IR2184 MOSFET driver high output damage

Question

everyone

So I'm trying to build a inverter circuit capable of driving most UPS transformers. Thus I'm building a full H-bridge inverter circuit with a 24V power supply being switched at a fundamental frequency of 50 kHz and SPWM generated from a PIC 24 at 15 kHz. My SPWM works as follows: For the first 500 Hz of the 15 kHz the PWM duty cycle stays the same at say 2 % for example and then from 500 Hz to 1000 Hz the duty cycle are increased to 10 %, from 1 to 1.5 kHz the duty increase to 21 % etc. This works for feeding the transformer so that it does not saturate (so no large currents drawn from this : 0.2 A with no load). Then I tried to add a low pass filter to remove the high frequency components at get a 50 Hz sinusoidal waveform of 50 Hz, which I got placing the filter directly after the inverter output, thus no transformer connected. The problem was adding a filter after my transformer caused the voltage to decrease as the current limit started to act. According to me this means that the inductor started to saturate causing the inverter to draw more power as the circuit on the HV side if the transformer started to have a smaller and smaller resistance (impedance).

What I'm unsure of is how this could have damaged my IR2184 MOSFET driver? I had a few of these drivers failing but don't know why. Only the high side output fails giving no output except maybe a bit of static noise, but the low side drive output still functions properly. According to me if the driver fails the one output would always be high and the other always low regardless of the input on pin 1? If anyone could maybe help that would be appreciated Below is my circuit diagram.

Answer 1

Assuming "X" is also ground somehow, you have the HO tied to ground through R90. When you switch high, this output rises to 12 volts above the source, or 36V. When you switch low, the HO output can be pulled down to ground through this resistor. In the data sheet, the "absolute maximum ratings" state that this voltage is only allowed to be 0.3 volts below Vs.

I'm not sure what you are trying to do with "Buck current FB", but the high side gate pull-down resistors should be tied to the source of the high side FET and not ground. R90 and R107 should be between the gate and source of the FET.

Answer 2

First, the line "For the first 500 Hz of the 15 kHz the PWM duty cycle..." confused me. I assume that you mean that the PWM period is (1/15 kHz) 66.7 us and that for the first (500/15000) 2.2 us... As for the blown IC, welcome to the power world, things blow, usually the FETs. Often the issue is transients caused by inductance, either explicit (the transformer) or parasitic (such as wiring). Adding to what winny said, try changing those snubbers such as C40/R104 to much lower, say 10 nF and 10 Ω. Something is probably causing HO (pin 7) to exceed VB (pin 8). Could it be "Buck current FB"? You are using the IR2184 which is the old old DIP package. Not only old (did I mention old?) but no DT (dead time) pin. That can be used to prevent shoot-through, where both Q17 and Q18 are on at the same time and causes transients and/or FET failure. Try to use the SOIC package and keep the wiring short and neat. In the data sheet, look at Figure 9C with a glance at 9A and 9B. Look at grounding. Grounds are not absolutes. Where does the current at "X" go? Currents cause voltage drops, especially in wires. If "X" voltage can be negative relative to "Earth", ICs can be damaged. BTW, the two VB (pin 8) can be connected together and only use one C42/D28 bootstrap. The two VBs better be about the same voltage.