Analog geography
No longer is it safe for the designer to hide in a cubicle and wait for specs to come over the wall from marketing.
Chris Mangelsdorf, Analog Devices -- EDN, 11/1/2004
The geography of analog design is changing rapidly. Design centers are springing up all over the planet. It is tempting to point to expanding economies or the availability of low-cost talent around the world, but the real reason resides elsewhere. The fundamental nature of the analog-IC business is changing.
In the past, analog ICs were relatively simple blocks, such as op amps, ADCs, or regulators. Although these blocks might have been internally complex, or might have possessed sophisticated technology, externally, their function was simple. ICs could generally embrace customer requirements with a handful of specs: bandwidth, number of bits, input impedance, and so on. A lone marketing engineer dispatched to a remote location could come back with the necessary information for the next generation of products. Design engineers never had to leave the safety of their cubicles.
Now, however, the clear trend is toward system-oriented analog components. Sometimes, it’s as simple as integrating analog pieces together with the digital, but more often than not, it implies the development of an entire analog system. Consider that camera and camcorder processing chips now integrate the whole signal-acquisition and timing chain, including correlated double sampling, gain control, black-level correction, analog-to-digital conversion, programmable timing generation for the CCD, and even drivers and power converters.
This level of system offers not only the conventional integration benefits of smaller size, lower cost, and reduced power, but also a demonstrable performance advantage. PCs, cell phones, audio components, wireless products, consumer goods, and portable devices of all kinds now use complex analog “system” chips that offer more than just one-package convenience; they enable a whole new level of performance.
This new type of analog system chip leads to a whole new set of problems in a global economy. No longer can an IC vendor define a device by a simple collection of specs. Sometimes, whole phonebook-sized spec sheets are inadequate. Many times, the project requires esoteric applications knowledge or proprietary system know-how of which a typical IC designer is unaware. How many IC designers are also experts in cell phones, for example, or display technologies? Sometimes the performance requirements on the analog system are so complex that even the customer does not know what the application needs. Beyond raw signal performance, the interaction of this complex chip with the application board can be difficult to predict, much less specify.
The result of all this new complexity is that the IC designer must work much more closely with the customer’s engineers to develop the whole system. No longer is it safe for the designer to hide in a cubicle and wait for specs to come over the wall from marketing. The IC team must engage with the customer’s development team on a regular-sometimes daily-basis. The IC team has to understand the application, appreciate the customer’s value proposition, predict application problems, and negotiate the delicate interaction between the silicon and the customer’s board. To achieve this goal, the IC team must be local and must speak the same language. There is far too much traffic to achieve success across several time zones in anything but your native tongue.















