The 991.2 Carrera Turbo and Turbo S continue to occupy the halo car position in the 911 lineup.  Just like prior “Turbo” 911s the new 991.2 utilize a 3.8 liter twin turbo flat 6 motor however unlike the 3.0 liter turbo 991.2 Carrera, the 3.8 liter turbo utilize variable geometry turbine to control boost pressure.  Because these cars produce vast quantities of torque, they are available only in AWD configuration.

Generally speaking, the entire range of 718 and 911 within the 991.2 generation utilize very similar control strategies and share many of the same tables and control logic strategies.  Where possible, similarities across all three types or vehicles 718, 3.0 liter 911, and 3.8 liter 911will be noted within this document.  This tuning guide focuses basic tuning strategies and defines tables for each major tuning category, such as load / boost control, fueling, and ignition timing. Over 500 tables exist within the tuning software.

991.2 Models

The Above chart outline some of the key differences in components between the 991.2 911 models. The drivetrain in the 3.0 liter cars is identical outside of increasingly larger turbochargers incrementing up from base to S and GTS.

Looking at the differences shown above we can see how the differences impact power and torque across the different models. The difference in turbocharger size can easily be shown by plotting them all next to eachother.

Load Based Engine Control

991.2 use a speed density-based load control strategy.  The ECU requests torque output from the motor based upon the accelerator pedal position.  This torque request is then processed through a volumetric efficiency model that determines target manifold pressure (boost control), air mass, fuel mass, and ignition timing.   Each of these systems has a variety of complex subsystems but the simple point is that these vehicles are calibrated as a matter of requested torque and all other processes are dictated by this.

Functional Architecture of a COBB OTS Calibration

As with any highly complex modern engine control unit there are many things understood and far more that are less well understood.  COBB calibrations are purposefully designed to create power and driving character improvements while maintaining as much of the stock architecture as is possible.  Utilizing this approach, intrinsic safety and emissions systems are preserved.  This guide is structured around explaining the rationale for changes in OTS calibrations and additionally around both the order of operations for a typical calibration.  

The first step in tuning a Porsche 991.2 Carrera is choosing a COBB Tuning Off-The-Shelf (OTS) calibration that most closely matches the mechanical components and modifications of the vehicle to be tuned

Stage1 calibrations are designed for vehicles with no aftermarket parts at all. COBB does not currently offer Stage 2 support for the 718 chassis.  We have found that additional exhaust components do not markedly improve the pumping efficiency of the 718 platform.  However on the 991.2 vehicles we've found taht exhaust modifications can make a difference.  For those inclined to utilize exhaust modifications, your “stage 2” impacts the pumping efficiency of the motor and as such it will have an impact on all major aspects of tuning (boost, fuel, and ignition).

Another important thing to keep in mind in the early stages of the tune is what fuel the car is currently on, as well as what fuel the customer intends to run the vehicle on.  There is little sense in tuning for 93 octane if the customer only has 91 octane. COBB Tuning offers calibrations for two different fuels 93 octane (98 RON), 91 octane (95 RON). Higher octane ratings indicate higher quality fuel that is less volatile and typically allow for higher cylinder pressures(temperature) before pre-ignition. This difference in fuel will determine how the car is tuned. Higher octane fuels support more ignition timing, higher boost levels, and leaner air-to-fuel mixtures compared to lower octane. Using a map designed for high octane with low octane fuels can result in damage to the motor.

It is normal to consider the type of intake the car is running.  These vehicle utilize a speed density based calculated mass airflow, but does not use actual MAF (Mass Air Flow) sensors. Changing intakes likely will not require changes to the calibration. The 991.2 Carrera does use the airflow calculations to limit boost pressure, so making changes to calculated airflow tables is critical to making the car run correctly. 

Torque Control System

After selecting the most appropriate initial calibration, prepare to test and refine the calibration on a chassis dynamometer. When creating a custom tune, it is best to begin testing under low load conditions by lowering the target torque tables.

In these vehicles the relationship between these tables is very direct in terms of controlling the difference between the torque command and output. Here you can see that the shape of the torque curve as measured by a chassis dynamometer very closely mimics the run using that map (Shown in blue)

As mentioned earlier, all 991.2 (3.0 and 3.8 liter 911) and 718 (2.0 and 2.5 liter Cayman /Boxster) are controlled via a torque request dictated by accelerator pedal position.  The requested torque tables are located in the “Torque Tables” folder.  Comparing an OTS calibration with stock highlights main tables used to add and subtract torque and power.  Within this folder there are actually very few meaningful tables that will need to be changed in the final calibration.  Most of these tables represent torque limits of two types.  The first type is an overall torque limits and the second type are those limits related to traction control and left foot braking, and therefore are moved out of the way (set to maximum values).

