On 17 December 2014 at 10:50, IVP wrote: > > The gates should have protection against EMF, hash etc from the motor. > A 15V zener paralleled with 270 ohm from gate to source is usually > sufficient. A fast Schottky diode across drain (cathode) - source (anode)= .. > A reverse-series (back-to-back) 15V zeners (suitably rated) across the > motor terminals with 0.1uF ceramics from terminal to grounded case > will suppress a lot of hash at its source. If noise is really bad you may > have to add a couple of ultrafast diodes too. > > FETs are very good at switching power but they can be quite frail > electrically. > > Most of that. The 270 Ohms across the g-s zeners may not be needed - and account needs to be taken of the fact that it also forms a divider with the driver to gate resistors usually specified. I've found that the gate-source zener is extremely effective at keeping FETS alive in applications with inductive loads, when miller capacitance can couple large transients onto the gate and exceed Vgs breakdown voltages. Driver to gate resistor was mentioned by someone as 100 Ohms. This high a value limits max gate drive currents substantially. (eg 10V drive, 100R -> Idrive max < 100 mA.) Gate drive peak currents of more like 1A are often useful. Among other things this deep-ends on PWM speed and gate capacitance. At the 50 kHz PWM rate mentioned if you want gate turn on to be say no more than For PWM in the say 10-100 kHz range I'd suggest R_driver-gate of around 10 Ohms. Low or no driver resistor can lead to nicely fast switching times with increased energy losses and more chance of parasitic oscillation. Too high driver-gate resistors reduce switching times with increase energy losses. Somewhere in between is the mythical best case. Joe's drain-source diode is essentially the same as the traditional 'across an inductive load' diode but with the power supply impedance in the return path. It also improves on the FET "body diode" if the FET is placed in reverse conduction as in overrun situations in some designs. Also 'can be useful' is a small gate-source reverse biased Schottky diode (cathode to gate) mounted close to the FET with minimum length connections to FET. This acts to sharply clip the negative transitions of any gate oscillations at a level below Vgsth of most FETS. Having their negative half cycles clipped is a good way of taking the wind from the sails (or energy from the waveform) of most parasitic oscillations. Contrary opinions or suggested modifications to the above welcome. Russell > --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .