Optimizing Sensor Performance with 1T-OTP Trimming

Update: Synopsys Expands DesignWare IP Portfolio with Acquisition of Sidense Corporation (Oct. 17, 2017)

By Jim Lipman, Sidense Corp.

Introduction

A sensor is a device that detects a change in a stimulus and converts it into an electronic signal that can be measured or recorded. The stimulus can be many things, including a physical property, environmental parameter, chemical composition or a location, to name just a few. All sensing elements have nonlinearities that include an intrinsic nonlinearity over sensing range along with offset and sensitivity nonlinearity variations over temperature.

Variations in component and circuit characteristics along with chip processing and packaging operations result in deviations of analog circuits and sensors from their target specifications. To optimize the performance of the systems in which these components are placed, it is necessary to "trim" interface circuitry to match a specific analog circuit or sensor. A trimming operation compensates for variations in the analog circuits and sensors due to manufacturing variances of these components.

The trimming requirement becomes more important as process nodes shrink due to the increased variability of analog circuit performance parameters at smaller processes, and to both random and systematic variations in key manufacturing steps. This manifests itself as increasing yield loss when chips with analog circuitry migrate to smaller process nodes, since a larger percentage of analog blocks on a chip will notmeet design specifications because of the variability in process parameters and layout.

Where Sensors are Found

Many electronic systems use sensors to interact with 'the outside world' and use the information from these sensors to optimize system performance and/or to provide useful information to a human being. The many different types of sensors include piezoelectric, magnetic, thermistor, MEMs capacitors, electrochemical, photodiode and geomagnetic.

Two of the most common systems that rely heavily on sensors are automotive and mobile communications, the latter best represented by the smartphone. Add to these the rapidly expanding Information of Things (IoT) ecosystem and you have some of the major markets that use a broad range of sensors to measure many different parameters.

Automotive Systems

The number of automotive sensors in a single vehicle has been steadily rising over time. According to the MEMS Journal, each new automobile has 60-100 sensors on-board measuring a very broad range of parameters, including temperature, humidity, light, pressure, fluid levels, positioning, engine combustion/detonation, acceleration, speed, lamp status, oxygen flow and compass direction (geomagnetic). The type of sensor a vehicle uses for a specific application depends on the parameter being measured and a single car will use several different types.

Many of these sensors operate under high temperature conditions and thus require that they and the electronic components with which they are used operate with a very high degree of reliability in their working environments, which include temperatures up to 150°C for automotive "under the hood" operation (AEC-Q100 Grade 0).

Mobile Communication Devices

With smartphones representing one of the highest growth segments of the mobile communication market, it is not surprising that the number of sensors found in new smartphone is also on the rise. Mobile device sensors measure parameters such as acceleration, position, barometric pressure, temperature, humidity, sound, geomagnetic position, gestures, proximity, light intensity, infrared and image recognition.

One example, theGalaxy S4 smartphone, integrates several different sensors, including an accelerometer, RGB light, geomagnetic, proximity, gyroscope, barometer, temperature, pressure, humidity and infrared. Embedded NVM can be used with all of these for a variety of applications to optimize the various smartphone subsystems.

IoT Ecosystem

IoT is poised to open the door for IP vendors, silicon foundries and software designers to develop products for literally tens of billions of connected devices over the next few years. In a recent article the Internet Business Solutions Group (IBSG) at Cisco predicts that there will be 25 billion devices connected to the Internet by 2015 and 50 billion by 2020.

The explosion of information that IoT devices will gather will require huge numbers of processors to process and manage the data from these IoT devices along with lots of low-cost, secure and reliable embedded non-volatile memory (NVM) for code storage, sensor trimming, device configuration, security keys and other storage functions. Virtually every one of the projected billions of IoT devices can use some amount of one-time programmable (OTP) memory. Requirements for this OTP include low cost, low voltage and power, field programmability, fast-start-up time and secure content storage.

The IoT ecosystem will impact a very broad range of market segments, including but not limited to industrial, consumer, retail, medical automotive, environmental, military, agriculture and automotive.

Storing Trim Parameters with 1T-OTP

Depending on the native system's requirements, there are several available technologies for storing sensor trim parameters, including flash memory, EEPROM, EPROM, eFuse, ROM and OTP. One technology that has gained broad acceptance is antifuse-based OTP memory. Antifuse OTP relies on permanent oxide breakdown to store trim bits and is easily programmed irreversibly on-chip during manufacture or in-system.

Sidense 1T-OTP IP is a good example of embedded antifuse OTP, providing many advantages compared to other non-volatile memory (NVM) technologies for sensor trimming. 1T-OTP uses Sidense's patented 1T-Fuse™ bit cell, a single split-channel transistor that encompasses both thick (IO) and thin (gate) oxide regions below the transistor's gate. The one-transistor bit cell results in a small memory footprint for minimal cost impact. In addition, since Sidense 1T-OTP is fabricated in standard logic CMOS processes, without any additional masks or process steps, there is no added wafer manufacturing cost. Low cost NVM is very important for cost-sensitive devices in the automotive and mobile communications markets.

When a sufficient programming voltage is applied to the split-channel transistor, either from an on-chip charge pump or from an external source, the oxide breaks down between the gate and channel, forming a conductive path and programming the bit cell. The programming is controlled and non-reversible. With this split-channel architecture, the breakdown and hence programming channel is only from the thin-oxide transistor gate to the channel, resulting in a very reliable and easy to read bit cell. The permanent programming results in long-term retention, even at higher operating temperatures such as the 150°C temperature for automotive under-the-hood operation. The antifuse technology Sidense uses is already proven in processes from 180nm to 20nm in polysilicon and HKMG processes, in addition to a number of High Voltage (HV) and power variants.

Another feature is the inherent high data security of 1T-OTP, making it ideal for applications such as storing encryption keys for secure transactions from a smartphone that cannot be hacked. The 1T-Fuse bit cell is practically impossible to reverse engineer even with advanced inspection techniques and since there is no charge storage, as is the case for some other NVM technologies, contents cannot be read using voltage or current scanning techniques. 1T-OTP macros also have additional features in the sensing circuitry and logic to further enhance data security against various forms of attack. 1T-OTP macros also support various read modes, including differential read modes, which can further improve data security.

1T-OTP macros provide a low-power NVM solution for applications such as mobile communication devices and IoT remote sensor circuits. The 1T bit cell means fewer transistors in the memory array and results in less power drain. A small OTP memory area means lower bit and word line capacitance, which reduces precharge and switching power. 1T-OTP macros have low read voltages and using a differential read mode lowers the read voltage and hence power further. 1T-OTP macros have additional circuit features to minimize power and using an integrated power supply macro optimizes OTP memory read and program voltages from the normal chip voltages.

The field-programmability of 1T-OTP macros supports recalibrating trim bit settings to account for sensor aging and possible changes in the ambient environment around a sensor. Operation at high-voltage and BCD processes, along with support for high temperatures expands 1T-OTP use into many sensor applications in automotive and industrial market segments.

1T-OTP: A Natural Fit

With such a wide range of sensor types, applications and environments it is important for the sensor and interface circuitry trim operations to be optimized for performance and power. Embedded 1T-OTP NVM provides enable low-cost, low power, high reliability, very secure and field-programmable trim parameter storage over a broad range of manufacturing processes and environmental conditions. Considering the many requirements of the various systems that incorporate sensors, 1T-OTP is a natural fit for sensor trimming and conditioning.