Traditional Approaches to Mixed-Signal Design
Principles of Top-Down Mixed-Signal Design
this technically is not a change in the requirements or the process, it has a similar effect on the design process. Rather than it being a change in the process, it is a change in the understanding of how the process affects the design, and the design team must react to this change midway through the design process.
2.6 Gigabit I/O
In many applications, one can make an economic judgement as to whether it is better to put the mixed-signal circuitry on the same chip as the rest of the system to reduce costs, or to put it on a separate chip to reduce risk. This is the case, for example, with most wireless systems. However, with high speed serial I/O there is no choice available, the mixed-signal circuitry must go on the same chip. Thus, the large system-chip inherits the risk associated with mixed-signal circuitry. Given the large opportunities and invest- ments1, and the short market windows generally associated with these chips, this is not acceptable. As mentioned before, a respin can be the difference between capturing a market and having no market at all.
3 Traditional Approaches to Mixed-Signal Design
At the Design Automation Conference in 1998, Ron Collett of Collett International pre- sented findings from a 1997 productivity study in which his firm analyzed 21 chip designs from 14 leading semiconductor firms. The study revealed a productivity gap of 14× between the most and least productive design teams. The study also revealed that developing analog and mixed-signal circuitry requires three to seven times more effort per transistor than designing digital control logic, though this factor was normalized out of the 14× ratio.
The reason for the poor productivity of those at the bottom of the scale are increasingly complex designs combined with a continued preference for bottom-up design methodol- ogy and the occurrence of verification late in the design cycle, which leads to errors and respins. There's a huge disparity in productivity between those mixed-signal designers who have transitioned to an effective “top-down” design methodology, and those who practice “bottom-up” design and rely solely on SPICE.
The traditional approach to design is referred to as bottom-up design. In it, the design process starts with the design of the individual blocks, which are then combined to form the system. The design of the blocks starts with a set of specifications and ends with a transistor level implementation. Each block is verified as a stand-alone unit against specifications and not in the context of the overall system. Once verified individually, the blocks are then combined and verified together, but at this point the entire system is represented at the transistor level.
1. A 0.25μ mask set costs $150K and includes 15 masks. A 0.18μ mask set costs $250K and includes 21 masks. A 0.13μ mask set costs $600K. At 90nm, masks are expected to cost $1.5M and at 65nm they will cost $4M .
The Designer’s Guide Community
9 of 31