Civil Engineering Reference
In-Depth Information
Fig. 2.36
Analysis of the bit timing in Saber
implement as it is only a software parameter that needs to be changed. Simulating
a virtual prototype quickly provides an overview of the quality and the robustness
of the CAN topology implementation. Problems related to the implementation can
be discovered early and the developer has a powerful solution available in order to
compare and validate different implementation options.
The previously discussed methodology demonstrated a simple case with just a
handful of data to evaluate. If the number of signals or amount of data is much
greater, it would be very time consuming to manually analyse all of the simulated
signals. Let us assume we had a topology consisting of 14 ECUs and just three
validation criteria (propagation delay, slew rate and settle time). This would mean
the network developer would have to analyse 14 × 14 × 3 = 588 discrete measure-
ments. It may also be necessary to visually inspect other signals in the system. And
if variations of stub lengths as well as worst-case configurations for the transceivers
or cable impedance were to be simulated, then the amount of data would increase
significantly. Obviously, a manual evaluation of the simulated data would be very
time consuming, error prone and would reduce the time-saving benefits of using
a virtual prototype. In Saber, the network developer can fully automate the entire
simulation and post-processing steps. Figure
2.37
shows a possible scenario.
The starting point of the entire process is an automated Round-Robin commu-
nication for which the developer specifies the desired communication pattern. The
entire simulation process is then put into a script and can be run via a single mouse
click. In the second step, the generated simulation data will be further processed/
modified according to the specific requirements of the automotive OEM. Due to
the fact that each OEM has different topology requirements and prefers a differ-
ent representation of the final data, it is infeasible to provide an automated post-
processing solution off the shelf that addresses everyone's needs. Therefore, Saber
