Case Study # 6

case study # 06

MEMS Motion Sensor Controller Integrated Circuit

To drive technology innovation and achieve success in the marketplace, you need vision, the agility to meet rapidly changing market needs and the ability to execute rapidly and reliably. Kapik has helped clients distinguish themselves in the marketplace by supporting their agility and ability to execute effecitvely. In this case our client, a 50-person semiconductor packaging and testing company, was seeking to move its business to the next level by adopting a brand new take on an existing product: they would bring to market a PZT-based (lead zirconate titanate) MEMS (Micro-Electro-Mechanical system) gyroscope.

Electrostatic-based MEMS gyroscopes and controller ICs had already been on the market for some time, however they were costly to produce; they required a specialized CMOS process to handle higher voltage and vacuum packaging, both of which drove cost and power requirements up. Our customer saw an untapped opportunity to meet a market need for lower-cost PZT gyroscope chips in the digital camera market and had cut a deal for a new MEMS device.

That said, they needed to do a lot of work to get the device into production and there was a notable technology gap: no controller chip on the market was even close to what they needed. And while the customer had the MEMS from a proof of concept board, the precision oscillator alone was estimated to multiply the cost of a critically cost-sensitive device by a factor of ten, thereby damaging the original value proposition. The company consulted with a local ASIC house for advice on IPs and how to proceed. The ASIC house referred them to Kapik.

This job was the kind of ground-up design challenge we love. It involved inventing the model for the PZT MEMS gyroscope sensor – both the tensor physics and the Verilog-A model – and devising a controller architecture that would match. It then involved the actual design. There were several challenges we had to navigate along the way, including:

New technology, limited existing knowledge: MEMS gyroscopes were still relatively new at the time and so documented engineering knowledge on circuit techniques was sparse in terms of dealing with potential process variation issues. In addition the IC required a number of new circuit development innovations, device modeling techniques and test schemes to evaluate its performance.

New foundry process, variable performance factors: The PZT MEMS gyroscope itself was manufactured using a new process and so no electrical model of the MEMS existed. MEMS sensitivity, resonant frequency and symmetry were expected to vary in the fabricated MEMS.

Power, area and cost constraints: To be viable for use in a variety of consumer devices, the IC design would need to meet rapidly evolving constraints on power and area. In addition, as part of the cost requirements, we had to use vanilla CMOS for the controller, not one of the high-voltage processes on the market.

There ended up being an additional, more significant challenge. Partway in to the project, the requirements significantly changed: the end market (which now included mobile phones), cost target, number of axes (from two to three), duty cycle and resolution. The characteristics of the MEMS device also changed radically several times as the customer tried different foundries. But we kept with our customer and remained committed to their success: we worked with the client to mitigate these challenges and managed the risk associated with the project by taking the following approach:

Developing a flexible electrical model and design: We obtained fabricated MEMS samples, tested them and developed our own electrical model from scratch. Recognizing the potential for processing variation, we created a design and circuit architectures that were robust enough to handle large variations in parameters and process

Sufficient testing and system redundancies: To help ensure the delivery of a high-quality, reliable IC, we ensured that there were sufficient viewable test points, test schemes to evaluate overall system performance and designed redundancies in the system

In the end, we achieved first-time silicon success, meeting our client’s challenging requirements and raising the bar on the size, cost and quality of gyroscopic ICs in ways that directly contribute to a vastly improved end-user experience.

The contribution that our design made in advancing MEMS gyroscopic technology has resulted in new work on a second-generation gyroscope controller. Unlike first generation gyroscopic ICs, which use a predominantly analog approach, we are taking a smart analog approach that enables us to develop a leaner, lower power/area design and utilize more sophisticated digital circuits. The new system architecture uses a single MEMS element that accommodates 3-dimensional actuation and detection instead of two elements that actuate in 2 dimensions. We are also taking a different approach to MEMS actuation by using a digital tracking loop and a digitally trimmed oscillator with feedback to reduce power and area requirements. Furthermore, we are using a smaller, lower power impedance amplifier to amplify the small currents produced by the MEMS.

Interested in finding out more about our expertise in mixed signal and smart analog design can help you?