Principles of Top-Down Mixed-Signal Design
Roles and Responsibilities
and interface. Areas of special concern, such as critical paths, are identified up front in the verification plan and simulated at the transistor level. The performance of the circuit is verified by simulating just the signal path or key pieces of it at the transistor level. Finally, if start-up behavior is a concern, it is also simulated at the transistor level. The idea is not to eliminate SPICE simulation, but to reduce the time spent in SPICE simula- tion while increasing the effectiveness of simulation in general by careful planning.
It is in this phase that the dynamic timing simulators (fast reduced-accuracy transistor- level simulators) play an important role. They often have the capacity to simulate large mixed-signal systems at the transistor level for a reasonable period of time. Again, even with timing simulators the simulations are generally only fast enough to provide limited verification. So use of a timing simulator does not offset the need for mixed-level simu- lation.
During the design phase, the test engineers should use top-level description of the design as a simulatable prototype upon which to develop the test plan and test programs. The availability of a working model of the system early in the design process allows test engineers to begin the development and testing of test programs early. Moving this activity, which used to occur exclusively after the design was complete, so that it starts at the same time the block design begins significantly reduces the time-to-production [10,11,27]. Bringing test development into the design phase can reduce post-silicon debug time by 50% and can eliminate a turn by finding chips that are untestable early. It can also improve tests, which then improves yield.
6 Roles and Responsibilities
Current systems-on-chip (SoC) designs are in general very complex with the mixed-sig- nal portion just a piece of a much larger whole. Within the mixed-signal design team, there are a number of individuals fulfilling leadership roles and responsibilities that are necessary to assure the success of the entire SoC. Those roles and responsibilities are attributed to individuals and described next. However, this is just one way they could be divided up between the team members. It is also not necessarily the best way, which would depend on the particular strengths of the team members as well as the type of design and size of the design team.
The team lead would supervise one, and only, one design project at a time. He or she works with the program manager to manage the schedule and the resources, acts as the technical interface to the customer, and is ultimately responsible for the technical execu- tion of the implementation. As the primary driver of the implementation, the team lead is responsible for the developing the simulation and modeling plans, including the pro- cess corners, and managing the relationships and communications between the block designers. He or she also interfaces with the design methodology group and is responsi- ble for keeping the process libraries and design kits updated. Finally, the team lead works closely with the system designer, the top-level designer, and the architect to spec- ify the overall architecture, to negotiate the overall specifications as well as the specifi-
20 of 31
The Designer’s Guide Community