See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/267600154

Temperature Effects on the Performance and Reliability of MEMS Gyroscope

Sensors

Conference Paper · January 2009

DOI: 10.1115/InterPACK2009-89370

CITATIONS READS

17 2,027

5 authors, including:

Chandradip Patel Joshua Davis

University of Maryland, College Park Johns Hopkins University
24 PUBLICATIONS   188 CITATIONS    7 PUBLICATIONS   47 CITATIONS   

SEE PROFILE SEE PROFILE

David Lemus
Universidad de las Fuerzas Armadas-ESPE
82 PUBLICATIONS   616 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Solder Reliability at High Temperature View project

Copper wire bonding View project

All content following this page was uploaded by Chandradip Patel on 02 October 2017.

The user has requested enhancement of the downloaded file.

 Proceedings of IPACK2009
InterPACK’09
July 19-23, 2009, San Francisco, California, USA

IPACK2009-89370

TEMPERATURE EFFECTS ON THE PERFORMANCE AND RELIABILITY OF MEMS

GYROSCOPE SENSORS

Chandradip Patel Allen Jones

Department of Mechanical Engineering Department of Mechanical Engineering
A. James Clark School of Engineering A. James Clark School of Engineering
University of Maryland, University of Maryland,
College Park, MD, 20742, USA College Park, MD, 20742, USA

Joshua Davis Patrick McCluskey David Lemus

Department of Mechanical
Engineering
A. James Clark School of
Engineering
University of Maryland,
College Park, MD, 20742, USA
Department of Mechanical
Engineering
A. James Clark School of
Engineering
University of Maryland,
College Park, MD, 20742, USA
Mechanical Engineer
TRX Systems, Inc.
7500 Greenway Center Drive,
Suite 820
Greenbelt, MD, 20770,USA

ABSTRACT and automotive markets for purposes including camcorder &

Elevated temperatures can significantly affect the
performance and reliability of MEMS gyroscope sensors. A
MEMS vibrating resonant gyroscope measures angular velocity
via a displacement measurement which can be on the order on
nanometers. High sensitivity to small changes in displacement
causes the MEMS Gyroscope sensor to be strongly affected by
changes in temperature which can affect the displacement of the
sensor due to thermal expansion and thermomechanical stresses.
Analyzing the effect of temperature on MEMS gyroscope
sensor measurements is essential in mission critical high
temperature applications, such as inertial tracking of the
movement of a fire fighter in a smoke filled indoor environment
where GPS tracking is not possible. In this paper, we will
discuss the development of a test protocol which was used to
assess temperature effects. Both stationary and rotary tests
were performed at room and at elevated temperatures on 10
individual single axis MEMS gyroscope sensors.
INTRODUCTION
MEMS vibrating resonant gyroscopes have the potential to
become one of the largest volume MEMS products in the
coming years [1,2,3]. Because of significant size and cost
reduction compared to macro-scale gyroscopes, applications of
MEMS gyroscopes are increasing everyday. Currently these
devices can be found in consumer electronics, defense, avionics
cell phone stabilization, vehicle stability control and missile
guidance. MEMS gyroscopes can also be used for inertial
navigation [1-6]. One such use is in tracking the movement of
the fire fighter in a smoke filled indoor environment where GPS
tracking is not possible. Here characterizing the performance
and reliability of MEMS gyroscope becomes important as they
are both affected by elevated temperature operation [3-10]. This
paper investigates the effect of temperature on the performance
and reliability of MEMS gyroscope sensors.
Almost all MEMS gyroscope sensors use vibrating
mechanical elements to sense rotation. The operating principle
of a MEMS gyroscope is a measurement of the transfer of
energy between two vibration modes of MEMS comb structures
caused by Coriolis acceleration. The vibrating elements are
made of silicon and the transfer of energy is measured by
capacitive sensing. This allows the devices to have high
resolution with low power consumption. Since these devices
don’t have rotating parts, they can be easily miniaturized and
mass-produced using standard MEMS fabrication processes. In
addition, the MEMS comb-structure is often connected directly
to signal processing circuitry on the silicon chip [4-10]. In the
application of interest in this work, MEMS gyroscopes are used
to track safety personnel in harsh environments, so maintaining
the accuracy of the MEMS gyroscope is critical. In this study,
we examine changes in the noise level and the signal level of

