Gasoline engine Calibration

[sprasad6]

Hi all, Just wanted to get some basic insights into gasoline engine calibration. So as I understand, once the engine is designed and validated it goes in Calibration. Just wanted to know the basic steps thereafter. Information on the following will be much appreciated:
1) What is base Calibration?
2) What is the Calibration software and who designs the algo/s? OEM or ECU supplier?
3) What do you mean by a locked ECU? Who has the key to unlock it? OEM or ECU supplier?

These are some basic doubts. I'll probably ask more if the discussion starts rolling.

[anonymous-user]

sprasad6 said:

Hi all, Just wanted to get some basic insights into gasoline engine calibration. So as I understand, once the engine is designed and validated it goes in Calibration. Just wanted to know the basic steps thereafter. Information on the following will be much appreciated:
1) What is base Calibration?
2) What is the Calibration software and who designs the algo/s? OEM or ECU supplier?
3) What do you mean by a locked ECU? Who has the key to unlock it? OEM or ECU supplier?

These are some basic doubts. I'll probably ask more if the discussion starts rolling.

an OEM engine program typically goes as follows:

1) the Research dept develops a prototype combustion system (often using a single cylinder test engine) in conjunction with a massive amount of computer simulation and design. That single test engine is run on a dyno to confirm it's performance

2) Research runs production studies, using the prototype engine, to confirm that it is possible (and sensible) to go to volume production with that engine and the tech it uses. Many, many 'brilliant on paper but non viable' designs die at this point

3) Research creates a specification and target document, that sets out the engines attributes in a production setting (ie power, capacity, masses, emissions, NVH etc etc etc). Senior management in the product planning dept use that document to set the spec for the next production project (ie, effectively the new model of a platform)

4) Product Development (PD) now take over, with the mandate to take the prototype design to production, with the following steps:

a) High level design and simulation;
b) Detail design and simulation
c) Attribute Prototype hardware - validates that the engine can indeed meet the targets set for it, using unique/custom/prototype hardware
d) Confirmation Prototype - second level prototypes that start to use off-tools parts (off tools are parts made with production tooling)
e) Validation Prototype - third level, all parts should be off-tools
f) Production Prototype - final level of validation hardware, using off-tools and off-process parts (ie, using production line parts and production line processes)
g) Production Prototype 2 - Parts built using all the planned methods and processes, often also used to make "press cars" or pre-launch demo vehicles etc.

Typically, the full process will be 5 years, with 2 years in R&D and 3 years in PD


During stage 4) the engine and calibration gradually matures over time, getting closer and closer to true production spec. As there are a huge number of interdependent factors, the process can be highly non-linear, and these days, DOE and automapping techniques are used to try to speed up the optimisation.

In terms of your questions:

1) Base calibration is fundamentally the calibration attributes that depend directly upon the base engine and will not change as the platform matures. Here we are generally anything that affects engine Torque, ie spark timing, cam timing, EGR, fuel mass and timing. Those factors will be optimised and set, and re-validated on each new hardware level. As the program matures, other things that must be optimised, ie exhaust emissions, depend upon having stable and optimium values for the base cal as soon as possibly, so this is always a primary aim early on for PD activities. It's also critical to get other departments something to start with, for example the chassis team, who need to get something to drive ASAP, often a mule, butchered up from an old model, but that needs to at least drive so they can get some attribute setting done. Delays to Base Cal have massive knock on effects to any program! If the base engine and base cal at say CP level fail to meet the targets set, it's often a difficult decision for management between delaying the project to meet those targets and carrying on with a reduced targets

2) The control strategy in the engine controller (strategy = the algorythms that run to effect the control, calibration = the data they use to calculate the output) is usually from a mix of sources. Simpler, lower performance projects will use strategy developed by the engine control supplier as there won't be that much 'clever' stuff going on, but for high perf projects or projects that are "lead" with regard to any particular tech, will run strategies developed by R&D during their pre-production work. Specific control strategies will generally always be finally written by the ems supplier, under direct control of the PD cal team in order to ensure overall quality.

3) All production ECUs are "Locked" ie the memory regions of those controllers are hidden behind an access function that (should) prevent simple reading and writing of that memory. The locking function is owned by the ecu supplier, and the Key to unlock access is, of course, granted to the PD team to allow them to modify that memory during the project. During the project, descriptor files are released by the ECU supplier that include the necessary key to allow the PD team to optmise the memory using the usual, industry standard, calibration toolsets (ETAS INCA, ATi Vision, etc etc)
Of course, nothing is 100% secure, various methods exist to bypass of circumvent the memory locks, and it simple depends on how much effort any 'attacker' wants to put into getting into the ecu (which usually depends on the financial reward for doing so ;-)