ASSPs — The Fruit Fly of the Automotive Industry
By Martin Mason, Actel Corp.
As an FPGA vendor, it warms the cockles of my heart to hear an automotive customer say: “I would never get any [ASSPs] in production. They’re the electronic equivalent of fruit flies.”
Martin Mason, Actel Corp.
By Martin Mason, Actel Corp.
As an FPGA vendor, it warms the cockles of my heart to hear an automotive customer say: “I would never get any [ASSPs] in production. They’re the electronic equivalent of fruit flies.”
Why fruit flies? In electronic system design terms, most ASSPs have a similar “blink-and-you-missed-it” life span as that of the fruit fly. Apparently to even mention an ASSP to the automotive component engineering department, your job is put on the line. Product obsolescence, qualification hurdles and life cycles are the real killers, right?
Imagine: When designing a model year 2009 automotive application, you specify an ASSP. You receive a product end-of-life notice just six short months before the start of production. Compound this with the almost-impossible effort involved in getting data from said supplier to support an extended temperature range or to meet AEC-Q100 qualification or the typical PPAP documentation process requirements.
So, where does the automotive engineer turn?
The answer used to be ASICs. However, by the time the ASIC is designed, validated, built and qualified, the model year 2009 target goal is unrealistic. Further complicating the matter, NRE expenses for even a simple digital ASIC can be cost prohibitive, particularly considering qualification expectations put on the supplier for this one device.
Ever-changing technology standards and greater time-to-market pressures are driving many automobile designers to turn to FPGAs instead of the custom ASICs, MCUs or ASSPs they have traditionally relied on.
A traditional hurdle for FPGAs has been cost considerations. However, that tide has begun to change.Customers have said: “I never thought we would have the budget for an FPGA. As a flexible and reprogrammable platform, the costs associated with qualification of a single FPGA device are spread over multiple projects, saving us money.”
As an FPGA vendor, clearly, this is exciting. However, not all FPGAs are created equal. Some are less suitable for certain classes of core automotive systems and safety applications. Unlike nonvolatile, flash-based FPGAs (www.actel.com/products/solutions/auto/default.aspx), SRAM-based devices are susceptible to neutron-induced firm errors, which can be potentially disastrous in certain safety applications, and a serious quality concern in others. Nobody wants to solve a short life span or availability issue with a long-term quality issue (see earlier blog: Quality, Reliability and Nonvolatile FPGAs for Core Automotive Systems).
Today, the economics of programmable logic, such as nonvolatile, flash-based FPGAs, are changing the part selection process. The ability to utilize a low-cost standard FPGA that is qualified to AEC-Q100 and PPAP offers the automotive market several significant flexibility, time and cost advantages. Not insignificantly, an FPGA offers a typical 10+ year product life cycle, easing the mind of those risk-averse automotive designers who’ve been burned by their share of ‘fruit fly’ devices.
With ever increasing demand for safety and body control features as well as the growing complexity of automotive electronics, semiconductor vendors must meet the needs of automotive designers without forcing them to also consider product obsolescence or life cycles in addition to the quality, reliability, flexibility, cost and time-to-market delivery factors critical for this market. As technology standards and regulations change and qualification costs and time-to-market pressures grow, designers are choosing nonvolatile, flash-based FPGAs over ASICs and ASSPs, the fruit flies of the automotive industry, ensuring product availability while maintaining product quality and reliability.
Martin Mason is the director of silicon product marketing, Actel Corp. www.actel.com
Advertisement