1 Copyright © 20xx by ASME

the angular velocity measurement as a function of temperature
through both rotational and stationary testing in order to ensure
device accuracy. The MEMS gyroscope used in our testing is a
programmable low power single axis gyroscope in an 11 mm ×
11 mm × 5.5 mm, laminate-based land grid array (LGA)
package. The device has an integrated thermal sensor, which
senses temperature of system during operation, and on-chip
signal processing that calibrates the angular velocity
measurement output over the temperature range of −40°C to
+85°C. Measurement of the angular rotation in °/s, applied
voltage (V), applied current (I), temperature in °C, and relative
angular position were accomplished using an evaluation system
test board specific to the component.
Fig.2.Stationary thermal test setup.
EXPERIMENTAL SET-UP AND PROCEDURE
In order to determine the effect of temperature on the noise
Next, rotary tests was performed in which a MEMS
levels and signal levels of a MEMS gyroscope sensor, it is
gyroscope was mounted on an evaluation board affixed to the
required to observe gyroscope performance under stationary
center of a rotating platform that turned at calibrated values of
and rotating conditions. In the stationary test, the MEMS
10, 20 and 40 rotations per minute (rpm). Again baseline noise
gyroscope sensor was mounted on the evaluation board which
and signal levels were established for each rotational velocity at
was affixed to a stationary table. The evaluation board was
room temperature. In this case the evaluation board was
connected to the USB port of a Palmtop which collected the
connected to a Palmtop computer for data collection as shown
data. First, testing was performed at room temperature to
in figure 3. The Palmtop computer was mounted onto the rotary
establish a stationary baseline noise and signal level using the
table using a vertical brace. Care was exercised in placing and
setup shown in figure 1. Each test was conducted until 10,000
mounting the laptop to avoid vibration during table rotation.
data points were acquired, and the tests were repeated on 10
The test was conducted until 10,000 data points were acquired,
different gyroscope samples. Then the sensor was then
and again the test was repeated on 10 different samples. After
subjected to five thermal cycles in air from 25ºC to 85ºC. In
this the gyroscope sensor was subjected to thermal cycling in air
order to achieve the thermal cycling conditions, a Temptronic
at 10, 20 and 40 rpm. Each velocity was tested for a number of
thermostream air system was used. This apparatus has the
thermal cycles, directly exposing the MEMS sensors to Class I
capability to supply constant temperature flowing air at
and II+ temperatures recommended by NIST and the National
temperatures ranging from -200ºC to 200ºC. The gyroscope was
Fire Protection Agency (NFPA) [11]. This included five thermal
placed under and at the center of the glass hood of Temptronic
cycles each from 25ºC to 85ºC, from 25ºC to 125ºC , from 25ºC
Thermostream apparatus such that it was directly exposed to hot
to 150ºC and from 25ºC to 175ºC. The test setup for rotary
air flowing from the nozzle at 5 standard cubic feet per minute
thermal test was identical to that for rotary baseline test, except
(SCFM) as shown in figure 2.
that the entire rotating platform was placed under the glass hood
of the Temptronic Thermostream apparatus such that the
gyroscope was centered directly under the hot air flowing from
the nozzle at 5 SCFM.

Fig.1.Stationary baseline test setup.

Fig.3.Rotary baseline test setup.

2 Copyright © 20xx by ASME

 Again, no permanent effects were seen in subsequent baseline
testing shown in figure-9.

Fig.4. Rotary thermal test setup.

Fig.5. (a) First stationary baseline test result; (b) Mean of
RESULTS AND ANALYSIS angular rate data with windowsize = 500.
All MEMS Gyroscope sensors were analyzed at both stationary
and rotary tests conditions. For both conditions the baseline test
was conducted first, followed by the thermal test. The primary
purpose of conducting baseline test was to observe the device
characteristic before temperature cycling to allow for later
comparison.

