Page 5 of 29

Calibrated mode:

After a certain number of switching events (32) the accuracy is considered to be quite

high. At this time the chip is switched into an averaging mode (= calibrated mode) where

only minor threshold corrections are allowed. In this mode a period of 32 switching

events is taken to find the absolute minimum and maximum within this period. Threshold

calculation is done with these minimum and maximum. A filter algorithm is implemented,

which ensures that the threshold will only be updated, if the adjustment value calculated

shows in the same direction over the last four consecutive periods. Every new

calculated adjustment value that shows in the same direction causes an immediate

update of the threshold value. If the direction of the calculated adjustment value

changes, there must be again four consecutive adjustment values in the same direction

for another update of the threshold value. Additionally there is an activation level

implemented, allowing the threshold to be adjusted only if a certain amount (normally

bigger than 1LSB) of adjustment is calculated. The threshold correction per cycle is

limited to 1 LSB. The purpose of this strategy is to avoid larger offset deviations due to

singular events. Also irregularities of the target wheel are cancelled out, since the

minimum and maximum values are derived over at least one full revolution of the wheel.

The output switching is done at the threshold level without visible hysteresis in order to

achieve maximum accuracy. Nevertheless the chip has some internal protection

mechanisms in order to avoid multiple switching due to noise.

Changing the mode:

Every time after power up the chip is reset into the initial mode. Subsequent modes

(precalibrated, calibrated) are entered consequently as described before. In addition,

two plausibility checks are implemented in order to enable some self-recovery strategy

in case of unexpected events.

First, there is a watchdog, which checks for switching of the sensor at a certain lower

speed limit. If for 12 seconds there is no switching at the output, the chip is reset into the

initial mode.

Second, the IC checks if there is signal activity seen by the digital logic and if there is no

outputswitching at the same time. If the digital circuitry expects that there should have

been 4 switching events and actually no switching has occurred at the output, the IC is

reset into the initial mode.

Reset:

There are several conditions, which can lead to a reset condition. For the IC behaviour

we have to distinguish between a “output hold mode”, a “long reset”, a “short reset” and

a “software reset”.

Output hold mode:

This operating mode means that the output is held in the actual state and there is no

reset on the digital part performed. This state will be released after the IC reaches his

normal operation condition again and goes back into the operating mode he was before.

The following conditions lead to the output hold mode:

• A drop in the supply voltage to a value less than 2.4V but higher than 2.0V for a time

not longer than 1µs .. 2µs.