In the book (and also in the lectures) it is explained that ZCS occurs when (for a full bridge) Q1 and Q4 are on and the current becomes negative (below resonance, current leads voltage, see section 19.3.1). The diodes D1/D4 then automatically take over the currents of Q1/Q4 and there is ample time to turn of the switches.
The part I don't completely understand is why the diodes would take over the current. That may be the case with ideal diodes that have forward voltage Vf=0V. In practice, a threshold voltage is required to forward bias the diodes, so Vf~0.7V. Since MOSFETs are bidirectional devices, they would conduct the negative current and their Ron*Isd would have to be larger than Vf before the diode turns on and takes over (some of) the current. Now suppose Ron*Isd is smaller than Vf, the diode would not conduct and the transistor would carry all the current
For kW type converters, ZCS may be easy to achieve, since probably Ron*I>Vf. But is it also practical to use ZCS in smaller converters?
Any comments would be much appreciated.
Regards,
Toby
4 comments:
My understanding is that, ideally when current goes to 0, turn off Q1/Q4 immediately. Of course, in practice accurate timing is difficult.
Once current crosses zero, Q1/Q4 can be turned off without losses at any time during the interval when current is negative. Diodes D1/D4 turn on as soon as Q1/Q4 (MOSFETs) are turned off. In the case when unidirectional switches are used, e.g. IGBTs or BJTs, diodes D1/D4 start conducting automatically at the zero crossing of the current. Q1/Q4 still need to be turned off before current becomes positive again.
OK, clear. So for bidirectional switches it's not like the current through the switch has to be zero upon turn-off, as the term ZCS suggests. Because the voltage across the switch does not change upon turn-off (besides Vf of the diode), since there is a parallel path in the form of a diode, switch-off is virtually lossless. I guess the term ZCS was invented when only unidirectional switches were used, because in that case the current through the switch really is zero when switching.
All correct Toby.
Post a Comment