Tactical grade MEMS accelerometer S.Gonseth 1, R.Brisson 1, D Balmain 1, M. Di-Gisi 1 1 SAFRAN COLIBRYS SA Av. des Sciences 13 1400 Yverdons-les-Bains Switzerland Inertial Sensors and Systems 2017 Karlsruhe, Germany 978-1-5386-3962-7/17/$31.00 2017 IEEE Pxx
Abstract SAFRAN COLIBRYS introduce the MS1000 accelerometer, a new class of high performance MEMS accelerometer specially designed for inertial applications. It is based on Colibrys long experience with MEMS technology and introduces an innovative design solution to meet tactical grade requirements. This paper introduces the MS1000 architecture, reviews in detail the key performances of the product, and presents qualification results. 1. Introduction For more than 20 years, SAFRAN Colibrys have been developing MEMS accelerometers for industrial, military, aeronautic, and safety applications. These sensors have proven to be highly reliable in aggressive environments, like firing shocks up to 20 000g when used in guided munitions and in the presence of heavy vibration with repetitive shocks. They can even maintain their high performances over extreme temperature ranges up to +175 C, making them suitable to serve the oil & gas market in applications like down-hole drilling. Key features of this MEMS include robustness in harsh environments, maintaining full performance over operational temperature for the sensor s entire life, low power consumption, and small size factor. A versatile accelerometer platform sensor has been developed and qualified to meet the requirements of inertial navigation, tilt measurement, vibration, and low noise acquisition, and can therefore cover the markets of aero & defence, automotive testing, railway, and structural monitoring. Table 1, below, presents the new 1000 series accelerometers, each designed to address different needs for different market. Table 1. 1000 series accelerometer platform Vibration market Tilt for oil & gas Inertial tactical Seismic for structural industry health monitoring High bandwidth 2.5Khz ±5% or 5Khz ±3dB Extended temperature up to 175 C Long term bias repeatability 1.2mg Very low noise 1μg/ Hz [2]
Basically, there are two ways of measuring acceleration with a MEMS sensor: either via an open or closed loop configuration. In an open loop configuration, the capacitance change in the MEMS is measured and amplified. In closed loop electronics the inertial forces are compensated by electrostatic forces [1]. Although closed loop systems allow for reaching better ultimate performance in terms of bias stability, linearity, and noise, there is a large price to pay in terms of power (needs to be very precise and high voltage), size (driven by the power supply requirements), and complexity (analogue and digital electronics). The 1000 series is a new generation of open loop sensor that, as compared to closed loop systems, has advantage in term of power and size, while still reaching the performances required for high demanding inertial applications. It is also significantly less complex and does not integrates embedded software. The 1000 series share the same ceramic leadless chip carrier LCC20 package as 9000 series with is small form factor and excellent long term reliability. This article reviews the MS1000 accelerometer which has been specifically designed for inertial applications. 2. Design and technology In order to achieve high performance and high reliability in harsh environments, a stable mechanical sensor is necessary. This can be made up of a MEMS device and its associated die attach technology. Safran Colibrys MEMS sensor, a proven capacitive accelerometer qualified in Mil/Aerospace products, is based on a bulk micromachining technology like that illustrated in Figure 1. In this configuration a bulk silicon proof mass is suspended by a spring and detects acceleration in the out of plane direction. It has already been shown that excellent performance can be reached with this approach [2]. Figure 1. Cross-section of a MEMS capacitive accelerometer. Out of plane acceleration will deflect the proof mass and change the capacitances between the middle, top, and bottom plates, respectively [3]
