The process of many steps requires significant documentation along the way to ensure efficient redesign later.
BY NIGEL HUGHES
The electronic content of a modern transportation platform, such as automobile wire harnesses, has become one of its key differentiators in the marketplace. Alongside the obvious problems in ensuring that the latest technologies are available is ensuring that the associated systems supporting their deployment can be developed and integrated into the final vehicle.
Problems are compounded by the pressure of ever decreasing design cycle times that have been reduced to between 24 and 18 months, with a goal for most automotive OEMs to reduce the time to 12 months by 2010. Within this cycle, the electrical distribution system supporting the electronic content must be designed, validated, and deployed. Doing so is increasingly dependent on developing new processes and tools to support those processes.
To preserve a reasonable margin, automotive OEMs must also ensure that an appropriate return is gained from the technology by specifying vehicle models with an appropriate set of options that support accurate pricing. Recently, the market has also seen a dramatic increase in the amount of optional content in a vehicle. This reflects the need to make the latest technologies available, while ensuring the basic models are still available in an extremely price-competitive market.
A key part of these new processes is increasing the amount of re-use of previous designs. Traditionally, this has involved a painstaking process of taking existing wiring designs and carefully updating them for new vehicle harness platforms. Such a manual process inevitably introduces errors, and demands significant validation and verification.
Re-use of previous designs tied to a particular platform is difficult. Almost all details of the wiring will change as it is integrated with a different physical platform (e.g., all of the wire lengths will need to change, locations of in-lines and splices will change). But logical systems can be designed that are re-usable and provide a sufficient level of abstraction to significantly increase the amount of direct re-use while still not requiring large amounts of manual re-work.
Process overview
We will consider each stage of this high-level process in more detail later, but first, let’s consider the broad context in which an electrical distribution system (EDS) is designed. First, systems are selected that will implement the required functionality from any available libraries. New vehicles will typically deliver at least some new content, so those systems must be developed (sometimes by a Tier 1 supplier).
All captured systems are then normalized (option and variant expressions from the particular platform applied), and shared connectivity identified. Then the designs must be placed in the context of the harness topology (the physical routes the wires will take through the vehicle), and the nets (signals) converted into actual wires inside harness bundles. Components, such as fuses and wires, can now be sized appropriately based on the amount of current they are expected to carry.
Once this step has been completed, individual wiring diagrams can be generated, which are passed on to service groups who will re-layout the diagrams for service manuals or other distribution channels. The harness designs are then passed onto the harness engineering groups, who will further refine the detail of the physical implementation by adding mechanical components (such as clips and grommets) and specifying terminals, harness dressings and coverings.
The starting point in the design flow is collecting existing system designs that will be re-used in the new vehicle model or platform. To make systems more re-usable than their physical implementation in a wiring diagram, a more abstract representation is needed.
Devices (or components) are placed together, and connected by nets that carry the signals (e.g., power or sensor readings) between the devices. An example is shown in the figure, “In-vehicle entertainment system design,” page 20.
As shown, signals run from the AUDIO_UNIT to the speakers, and provide I/O from separate elements, such as the CD CHANGER. With the connectivity information captured in this abstract form, it is possible to import it from a library into a specific vehicle platform.
These systems must also identify their optional content. For the in-vehicle entertainment system, for example, you can see a number of areas that have been marked out—the additional two speakers (marked 6 speaker), and the sub-woofer. These are purely visual indications that the system contains optional content.
Designing for options
What is optional or standard is determined by the marketing plan for a particular vehicle (a high-end platform may include all content as standard, while a low-end may have everything as additionally purchased options). Therefore, the content of the design will have to be marked with the applicable option tags for any given platform.
Other considerations include power and grounding. Note in the figure that all devices have their own grounds. These are logical grounds that may be combined into a single physical ground when placed inside the vehicle. This process will happen during the physical integration stage, described later.
Power and the attendant fusing also provide challenges. Each device requires power, but you may not wish to represent the ignition system on every diagram, and you certainly cannot specify the correct fusing implementation in a re-usable system design. As a result, a typical process involves placing shared nets (nets that appear in many diagrams but are actually a single logical entity) that supply the power. These can be seen in the figure labeled PWR_IGN_RUN and PWR_IGN_ACC.
