The apparently simple circuit that Joe posted is so useful that it merits some rambling on re: On 7 January 2014 20:02, IVP wrote: > > Zener gate to ground:........ I've seen it seldom elsewhere >> > > I notice that Silicon Chip are using it lately, and push-pull drive > > Attached is what I'd typically use between a PIC and a FET > > Looks familiar :-). That driver works very well and simpler you cannot get. > Q1 and Q2 can be ordinary BJT, eg BC337, BC327, or low Vsat > > fast-switching high-current like Zetex ZTX450, ZTX550. Low Vsat usually not too important here as long as you are not marginal for drive voltage. BC337/327 TO92 or the BC817/ 807 SOT23 direct equivalent work well here. I use these as my jellybean bipolars of choice as they have better drive specs than almost any other common type and usually cost as little as most. Buy 100 (or more) of each when you find them at a good price. (Or about 1 cent in Asia in modest production volumes). Buy the BCXXX-40 parts (average beta of 400 (range 250-600)) as they usually cost no more and you end up with a really useful jellybean part (500 mA+, good beta, ...). R3 limits peak current to gate and actually reduces switching losses which can occur if gate is switched too fast !. R3 also lightly 'decouples' gate from the stiffish driver pair and allows the zener to deal with high voltage pulses couple from drain by Miller capacitance. Without R3 the transistor pair would get involved with large gate transients but I'm not sure how benign this would be. There is a ~=3D 2 x Vbe dead band in the drive current as Vin transitions past Vgate. This seems to be [tm] enough to make this arrangement resistant to serious shoot-through effects that may otherwise be expected with two drivers and not much else connected across the supply. Vgate will drive only to within about 1 Vbe of either rail - this tends to matter only when drive voltage is marginal for the FET concerned. [eg using the fabulous CES2310 low Vgsth N Channel MOSFET (30V and a few amps rated) an MC34063 SMPS IC running at 3V will (just) provide enough drive using this circuit]. [With MC34063 drive transistor bases directly from IC with about 1k pulldown as output is active pullup / float low.] > .... so R2 can be anything over 270 ohms, keeping in mind the PIC pin's 20mA Both transistors are operating as emitter followers so R2 can be as low as desired if there are reasons that lower is desired. For most FETs I use ZD1 =3D 12V as Joe says. It can be any value high enoug= h above the drive voltage as to not conduct in normal operation. In the case of FETs with very low Vgs max (usually ones with super low Vgsth and so very thin gate oxide) a lower voltage zener may be wise to keep Vgate max under spike conditions well clear of Vgsmax. For extra points (seldom needed) connect a reverse biased Schottky diode across the zener. Zener (and Schottky if used) should be as physically close to FET g & s as possible. If Murphy decides that the gate should experience parasitic oscillation (as happens when it really really mustn't [go around power, please ! - oscillate] the Schottky will clamp the negative gate excursions and make the gate oscillation very asymmetric and dissipate its energy. For extra extra points on TO220 or similar packages, slide a ferrite bead of characteristics that feel good onto the drain lead. Knocks the edges off high frequency ringing. May knock the edges off your signal in high enough frequency cases (rare). You could do this with eg D-Pak pkgs if you arranged to use the truncated middle pin as drain connection and di not use the tab electrically, but takes special effort. Russell Miller effect: http://en.wikipedia.org/wiki/Miller_effect Images: https://www.google.co.nz/search?q=3Dmiller+capacitance&tbm=3Disch&tbo =3Du&source=3Duniv&sa=3DX&ei=3D8s_LUrilO8KVlQWuk4DoAQ&sqi=3D2&ved=3D0CDcQsA= Q&biw=3D1920& bih=3D1074 --=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 .