Overall Torque Limits

• Actual gear for enabling torque limitation

• Initialization torque for vehicle speed profile

• Lower vehicle speed threshold for enabling torque limitation

• Max engine torque depends on pedal value and brake pressure / brake torque limitation

• Maximum Reference Indicated Torque

• Minimum torque for vehicle speed-based torque limitation

• Upper vehicle speed threshold for enabling torque limitation

• Torque limit for TCU based on variant coding error

Just 4 tables are used to dictate increased or decreased torque and power in most 991.2 and 718 vehicles.  These tables are the MAIN tables you can use to manipulate the power curve

• Maximum torque at clutch for overboost in normal mode 1

• Maximum torque at clutch for overboost in normal mode 2

• Maximum torque limitation at clutch for overboost in sport mode 1

• Maximum torque limitation at clutch for overboost in sport mode 2

These tables are separated into Normal and Sport mode.  Normal and Sport mode are accessed by the driver through a selector on the steering wheel.  COBB OTS maps have these modes set to the same values across all maps so the car makes the SAME power in Normal and Sport modes.  However, it is entirely possible to set normal and sport modes to different power output.  Some customers for example may want a low power tune for valet or use by an unskilled driver.  While others may want separate street and track modes.

Normal and sport modes are separated into “1” and “2”.  At this time differences are unknown and tables 1 and 2 should be set to the same values.  

Torque Limits

Torque requests are not without an upper bounds.  Finding and raising torque limits comprises a large component of successfully calibrating 718 and 991.2 vehicles. 

Some torque limits are simple / static and are related simply setting an upper bound for torque.  Many of these limits are simply moved out of the way in COBB OTS maps as previously mentioned in “Torque Request” above.  Other torque limits are highly dynamic and related to current conditions. Conditions that will create a torque limit include. 

• Exhaust Gas Temperature (modeled)

• Fuel limits (injection window, Fuel pump set points) (measured)

• Detonation and Octane learning (calculated / measured)

• Coolant Temperature (measured)

• Oil temperature (measured) 

• Charge Air Temperature (measured)

• Intake Air Temperature (measured)

• Engine Compartment Temperature (modeled)

• Turbo Rotational Speed (modeled) 

• Fuel Injection mode 

• Barometric Pressure (measured)

The vast majority of these static and dynamic torque limits are well out of the way for a stock turbo configuration supported by COBB OTS mapping. Details regarding each of these dynamic torque limiting scenarios are outlined in their respective specific subsystem folder below. 

Simple / Static torque limits are contained within the “Torque Control” tuning folder.  A few specific Torque limiting tables are here and worth describing.  

Torque Limit – Intake air temperature.  

This table is tuned very conservatively from the factory and will serve as an active torque limit when the car is run repeatedly or driven under anything but perfect conditions.  The Z data here isa  torque limit and is defined by engine speed and Air Temperature.  This is intake air temperature as measured at the manifold.  It is not unusual for intake air temperature to be actively limited in a single dyno pull with these values set to factory.  It is recommended to allow this function to limit torque but at slightly higher air temperature.  This will make the car more consistent run to run.  

Intake Air Temperature Base Torque limit 

In this example (991.2 Carrera) you can see two power curves.  The dyno plot shown in red was done with factory settings for the Air Temp based Torque limit.  The Blue Dyno plot shows the results with the same car under the same conditions but with the Air Temp based torque limit increased to allow slightly higher temps before intervention. This can be observed in the second plot which graphs Engine Torque Limitation.  The green line if the higher allowable torque generated by increasing allowable torque related to air temperature.  This is just one example of a dynamically generated torque limit.  In this case the limit was logically

The maintenance of torque entering and exiting dynamic limiting conditions is controlled through a PID system.  The I and P terms related to increased and decreased torque are held in Torque Control under the Torque Control PID subfolder.

Overboost Function

The factory ECU contains logic permitting “overboost”.  This logic is turned ON and OFF with a simple 1 (on) or 0 (off).  COBB OTS calibrations follow the OE calibration and DO NOT use this function.  Use of this function creates a large torque request at lower RPM that is outside of ideal for these vehicles.  

Torque request and offsets for overboost functions

The additional torque requested during overboost conditions is managed by a separate set of torque request tables within this folder.  

• Maximum torque limitation for overboost in normal mode

• Maximum torque limitation for overboost in normal mode 1

• Maximum torque limitation for overboost in sport mode

• Maximum torque limitation for overboost in sport mode 1

These torque requests are used as adder / limiter to standard torque request.  In our experience these tables are ONLY utilized if the Switch for disabling the calculation of torque request for overboost is set to ON (1). 

An alternate form of overboost is used conditionally in Sport mode.  This additional torque is requested only in sport mode and only conditionally for a period of time. This additional requested torque is defined by ambient pressure and engine speed.

