|back to main page|
Bipolar Drive Signals
This section talks about why bipolar drive signals on the FET gate in a half bridge are useful.
Bipolar Drive Advantages
The main advantage in driving a GDT with a +/-Vcc drive is the reduction of "shoot through", otherwise known as cross conduction.
Shoot Through Reduction
A set of simulations were done in SIMetrix Spice (demo version) to look at shoot through in a half bridge. The basic model is shown in the below schematic.
The model consists of two IRF530 MOSFETs (all that was worth using in the NMOS library in the SIMetrix demo) driven by two pulse sources at a frequency of 100kHz with a duty cycle of 50%. The pulses are from 0V to 12V with a rise time of 50ns...
...and this is the result. The top waveform (green) shows the gate drive waveform and the bottom waveform (red) shows the current through Q1 drain (top MOSFET). The 5A steps are a result of the switching of 50V across the 10 ohm resistor. However, there are large current spikes on every edge, over 30A in some cases, which are caused by shoot through. This is why the spikes occur...
This is a zoom of the two MOSFET drive waveforms as one goes high and the other goes low. The problem is that there is a region where both of the MOSFET gates are above 4V relative to their sources, resulting in them turning partially on - enough to conduct a large pulse of current through both of them.
These spikes will cause large amounts of supply ripple that could affect other circuits, not to mention causing heating in the MOSFETs due to P=I^2R losses. There is also a risk of exceeding the pulsed current rating of the device.
Changing the drive signals from unipolar (0 to 12V) to bipolar (-12V to +12V) means this can be reduced
The above plot shows the two out of phase drive signals, green for the top FET and red for the bottom FET. Notice they now swing from -12V to +12V.
Now look at the current waveform through Q1 (the top FET) with a noticeable improvement in shoot through performance with hardly an spikes visible at all. The reason is below.
The waveforms now cross over at 0V so that one MOSFET is always off when the other begins to turn on - with no need for deadtime! This isn't the end of the subject as leakage inductance can also cause cross conduction.