The MS1000 is based on this same robust technology and introduces an innovative MEMS device design and die attach technology which reduces its sensitivity to mounting stress and therefore further improving bias stability. The second key element is a custom designed electronic circuit which was developed with a special emphasis on bandwidth, noise, linearity and to provide stability for operational in various environmental conditions, i.e. over the full temperature range. The capacitive signal from the MEMS sensor is measured by the differential charge balancing loop (C2V) block. In a control loop the capacitor bridge is balanced. This concept [3] has already been successfully used in a previous integrated circuit found in current Colibrys products. In order to improve stability and noise, a fully differential configuration is used which provides a positive and negative output signal. The acceleration signal is the difference between these two. Figure 2. Functional block diagram of the MS1000 The charge balancing loop also includes features to adjust the linearity caused by stray capacitances and sensor nonlinearity. A significant advantage of this concept is that the electrostatic forces applied on either side of the proof mass are always equal, independently of the plate position. This significantly reduces measurement errors caused by electrostatic forces, especially for low-g accelerometers. The signal from the charge balancing loop is fed through a programmable gain amplifier (PGA) to provide the user with an easy to use output voltage. The differential output assures that zero-g corresponds to positive and negative output being equal, independently of any reference voltage. Also, the gain is ratiometric to the power supply voltage. I.e. if the power supply is used as a reference for the subsequent A/D converter, the accelerometer output becomes largely independent of any power supply voltage variation. [4]
The service blocks include a One-Time Programmable (OTP) memory to store the calibration data, an RC oscillator (clock), a power management function (POR), and a reset block (Reset). The integrated circuit is designed to operate between -55 C and 125 C and includes a temperature sensor that can be used for temperature correction by the user. It operates at 3.3V, and uses less than 10mW. 3. Performance and qualification results The performance of the MS1000 has been validated through the acceptance test procedure (ATP) which is performed on 100% of devices during manufacturing, and through complementary characterisations done in qualification. The MS1000 series production test equipment have been totally renewed to enlarge the test coverage and to include all standard calibration steps related to bias, scale factor, linearity, as well as the measurement of the temperature sensor, frequency response, bias vibration rectification error, noise, and a full temperature characterisation over the range of -40 C to 85 C. The following paragraphs present MS1010 (10g sensor) key performances. 3.1 Linearity Linearity is important in inertial applications in the presence of significant vibration, which will translate into a bias shift through rectification. Therefore, all sensors are calibrated thanks to the charge balancing loop which includes two parameters to adjust the even and odd nonlinearity components independently. A calibration method based on IEEE std 1293-1998 [4] is performed on a shaker against a piezoelectric reference accelerometer. This method is based on fitting the response curve with a 3rd order polynomial and subtracting the constant term (bias) and the first order term (sensitivity). The vibration rectification error is measured on a shaker at 50% of the full scale amplitude and from 50Hz to 2kHz. Typical nonlinearity curves of calibrated accelerometers and vibration rectification error measurements are shown in Figure 3. [5]
42 units 24 units Figure 3. MS1010 (10g sensor), IEEE non-linearity (left) and vibration rectification error (right) 3.2 Noise The noise spectrum performed on an extended frequency range shows in Figure 4 a uniform white noise within the operational range. From this graph, we can extract a white noise of 9 µv/ Hz or 34 µg/ Hz for a 10g sensor and a flicker noise corner at 0.01 Hz. 6 units Figure 4. MS1010 (10g sensor), noise over a full frequency range Figure 5 shows an Allan Variance plot that is in good agreement with the noise spectrum measurement and shows a typical flicker noise corner at 5µg at 100s. [6]
6 units Figure 5. MS1010 (10g sensor), allan variance plot 3.3 Temperature performance The MS1000 has been designed to target tactical grade inertial performance where key parameters are: compensated bias and scale-factor over temperature, and long term repeatability. Temperature performances are evaluated over a temperature cycle of -40 C to 85 C for the bias (K0) and the scale factor (K1). In both case, a 3 rd order polynomial curve is fitted over the parameter models (K0, K1) and resulting residues are extracted. The Figure 6 shows linear behaviour of the bias over the full temperature range with typical slopes of 375µg/ C and residues after 3 rd order polynomial curve fitting below 0.7mg 24 units 24 units Figure 6. MS1010 (10g sensor), Raw Bias (K0) with temperature slope (left) and bias residues (right) [7]