• Torque offset for overboost 1

• Torque offset for overboost 1

Torque Refence Setpoints

          The torque control system in all 991.2 and 718 is calibrated around a set of internal standards that relate Torque to airflow and Airflow to Torque. Two sets of 64 tables serve to establish this relationship.  The Torque reference tables are separated into 4 sets of 16 tables.  The Z data for the table is reference torque and the axis are engine speed and airflow. 

These tables are set up with three different paramters that denote different states for various active cam positions.  This is represented by the three digits at the end of the name and are divided into three sections, with four possible values (0-3) for each position.

All setpoint values are weighted by the actual camshaft degrees.  For example, intake cam degrees of 10 and exhaust cam degrees of 10 would interpolate 50% between intake cam index 0 and 1.  Exhaust cam position is similarly weighted across index tables.  Exhaust camshaft of 35 degrees would use a weighted value of 50% from index 2 and 50% from index 3.

For all 3.0 liter 991.2 Carrera and 2.0 liter 718 only the first 32 of these 64 tables is utilized because these cars do not have exhaust valve lift. (Index 1 is active for 0 and 1 only (all low lift and Intake low lift exhaust standard lift).  In contrast, 2.5 liter 718 and 3.8 liter 991.2 variants ALL 64 tables are used as these cars have high and low lift intake AND exhaust as well as variable camshaft timing. 

For stock turbo-based calibrations these reference tables can be left completely stock.  These torque to airflow reference points are intrinsic to the vehicle performance.  If these tables are extended or changed you MUST properly calibrate the Airflow Reference tables that are calibrated as the inverse of these tables - Airflow as Z data and Torque as X axis. 

Ignition Timing

The ignition control strategy in the Porsche 991.2 Carrera is very dynamic and has many contributing variables to determine the overall ignition timing value.  There are a total of 64 main ignition tables currently in software that are grouped into 4 folders of 16. 

These tables are set up with three different paramters that denote different states for various active cam positions.  This is represented by the three digits at the end of the name and are divided into three sections, with four possible values (0-3) for each position.

All setpoint values are weighted by the actual camshaft degrees.  For example, intake cam degrees of 10 and exhaust cam degrees of 10 would interpolate 50% between intake cam index 0 and 1.  Exhaust cam position is similarly weighted across index tables.  Exhaust camshaft of 35 degrees would use a weighted value of 50% from index 2 and 50% from index 3.

For all 3.0 liter 991.2 Carrera and 2.0 liter 718 only the first 32 of these 64 tables is utilized because these cars do not have exhaust valve lift. (Index 1 is active for 0 and 1 only (all low lift and Intake low lift exhaust standard lift).  In contrast, 2.5 liter 718 and 3.8 liter 991.2 variants ALL 64 tables are used as these cars have high and low lift intake AND exhaust as well as variable camshaft timing. 

For stock turbo-based calibrations these reference tables can be left completely stock.  These torque to airflow reference points are intrinsic to the vehicle performance.  If these tables are extended or changed you MUST properly calibrate the Airflow Reference tables that are calibrated as the inverse of these tables - Airflow as Z data and Torque as X axis. 

Reference Ignition

The reference ignition tables represent the car running at the ideal timing for the engine and is the timing at which the reference torque levels are assumed to exist.  Under most conditions these tables will not need to be touched.

Knock Control System

The knock control strategy on the Porsche 991.2 Carrera is very complicated and uses individual cylinder knock control to make changes to the ignition timing at all times. The system is very sensitive and thus almost always has some sort of feedback. You can monitor knock retard and ignition correction in each cylinder. This is the best way to check for detonation, and to ensure a safe running vehicle. The closer to 0 the better, but if it sways into the negatives (-2 and below) this is the car registering detonation. Since the car is very dynamic it is made to be sensitive. If you are getting values past -6 you will want to try and lower ignition timing, or add more fuel to try and see if you can bring the values back closer to 0. Running a race fuel will also help in getting the knock control values closer to zero.

Porsche individual cylinder Knock feedback systems.  Ignition timing is adjusted per cylinder in response to engine noise consistent with detonation.  This graph presents overall ignition timing (blue) and individual cylinder ignition adjustments (decrementing negative values) during a full load pull from 2000 RPM to redline. 

