗DISP∗MMA
Estimated Mass Air ( g )=
R∗IAT

Ideal Gas Law – Mass Airflow

To determine the air entering the engine per unit of time, we add RPM as an input. Because the crankshaft ro-
tates 720 degrees (i.e. two revolutions) for a full stroke, the number of times air is entering the cylinders per sec-
ond is given by:

RPM
2∗60

We combine the above with our mass air equation to get, ultimately, an estimated mass air in grams/second:

RPM ∗MAP∗VE ∗DISP∗MMA

Estimated Mass Airflow ( g / sec)=
120∗R∗ IAT
Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 41
Ideal Gas Law – Engine Load
The factory ECU deals in terms of load in grams per crankshaft revolution which is calculated from MAF sen-
sor-based airflow as follows:
Mass Airflow∗60
Calculated Load ( g / rev)=
RPM

From the perspective of the ideal gas law, we are able to calculate load directly without first calculating mass air-
flow. If you replace the Mass Airflow input in the equation above with our Estimated Mass Airflow equation, it sim-
plifies to the following:

MAP∗VE∗DISP∗MMA
Estimated Calculated Load ( g / rev )=
2∗R∗IAT

Ideal Gas Law – SD Reference Load

The non-linear charge temp correction is inherent to the equation above in determining the estimated load. How-
ever, it would be useful to be able to tweak the charge temp correction because our IAT sensor input may not al-
ways be exactly representative of actual charge temp. This is highly dependent on the placement of the IAT sen-
sor for a given car.

To allow for tweaking of the charge temp correction, the SD ECU first calculates a reference SD load using a
constant for charge temp (30 deg. C, 86 deg. F). Then the IAT table correction is applied to the reference value.
The default values of the IAT table are set-up with the ideal gas law in mind given our reference temperature.

Consider the following example:

MAP = 1277.359 mmHg (native units of the ECU, which is the same as 24.7psia)

VE = 0.8 (i.e.80%)

DISP = 2.0 liters (configurable table value in software)

MMA = 28.97 g/mol

IAT = 303.15 Kelvin constant (our reference temp of 30 C)

R = 62.363669 (given the mmHg native ECU units for MAP)

For our reference calculation, all the constants can be rolled up into a single constant as follows:

MMA
constant =
2∗R∗ IAT

constant =
28.97 =.000766177
2∗62.363669∗303.15

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 42
So, the ECU calculates the reference load as follows:

Estimated Calculated Load ( g / rev )=VE∗MAP∗ DISP∗constant

Estimated Calculated Load ( g / rev )=0.8∗1277.359∗2.0∗0.000766177=1.57 g / rev

SD Final Load
We have calculated the SD reference load above. However, this load is only valid at our given reference temp of
86 degrees F. We must apply the IAT compensation which, with the default values in the tables, follows the ideal
gas law the same as if we had originally plugged the current IAT value into the estimated calculated load equation
that was first described in this section. Let’s assume that the current IAT is 122 degrees F. Looking at our IAT
compensation table:

We see that the table calls for -6.19% correction at 122F IAT. We can convert -6.19% to a multiplier of 0.9381
((100+x)/100). From our reference load example, we calculated a reference load of 1.57 g/rev. We apply the IAT
correction from our table as follows:
SD load =SD reference load∗ IAT correction ( multiplier)

SD load =1.57 g / rev∗0.9381 =1.47 g / sec

If we were to plug in the IAT of 122F (50C) into the original equation given our current example:

MAP = 1277.359 mmHg (native units of the ECU, which is 24.7psia)

VE = 0.8 (i.e.80%)

DISP = 2.0 liters (configurable by tuner)

MMA = 28.97 g/mol

IAT = 323.15 Kelvin (example IAT of 122F or 50C)

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 43
R = 62.363669 (given the mmHg native ECU units)
MAP∗VE∗DISP∗MMA
Estimated Calculated Load ( g / rev )=
2∗R∗IAT

1277.359∗0.8∗2.0∗28.97
Estimated Calculated Load ( g / rev )= =1.47 g / rev
2∗62.363669∗323.15

We end up with the same load calculation either way, but our “external” IAT compensation table allows us to
tweak it in cases where the IAT input may not be representative of the actual charge temp.

Our final SD load is determined after the ECT and barometric compensations have been applied.

Estimated VE Calculation
The SD ECU has the “SD VE Estimated (MAF)” monitor which allows you to determine an estimate of VE if you
have a properly installed, functioning, and calibrated MAF sensor in the car. The math for estimating VE is simply
solving for VE in the estimated calculated load equation given above. We are using the calculated load as deter-
mined by the MAF sensor, “MAF Calibration” table, and RPM as an input to our estimated VE calculation:

Load ∗2∗ R∗IAT

VE =
MAP∗ DISP∗MMA

Let’s take the previous example and solve for VE given our previous estimated load of 1.47 g/rev:

1.47∗2∗62.363669∗323.15
VE = =0.80( or 80%)
1277.359∗2.0∗28.97

As you can see, we end up with the same VE calculation as used in the original estimated calculated load exam-
ple.

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 44
 APPENDIX

Appendix – Aftermarket IAT Sensor Install

The following describes how to install an aftermarket intake air temp (IAT) sensor by modifying the factory wiring
harness. This information is applicable to the 2002-2012 WRX and 2004-2012 STi:
• Locate and unplug your factory MAF sensor wiring harness.
• For GD Subarus (2002-2007 WRX and 2004-2007 STi), cut the Brown and Green/Red wires of the factory
wiring harness so that you have a suitable length.

• For GR Subarus (2008-2012 WRX and STi), cut the Yellow/Red and Yellow/Blue wires of the factory
wiring harness so that you have a suitable length.

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 45
 • Strip and solder together the two wires of the aftermarket IAT sensor pigtail and the engine wiring harness.

• Plug the MAF sensor harness back in (if MAF sensor is not being removed).
• Modify the “Intake Temp. Sensor Calibration” table in the AccessTUNER software (under “Sensor Calibra-
tions” group) so that the appropriate calibration is achieved for your aftermarket IAT sensor. Note: Recom-
mended calibrations for some aftermarket IAT sensors can be found in the “Tuning SD – Additional Top-
ics” section earlier in this document.
• Save your map and reflash to car.

Copyright © 2012 COBB Tuning Products LLC. All Rights Reserved | www.cobbtuning.com 46
 • Before tuning, verify that the new IAT sensor is reading properly by monitoring Intake Temp via the Ac-
cessTUNER software or the AccessPORT.

Copyright© 2012 COBB Tuning Products LLC. All Rights Reserved 1 www.cobbtuning.com 47