Stationary test result:

A first stationary baseline testing was conducted to observe
the gyroscope sensor output at room temperature. The top plot
in Figure-5 shows sensor output for the first 10,000 angular
velocity measurements. The scattering of data was easily
analyzed by grouping each set of 500 data points into a single
point and plotting the raw output as shown in bottom plot of
figure-5. The selection of 500 data points as the “window size”
or the reduced level of resolution was driven by a tradeoff Fig.6. (a) Stationary thermal test results from 25°C to 85°C; (b)
between noise reduction and loss of signal. After the first Mean of angular rate data with windowsize = 500; (c) Thermal
stationary baseline test, the sensor was then subjected to five cycles from 25°C to 85°C.
thermal cycles in air from 25ºC to 85ºC. Figure-6 includes three
different plots representing raw angular velocity data, the mean
of angular velocity with window size 500 and a temperature
cycle plot. The raw data plot when compared with the
temperature plot indicates that mean angular velocity drifting
higher with temperature. This is verified from the mean of
angular velocity plot with window size 500. This effect appears
only when temperature reaches its peak value. The Noise level
returns to its normal level when the temperature returns to room
temperature with no apparent decrease in durability as shown in
subsequent baseline testing plotted in Figure-7.
Air temperature was then further elevated, as the sensor
was subjected to air cycles form 25ºC to 125ºC as shown in
figure-8. Here, we observed an even grater increase in mean
rate level which again returned to normal value with
temperature dropping below 85ºC. This drift does not Fig.7. (a) Second stationary baseline test result; (b) Mean of
correspond to any actual motion as the sensor is stationary. angular rate data with windowsize = 500.

3 Copyright © 20xx by ASME


Fig.8. (a) Stationary thermal test results from 25°C to 125°C;
(b) Mean of angular rate data with windowsize = 500; (c)
Thermal cycles from 25°C to 125°C.
Rotary test result:
The rotary test conditions were similar to the stationary test
condition, except the senor was also exposed to two further
elevated temperature cycles from 25ºC to 150ºC and 25ºC to
175ºC. All 10 MEMS Gyroscope sensors were tested at 10, 20
and 40 rpm with 5 different thermal cycles, but for simplicity
only sensor performance at 10 rpm with 5 different thermal
cycles discussed.
In the rotary test, a baseline test was first conducted to
observe the gyroscope sensor output at room temperature
condition as shown in figure-10. This figure shows raw angular
velocity data and a mean of angular rate plot with window size
500. After the first baseline test, the sensor was subjected to five
thermal cycles in air from 25ºC to 85ºC as shown in figure 11.
There was a slight drift in the angular velocity as observed from
the mean of angular rate plot with window size 500 in same
figure. Again a baseline test was performed. The air temperature
was elevated and thermal cycle was performed from 25ºC to
125ºC as shown in figure-13. Here, the angular velocity as was
increased monotonically with temperature without any
corresponding change in the actual angular velocity of the turn
table. This was also true for thermal cycling from 25ºC to
150ºC as shown in figure-15. This is due to strain on the sensor
which alter its output. When the sensor was exposed to thermal
cycles from 25ºC to 175ºC, it was observed that the device
ceased to function as shown in figure-17 that again returned to
normal when the temperature dropped back to room
temperature shoeing no permanent damage.

Fig.9. (a) Third stationary baseline test result; (b) Mean of

angular rate data with windowsize = 500.

The statistical values for stationary baseline tests are mentioned

in table-1.

SL Mean Standard
Tests
No (°/s) Deviation (°/s)
1 First Stationary baseline test 0.69 0.44
Fig.10. (a) First rotary baseline test result; (b) Mean of angular
2 Second Stationary baseline test 0.55 0.45 rate data with windowsize = 500.

3 Third Stationary baseline test 0.43 0.44

Table.1.Stationary baseline test data summary.

4 Copyright © 20xx by ASME

Fig.11. (a) Rotary thermal test results from 25°C to 85°C; (b)
Fig.14. (a) Rotary thermal test results from 25°C to 175°C; (b)
Mean of angular rate data with windowsize = 500; (c) Thermal
Mean of angular rate data with windowsize = 500; (c) Thermal
cycles from 25°C to 85°C.
cycles from 25°C to 175°C.

