Multimedia Systems Design
By Dr. Uwe Lautenschlager, Continental Automotive GmbH
Continental’s Automotive Group, a system supplier for international automotive OEMs, is comprised of 3 divisions: Chassis and Safety, Powertrain and Interior. The division Interior is itself split into 4 business units: Instrumentation & Driver HMI, Infotainment & Connectivity, Body & Security, and Commercial Vehicles & Aftermarket.
The business unit Infotainment & Connectivity covers Radios, Connected Radio and Entry Navigation, Multimedia Systems, Telematics, Device Connectivity, and Software and Special Solutions, with a particular emphasis on Entertainment, Information, Navigation and Communication.
Multimedia Systems with Audio, Navigation, Internet, Telephone, TV & Rear-Seat Entertainment Functions
One of the greatest challenges for automotive multimedia systems is that they have to operate under a wide range of environmental conditions:
- Temperatures ranging from -40oC to +100oC
- Wide range of power supply voltages (~4.5 to 16.0 V)
- Superimposed alternating voltages
- ESD polluted environment (ESD = Electrostatic Discharge up to +/-15 kV)
- Full range of humidity (0% - 100%)
- Chemical influences
and with stringent design criteria:
- Mechanical strength and stability (static /dynamic): Vibration & mechanical shock
- Strong limitations on size and weight
- Strong requirements regarding risk of injury (Head Impact) and product liability
A World Away from Physical Prototyping
Designing systems to meet such requirements requires a lot more than physical prototyping after the design is complete. Indeed, Continental’s simulation vision is to Get the Product Right the First Time, and their strategy to achieve this has been through Simulation-Based Design Decisions. As a result, simulation is now very highly integrated into the design process.
Continental undertakes full 3D system-level modelling, with all thermally-relevant and air flow parts included in the simulation. The geometry representation based on mechanical CAD data, with export/import of all relevant parts and modules from CAD-System. However, the geometry representation can involve both simplification and idealization. The complexity used depends on the simulation objective: less detail would be used for the analysis of various initial concepts, whereas more details would be included in the model for a mature design.
“Know How” – Continental’s Competitive Edge
Continental has developed considerable in-house ‘know how’ associated with building and validating thermal simulation models (FloTHERM models) of their complex electronic systems and sub-systems, covering both the physical hardware, and potential cooling solutions.
|Housing: 1-DIN, 2-DIN, customer-specific CD-, HDD-, DVD-Drives (ROM, Video), Changer Shielding frames Front displays (LED-Backlight) Connectors and cables Components: Main-CPU, Processors, Memory PCBs Modules: Power Supply, Drives, GPS, Telephone, Amplifier, Tuner, Display, etc.||Fan (type, characteristic curve, direction) Contact resistances (conductive paste, etc.) Vents, filters (free area ratio, pressure loss) Heat sinks, heat pipes, heat spreaders IC Packages (type, size, thermal model type) Thermally conductive gap pads, etc. Environmental conditions, use cases Power dissipations: min., typical, max., use case Location of sensors for measurement|
Uncertainties arising from the various unknowns at each stage in the design process, modelling assumptions and simplifications are mitigated through validation, as the following two examples illustrate.
Example: Tuner-Module Modelling
This can be modelled as a single block with homogeneous power dissipation distribution, as a detailed model with all of the internal electrical components and shielding represented, or anywhere in between. In Continental’s experience, a simplified model is often sufficient within a complete system analysis. However, conclusions on the internal component temperatures are drawn from a reference simulation as depicted below.
Single Block (System-Level) and Detailed (Reference) Representations of Tuner Module
Example: Chassis Model Verification
A part of the chassis is used as heat sink for amplifier cooling. An important aspect of the chassis is the extent to which the brackets on the top and bottom covers are in thermal contact (related to manufacturing tolerances). Verification of modelling assumptions is the key to improving model fidelity. Therefore, a prototype of the product was measured and the results used to compare against the simulation model. Scenario 1 (g1) considers the gap to be fully closed (perfect thermal contact) whereas Scenario 2 (g2) has a gap of 0.3mm, so there is effectively no thermal contact between the covers. By comparison with the measured case temperature the results for Scenario 2 (with gap) seems to be the best representation of the bracket’s contact.
Verification of Case Thermal Contact Modelling Assumption
Beyond High Fidelity Simulation
Continental’s focus on simulation does not stop at building and validating thermal models. Design decisions made to improve the thermal design can impact the mechanical, electrical, and EMC performance of the product. Faced with this problem, a major question facing Continental’s designers is:
“How can we achieve design flexibility and enable better design decisions before the freedom for such decisions is eliminated?”
The answer requires analysis of the product’s behaviour for all disciplines as well as the identification of independent and coupled system variables. Multidisciplinary Design Optimization (MDO) techniques with simulation, optimization (with discipline-dependent objectives and constraints) and Design-of-Experiments (DOE) & Response Surface Methods are suitable means to solve this design problem. Indeed, DOE tools are the foundation for concept exploration and robust design. Two major aspects of DOE are the planning and statistical analysis of the numerical simulations.
The possible number of discrete and continuous design variables is extremely high. Screening simulations support the selection of the major factors (design drivers) to be included as design variables in the MDO. Even so, the computational effort required can be immense.