Analog Devices manufactures a range of micro-machined silicon chip accelerometers. They contain a capacitor with a movable plate acted upon by the accleration, and an equal fixed capacitor, normal to each direction in which they sense acceleration. An on-chip oscillator produces a square wave that compares the capacitances of the movable and reference capacitors, a synchronous detector, and circuits that produce a voltage proportional to the acceleration. There are 3-axis and 2-axis devices. We shall look at a 2-axis (x and y) device, the ADXL322, that has a range of ±2g, whichis well suited to determining the attitude of the chip in space with a constant gravitational acceleration. A data sheet for this device can be downloaded from the link given in the References.
The ADXL322 is provided in a very small surface-mount package only 4 mm square and 1.45 mm thick, a 16-pin LFCSR. It is impossible to work with this package without special tools, but fortunately SparkFun Electronics supplies a breakout board, SEN-00849, on which the chip is mounted together with three 0.1μF capacitors required for its use. This board has six plated-through holes in which wire-wrapping posts may be soldered so that the board can be pushed in a solderless breadboard, and connections easily made. When soldering the posts, avoid making solder bridges. The connections are clearly marked Vcc, GND, X, Y, Z and ST. There is no connection with the Z. This board is also quite small, 18 mm square. The data sheet refers to Vcc as Vs, and GND as COM.
Connect Vcc and GND to a 5V supply. The chip operates from 2.4V to 6.0V, so this is safe although the normal supply seems to be 3V. Connect two DMM's to measure Vx and Vy. Note that the directions of x and y are marked on the board. When the chip is horizontal, the acceleration is zero in the x and y directions, and the DMM's should read something close to Vcc/2. The typical supply current at 5V is 0.70 mA.
Now rotate the board 90° in directions that give ±g in each of the x and y directions. Find the sensitivity in V/g, which should be about 0.750 V/g. The sensitivity is proportional to the supply voltage.
The ST (self-test) connection can be left open if it is not used. When it it connected to Vcc, an electrostatic force will be applied to the movable plates causing a change in output voltage. Try this, and note the output voltages.
Measure Vx when the chip is horizontal (g = 0) and when the x-axis is vertical (g = g0). I found 2.50V and 3.25V in the two cases, so the sensitivity is 0.750 V/g. Using a 30-60 triangle, support the chip at 30° with the y-axis still horizontal, and note Vx. I measured 2.87 V, so 3.25 - 2.87 = 0.37 corresponds to g0 sin θ. Therefore, the angle from the accelerometer data is θ = sin-1 (0.37/0.75) = 29.6°.
Attitude in space is easily determined from measured accelerations. If the chip is rotated through an angle θ about a perpedicular axis from the position of zero acceleration (vertical), the effective acceleration is g = g0 sin θ, so the angle is θ = sin-1(g/g0), as we have used above. The two axes permit the measurement of pitch and roll, as usually termed.
3-axis accelerometers have been used to detect g = 0, which means the object containin the accelerometer is in free fall. This allows the safe preparation of a hard disc drive, for example, before the laptop hits the floor.
A recent article in the American Journal of Physics (see References) describes how an accelerometer was provided with radio remote sensing so that no wires are required to connect it, so that it can be attached to wheels and so forth in the teaching laboratory.
The ADXL322 is very easy to use, and requires a very small power to operate it, so it is easy to apply in nearly any circumstances. It can be powered with a single AA cell with a switching converter to 3 or 5 V.
Visit Analog Devices to get a data sheet for the ADXL322.
The breakout board SEN-00849 can be ordered from Sparkfun Electronics
The use of an accelerometer with XBee 2.4 GHz transceivers providing wireless operation is described in E. Ayars and E. Lai, American Journal of Physics (78), July 2010, pp. 778-781.
Composed by J. B. Calvert
Created 3 July 2010