Take care, MOSFETs are fast! They are a bit slowed down by the gate resistors, which's original purpose is to prevent oscillation. Nevertheless, you will need good power supply decoupling at the BS107 (not shown in the scheme) and, of course, low inductive connections between the micro's and the FETs' grounds!! A solid ground plane is strongly recommended.

The pull-up at the micro is needed, because the micro turns-off its internal strong active pull-up after two oscillator periods, when emitting high state. (See also http://www.8052.com/faqs/121619). The resulting weak active pull-up can present a rather high source impedance, which can make the BS107 become susceptible to noise and glitches. Also, when using a Port0 line for Px,y, with its open drain topology, then a pull-up is obligatorily needed.

4k7 is a good compromise, because lower values can result in a higher output low voltage at Px,y, on the other hand, increasing the risk of turning-on the BS107 due to noise.

If noise is a problem, a cap in the 100pF range can be connected from the gate of BS107 to its source. In any case, the two 4k7 resistors and the BS107 must be mounted as close to the micro as possible.

As long as SOAR isn't an issue, there's no need to insist in ultra fast switching. The relay cannot be turned-on ultra fast, anyway, due to its intrinsic inductance. And turning-off an inductance ultra fast would only result in an ultra high inductive kick voltage. The gate capacitance of IRF540N is about 1.2nF, which gives a turn-on time of roughly 2...3µsec in conjunction with the two 1k resistors. The turn-off time is about the half. This should allow the ultra fast recovery free wheel diode BYV79 to turn-on smoothly without any relevant spike.

Don't forget good cooling. In automative applications a must!


Kai