SL Mean Standard
Tests
No (°/s) Deviation (°/s)
1 First Rotary baseline test 59.87 0.84

2 Second Rotary baseline test 59.98 1.49

3 Third Rotary baseline test 59.68 0.82

4 Fourth Rotary baseline test 59.99 1.27

5 Fifth Rotary baseline test 59.79 1.28

Fig.12. (a) Rotary thermal test results from 25°C to 125°C; (b)
Mean of angular rate data with windowsize = 500; (c) Thermal
cycles from 25°C to 125°C.
Table.2. Rotary baseline test data summary.
SUMMARY AND CONCLUSION
The following have resulted from this study:

1. A new test procedure has been developed for evaluating the

effects of temperature on MEMS gyroscopes operated beyond
the manufacturer’s recommended temperature range.

2. There was a monotonic increase in the angular velocity

measured by the gyroscope with temperature when the
gyroscope was operated above 85°C during rotary testing, that
did not correspond to any actual change in the angular velocity
of the table. This increase was a temperature related drift in
performance which must be accounted for in using the angular
velocity measurements for tracking.

3. There was an increase in the noise levels when the gyroscope

Fig.13. (a) Rotary thermal test results from 25°C to 150°C; (b) was operated at temperatures above 85°C during stationary
Mean of angular rate data with windowsize = 500; (c) Thermal temperature cycling. The noise level returned to normal when
cycles from 25°C to 150°C. the temperature dropped below 85°C with no apparent decrease
in durability as a result of the effect.

5 Copyright © 20xx by ASME

4. When the MEMS gyroscope was subjected to 175ºC, the
gyroscopes cease to function.

ACKNOWLEDGMENT
The authors would like to acknowledge the financial and
technical support of the Maryland Industrial Partnership
Program and TRX Systems, Inc. The authors would also like to
thank the CALCE Electronic Products and Systems Center for
the use of their laboratories in the conduct of this work.

REFERENCES
[1] Steven Nasiri, "A Critical Review of MEMS Gyroscopes
Technology and Commercialization Status," Inven Sense.
[2] B. Stark, “MEMS Reliability Assurance Guidelines for
Space Applications,” JPL internal document, JPL D-16460,
October 30, 1998
[3] Shcheglov,K.,etal.,"Temperature dependent characteristics
of the JPL silicon MEMS gyroscope, Proceedings 2000
IEEE Aerospace Conference, Big Sky, MT, USA, March
18-25, 2000, pp. 403-411.
[4] Jin-Won Joo and Sung-Hoon Choa, "Deformation Behavior
of MEMS Gyroscope Sensor Package Subjected to
Temperature Changes," IEEE Transactions on Components
and Packaging Technologies, VOL.30,No.2, June 2007
[5] P. Monajemi, P. J. Joseph, P. A. Kohl, and F. Ayazi, "A Low
Cost Wafer-Level MEMS Packaging Technology,"
International Conference on Microelectromechanical
Systems (MEMS 2005), pp. 634-637, 2005.
[6] Z.Szűcs, G.Nagy, S.Hodossy, M.Rencz, A.Poppe,
”Vibration Combined High Temperature Cycle Tests for
Capacitive MEMS Accelerometers”, Therminic 2006,
September 2006
[7] P. Qi Lin; Stern, "Analysis of a correlation for thermal
noise reduction in a mems gyroscope." In System Theory,
2002. Proceedings of the Thirty-Fourth Southeastern
Symposium on, pages 197{203, 2002.
[8] L.Felton, N.Hablutzel, W.Webster, K. Harney, “Chip scale
packaging of a MEMS accelerometer.” Electronics
Components and Technology Conference, vol. 1, pp. 869-
873, June 2004.
[9] M.I.Ferguson, D.Keymeulen, C.Peay, K.Yee, D.Li, “Effect
of Temperature on MEMS Vibratory Rate Gyroscope”, in
Proceedings of the IEEE Aerospace Conference, Big Sky,
March 2005.
[10] L.F.Wu, F.X.Zhang, “Effect of Temperature on the
Performance of a Silicon Micro-machined Gyroscope,”
Chinese Control and Decision Conference, 2008
[11] Donnelly M. K., Davis W. D., Lawson, J. R., and Selepak,
M. S., “Thermal Environment for Electronic Equipment
used by First Responders,” National Institute of Standards
and Technology, Gaithersburg, MD, NIST Technical Note
1474; January 2006.

6 Copyright © 20xx by ASME

View publication stats