At a later stage, a power system design will be created that connects all of these shared power-supply nets to the fusing architecture being used in the vehicle and associated power-supply components (e.g., batteries and alternators). For example, you might want to progress to initial topology creation and wire synthesis before coming back to finalize the specification of the fuse architecture. These cycles will require multiple versions of a system during its lifecycle.
The process of managing multiple versions of the system design will be heavily dependent on the requirements of the OEM; however, there are some consistent themes. When releasing a design, it becomes read-only and, therefore, cannot be edited itself. The intention is that only released designs will be taken forward to the integration stage.
It is unreasonable to assume that no further changes will take place. So, there may be a need to produce a new revision of a design (a copy of an existing design, where the relationships between content in the new design are traceable to the original design), which can then be edited and the changes re-integrated into the vehicle. An example of where this release-revision-release cycle should be expected is in the power system design described earlier, which is expected to change as the systems are integrated into a particular platform.
Looking back at the design flow, you have collected the existing designs being re-used, and developed the additional system designs required for this platform. Once a set of designs has reached a sufficient level of maturity, they can be released and grouped together into a set, or build-list, of versions that are compatible with each other. A change to the power design may necessitate a revision in another system, and only those two versions are compatible with each other. This build-list can then be passed on to the integration stage.
Harness integration
The systems are now integrated into a harness topology. Once again, the source of the topology will vary from one OEM to another, but a common source is from an MCAD tool, such as UG’s NX or Dasault’s CATIA V5. Harnesses must be bridged into the design tool and connected together.
The bridging process in modern software tools is no longer a complex and delicate procedure, often allowing manipulation of the data as it is converted from the 3-D world into the 2-D topology. The topology could also be drawn manually in 2-D and later synchronized with a 3-D model as it became available.
Now that the topology is complete, you can move to the next stage of the flow and place devices from system designs into the harness topology at their correct locations.
For example, the stereo would be placed into the center console, perhaps attached to the instrument panel harness. Some devices may have multiple locations depending on options or variants. For example, left- or right-hand-drive configurations will cause an instrument binnacle to be placed on one side or the other.
Rules can be used to control the automatic placement of components. A rule is a formal description of a heuristic used by an actual wiring engineer, such as “place all fuses connected to devices on the engine harness in the engine fuse-box.” The benefit of capturing this information in a software tool is that the expertise can be repeatedly and reliably re-applied. Rules are used not only to constrain where devices are placed, but also to show how wiring is generated—the next stage in the flow.
For synthesis, appropriate rules might include limiting the number of wires that can be spliced together (perhaps changing that limit depending on whether it’s a wet or dry harness), the distances between splices and their devices, and the maximum length of a wire. Rules applied to the released systems, which have been integrated into the harness topology, can be used to complete the next stage of the flow—the synthesis of the actual wiring.
Generating wiring
Before individual harness derivatives supporting particular vehicle models can be produced, the 150% or composite wiring must be generated, supporting all allowable combinations of options. This composite wiring will never be used in a single car delivered to a customer, and as such is “un-buildable.” To actually define the different buildable configurations, you must generate derivatives of each harness that contain a subset of the generated wiring (and splices).
The initial set of configurations may produce more derivatives than can be cost-effectively manufactured and, therefore, decisions will be made about options that may have their associated wiring given away. It might mean that the wiring for a six-speaker system is always included in all harnesses regardless of whether the customer has selected that option. Typically, each harness in an automobile will have 10 to 20 different manufactured derivatives.
Although there may be a number of cycles back through the process, such as applying changes to system diagrams and regenerating the wiring, your full EDS design is now complete. You can now hand off automatically generated wiring diagrams of systems, or parts of systems, to service documentation groups for inclusion in electronic or printed service manuals.
NIGEL HUGHES is product development director with Mentor Graphics Corp. (www.mentor.com/electrical). He is based in Mentor’s Berkshire, UK office and can be reached at: nigel.hughes@mentor.com