Adaptive Ignition Timing and Octane Learning 

The ignition control systems in 991.2 and 718 utilize a mechanism to learn fuel octane and adapt overall ignition strategy.  This is a first for Porsche 911 as all earlier cars responded to fuel quality with highly sensitive ignition on an immediate basis but did not store this information in as a means predict fuel quality.  991.2 remember fuel quality through a not yet fully understood mechanism that translates immediate knock responses into an octane learning in the form of a number from 0 to 1 – Ignition RON.  Ignition RON is used to calculate ignition timing offsets and, in extreme cases, reduce requested torque.  In the OE calibrations this value is not reset.  This means that a bad tank of fuel will be remembered ever after the fuel quality is increased.  The core parts of this system remain intact within COBB OTS calibrations (ignition adjustments and the potential to reduce requested torque). COBB OTS maps however utilize factory logic to RESET the learned “Ignition RON” through each key cycle.  This factory logic is located in Ignition RON reset.  A value of “1” enables the key cycle reset while “0” maintains this learned octane value through key cycles.  You can determine if the ignition RON has changed as this is a log parameter.  0 is high octane.  1 is low. 

The adaptive octane learning uses two thresholds to determine whether to move RON value, to determine the direction of the move (toward 1 or 0) Max ignition for RON adjustment is the threshold for an increase in RON if ignition corrections is above this value. Minimum ignition for RON adjustment – if ignition correction is lower than this threshold the learned RON value will move toward 0 (higher octane).  This means that the ECU is self-tuning to keep ignition correction between these Minimum and Maximum values. 

The learned octane (RON value impacts overall ignition timing through a second table Table for ignition angle correction due to global adaptive circuit. 

Learned Octane based Ignition reduction = (RON, 0 to 1) x (Ign Angle correction due to global adaptive circuit)

So, with a RON value of 0 (high octane), this global timing reduction is without impact.  If ignition correction is more than Max ignition for RON adjustment then the RON will move from 0 toward 1.  Any movement away from 0 will produce a timing reduction with 1 producing the largest timing reductions at high airflow. This background octane learning system does not reduce instantaneous ignition corrections. In fact, this system depends upon knock responsiveness in order to learn fuel quality. 

This system not only impacts timing adjustment but also can decrement power as ignition corrections indicate lower fuel quality.  It’s a logical system that in the absence of acceptable fuel, turns down the power to protect the motor. 

The magnitude of this adaptive system is indicated by the a table in the Torque Control folder called Minimum Value For Torque Limitation due to IGA retard

Minimum value for torque limitation due to Ignition Correction - Requested torque values above these thresholds will be limited as a function of knock correction. The higher the observed knock correction, the lower the allowable torque before limits are applied.  IE, under conditions of high knock, torque will be limited

Knock detection Sensitivity If running built engines, you can change the knock detection thresholds in the software. By raising these tables, you are de-sensitizing the knock control system. THIS SHOULD BE DONE WITH GREAT CARE AS DOING THIS CAN RESULT IN ENGINE DAMAGE AS KNOCK CONTROL BECOMES LESS ACTIVE!!

Ignition TIming Control Beyond Factory Strategy:  Combustion Mode Timing Offset

One common challenge in newer Porsche calibration is ignition timing control.  Starting with the 991.2 generation of vehicles ignition timing became much more complex.  Each vehicle is mapped with 64 total tables grouped in 4 folders of 16 tables.  These tables are segregated by valve lift.

• All lift (both intake and exhaust valves are high lift)
• Exhaust lift (only the exhaust is high lift)
• Intake lift (only the intake is in high lift)
• Standard lift (both intake and exhaust are low lift)

Within each of these groups of 16 the timing tables are further divided by their intake and exhaust cam position.  The final ignition timing is calculated as a weighted blend of more than one table as cam position is not exactly that indicated by the index. 

Ignition timing depending airflow and engine speed.  References by valve lift, Intake cam position, and Exhaust cam position.   [Index 1 is intake cam advance] 0 = 0, 1 = 20, 2 = 40, and 3 = 50 degrees.  [Index 2 = exhaust cam position] 0 = 0 degrees, 1 = 20, 2 = 30 and 3 = 40 degrees.  All ignition values are weighted by the actual camshaft degrees.  For example, intake cam degrees of 10 and exhaust cam degrees of 10 would interpolate 50% between intake cam index 0 and 1.  Exhaust cam position is similarly weighted across index tables.  Exhaust camshaft of 35 degrees would use a weighted value of 50% from index 2 and 50% from index 3.  See tuning guide for further description.

One way to get around this difficulty is to allow the factory to do most of the timing calculations and then use the “Ignition Angle Correction Combustion Mode 0-3” These tables are defined by RPM and airflow and have a huge amount of granularity as you can see. These tables are NOT used from the factory and are populated with “0”. However, careful application of timing (both positive and negative numbers can be used here) is an excellent way of managing FINAL timing output. 

Here you can see a dyno plot from a 991.2 GTS with a turbo upgrade.  Without using these timing control strategies and using the stock system the car makes 45 fewer WHP and the power is very “choppy” as you would expect on an under-timed engine (RED).  When timing is added with the combustion mode tables, the power is smooth and substantive.  This car was tuned with 100 octane unleaded fuel and was still well below MBT using this approach.