Scale factor over temperature also exhibit excellent uniformity and linear behavior over temperature. The Figure 7 shows a typical scale factor temperature slope of 120ppm/ C and residue of 120ppm. 24 units 24 units Figure 7. MS1010 (10g sensor), Raw scale factor (K1) with temperature slope (left) and scale factor residues (right) 3.4 Long term performance The long term performance of the MS1000 is evaluated using the characterization of the bias and scale factor after applying the environmental conditions found in Table 2: Table 2: Environmental test plan Environment TurnOn / TurnOn Short-term stability Low temperature storage High temperature storage Temperature cycling High Temperature Operating Life Vibration Shock Conditions 100x Powered, min 8h 72h / -55 C 10days / +85 C 100 x / [-40 C ; +125 C] 10 days / +85 C / powered 20grms / 10-2 000Hz 5 x 500g / 0.5ms / 6 directions An initial measurement is done on all sensors to characterise the bias and the scale factor over the temperature and define a unique model for each accelerometer. This reference model is used during the whole campaign to extract the bias and scale factor residue after each intermediate measurement of the long term plan. Figure 8 illustrates the bias and scale factor residues for a sensor calculated from the reference model before starting the ageing [8]
plan. The bias shows a limited hysteresis with a maximum bias residue of 0.3mg and a maximum scale factor residue of 100ppm which highlight the efficiency of the new MEMS design and is die attach process. 1 unit 1 unit Figure 8. Bias and scale factor residue on a sensor extracted from the reference model on the initial measurement. The bias and scale factor are characterized after each step of the ageing plan and the residue is calculated with reference to the initial model. Figure 9 illustrates such measurement for a sensor where the maximum value over the full temperature range has been defined after each ageing condition. The bias hysteresis remain stable (<0.3mg) over the whole plan while the repeatability increase to a value of 0.7mg. 1 unit 1 unit Figure 9. Bias and scale factor residue calculated from the reference model for each ageing steps Figure 10 shows the maximum bias and scale-factor residue at each step of the ageing plan for 24 accelerometers. The high performance and stability of the new MS1000 accelerometer is confirmed with a long-term bias repeatability < 1.2mg for a 10g sensor. [9]
- - - individual part median ɪ 1 sigma 24 units 24 units Figure 10. MS1010 (10g sensor), maximum bias (left) and scale-factor (right) residue over the full long term plan 3.5 Performance summary The MS1000 s overall performances on available range are reviewed in the table below. The family will be complete in the near future with the introduction of the 5g, 30g and 100g Table 1. MS1000 Key parameters Performance, typical values MS1002 MS1010 Unit Full-Scale acceleration ± 2 ± 10 g Residual Bias modeling error 0.14 0.7 mg Long-term Bias repeatability 0.24 1.2 mg In run bias stability 3 15 µg Residual Scale factor modeling error 120 120 ppm Scale Factor Sensitivity 1350 270 mv/g Misalignment 10 10 mrad Resolution (1Hz) 7 34 µg rms Non Linearity (IEEE norm) 0.3 0.3 % FS Operational temperature -40 to +125-40 to +125 C Operating power consumption 10 10 mw Size 9 x 9 9 x 9 mm 2 [10]
4. Conclusion SAFRAN COLIBRYS have more than 20 years of experience in manufacturing accelerometer for industrial, military, aeronautic and safety applications. The MS1000, a new class of high performance MEMS accelerometer specially designed for inertial application, introduce innovative design solutions to meet tactical grade requirements. Excellent thermal behaviour have been reported on a 10g sensor with bias residues of 0.7mg and scale-factor residue of 120ppm. The bias long term repeatability is one of the key parameters for inertial application and the MS1000 shows a stability of 1.2mg. The achieved performance shows the potential of MEMS technology to compete with quartz servo accelerometers. Safran Colibrys has demonstrated is position in the high-end market, ideally positioned to serve demanding applications in terms of performance and high reliability under harsh environments. [11]
5. References [1] P. Zwahlen, D. Balmain, S. Habibi, P. Etter, F. Rudolf, R. Brisson, Open-loop and Closed-loop high-end accelerometer platforms for high demanding applications, Plan 2016, April 11-14 2016, Savannah, USA [2] J-M. Stauffer, O.Dietrich, B. Dutoit, RS9000, a Novel MEMS Accelerometer Family for Mil/Aerospace and Safety Critical Applications, IEEE/ION Position Location and Navigation Symposium (PLANS), May 4-6 2010, Indian Wells, California, USA [3] R. Le Reverend, Device for measuring a force with the aid of a capacitive sensor using charge transfer, US patent US5821421 [4] IEEE std 1293-1998 IEEE Standard Specification Format Guide and Test Procedure for Linear Single-Axis, Nongyroscopic Acclerometers [12]