UCC27524A1-Q1: how to judge when the body diode will be on and how to confirm the risk of overshoot or undershoot of current and voltage on the output

Hello Cheng,

chengdp said:

Q1：We have test the current from OUTA of UCC27524A1-Q1. As shown in the following figure,the voltage of the 1ohm resistance was measured of which the value is equal to the output current.The peak value is 0.438A,could you please help to confirm the risk? Thanks!

So that is the output current from OUTA/OUTB? It looks like that is the normal gate turn-on/off current from driving the FET? It looks like a normal turn on and off with almost no reverse current, so I don't see any risk. Is this from the same setup as shown in the other captures?

chengdp said:

Q2：We have test the voltage of pin OUTA and pin OUTB of UCC27524A1-Q1.As shown in the following figure,the peak value is -2.85V when the low side FETs are about to switch.And  it lasts 44ns. 

It breaks the datasheet spec, but, that pulse does not contain very much energy. With these drivers, the output stage is designed to go from pull-down to pull-up and vice versa with no dead-time time. Therefore, the output should never be Hi-Z and conduct through the body diode. 

So in this case, Rol should still be on when that negative pulse happens.  Regardless, the instantaneous power at that time is still V*I, and the integral of that is the energy. You could determine the energy in that pulse using a current measurement, but there is no easy way to say what level of energy or power would damage the device.

chengdp said:

Q3:We have test the voltage of pin OUTA and pin OUTB of UCC27524A1-Q1.As shown in the following figure,the peak value is 20.25V when the high side FETs are about to switch.Could you please help to confirm the risk? We also need to know about when the body diode will be on and conducts the reverse current, how can we judge? Thanks!

This ringing is more of a risk because it has a higher amplitude and duration. You can see that noise is coupling into VDD through the pullup of the device. This is one way to tell that reverse current conduction is happening, although it is hard to tell from the normal discharge from charging the gate of the FET.

To summarize, the body diode does not have to conduct for there to be reverse current, and there is no easy way to tell if current went through the channel or body diode for this device. This noise could damage the driver, and the best way to address it would be to add gate resistance or capacitance. This will reduce the coupled current, and also slow the dv/dt on the switch node that is injecting current in the first place. Diodes can also help shunt the noise away from the driver, but need to be low inductance to help. 

Thanks,

Alex M.