www.austriamicrosystems.com/AS5145

Revision 1.10

27 - 36

AS5145

Data Sheet - A p p l i c a t i o n I n f o r m a t i o n

The angular displacement (Q) of the magnetic source with reference to the Hall sensor array may then be modelled by:

± 0.5º

(EQ 3)

The ±0.5º angular error assumes a magnet optimally aligned over the center of the die and is a result of gain mismatch errors of the AS5145.

Placement tolerances of the die within the package are ±0.235mm in X and Y direction, using a reference point of the edge of pin #1 (see Figure

21).

In order to neglect the influence of external disturbing magnetic fields, a robust differential sampling and ratio metric calculation algorithm has

been implemented. The differential sampling of the sine and cosine vectors removes any common mode error due to DC components introduced

by the magnetic source itself or external disturbing magnetic fields. A ratio metric division of the sine and cosine vectors removes the need for an

accurate absolute magnitude of the magnetic field and thus accurate Z-axis alignment of the magnetic source.

The recommended differential input range of the magnetic field strength (B(X1-X2), B(Y1-Y2)) is ±75mT at the surface of the die. In addition to

this range, an additional offset of ±5mT, caused by unwanted external stray fields is allowed. The chip will continue to operate, but with degraded

output linearity, if the signal field strength is outside the recommended range. Too strong magnetic fields will introduce errors due to saturation

effects in the internal preamplifiers. Too weak magnetic fields will introduce errors due to noise becoming more dominant.

9.6 Failure Diagnostics

The AS5145 also offers several diagnostic and failure detection features:

9.6.1 Magnetic Field Strength Diagnosis

By software: the MagINC and MagDEC status bits will both be high when the magnetic field is out of range.

By hardware: Pins #1 (MagINCn) and #2 (MagDECn) are open-drain outputs and will both be turned on (= low with external pull-up resistor)

when the magnetic field is out of range. If only one of the outputs are low, the magnet is either moving towards the chip (MagINCn) or away from

the chip (MagDECn).

9.6.2 Power Supply Failure Detection

By software: If the power supply to the AS5145 is interrupted, the digital data read by the SSI will be all “0”s. Data is only valid, when bit OCF is

high, hence a data stream with all “0”s is invalid. To ensure adequate low levels in the failure case, a pull-down resistor (~10k

Ω) should be added

between pin DIO and VSS at the receiving side.

By hardware: The MagINCn and MagDECn pins are open drain outputs and require external pull-up resistors. In normal operation, these pins

are high ohmic and the outputs are high (see Table 9). In a failure case, either when the magnetic field is out of range of the power supply is

missing, these outputs will become low. To ensure adequate low levels in case of a broken power supply to the AS5145, the pull-up resistors

(~10k

By hardware: PWM output: The PWM output is a constant stream of pulses with 1kHz repetition frequency. In case of power loss, these pulses

are missing.

9.7 Angular Output Tolerances

9.7.1 Accuracy

Accuracy is defined as the error between measured angle and actual angle. It is influenced by several factors:

- The non-linearity of the analog-digital converters

- Internal gain and mismatch errors

- Non-linearity due to misalignment of the magnet

As a sum of all these errors, the accuracy with centered magnet = (Errmax – Errmin)/2 is specified as better than ±0.5 degrees @ 25ºC (see

Figure 23).

Misalignment of the magnet further reduces the accuracy. Figure 22 shows an example of a 3D-graph displaying non-linearity over XY-

misalignment. The center of the square XY-area corresponds to a centered magnet (see dot in the center of the graph). The X- and Y- axis

extends to a misalignment of ±1mm in both directions. The total misalignment area of the graph covers a square of 2x2mm (79x79mil) with a

step size of 100µm.

For each misalignment step, the measurement as shown in Figure 23 is repeated and the accuracy

(Errmax – Errmin)/2 (e.g. 0.25º in Figure 23) is entered as the Z-axis in the 3D-graph.

Y1

Y2

–

()

X1

X2

–

()

------------------------