UCC27714: 2-SWITCH FLYBACK TOPOLOGY, LOAD SWITCH PROBLEMS

Hi Eddie,

Thanks for your interest in UCC27714! I'm glad to see your project is close to completion.

Eddie Loy said:

QUESTION 1 : STARTING WITH THE KNOWN PARAMETERS LISTED ABOVE, ARE THEY GOOD NUMBERS? FROM WHAT I GATHERED FROM THE DATASHEET THEY AREREALLY CLOSE.

The HB-HS voltage drops from 16.4V to 12.4V, considerably lower than normal for driving IGBTs (typically we see at least 15V at all times). The gate charge of the selected IGBT is less than 100nC, and since C = Q/V, this suggests the value of the bootstrap capacitor is smaller than expected (around 0.025µF, vs 0.1µF in the schematic diagram). What is the construction and voltage rating of the nominal 0.1µF bootstrap capacitor? For example, a small 0603 ceramic capacitor with 25V rated voltage? The total capacitance of SMT capacitors can shrink with applied voltage, for example see this Murata FAQ.

Please try replacing this capacitor with a larger value and check if Q1 is fully turning on in that configuration.

I believe the above answer should cover question 2 as well.

Eddie Loy said:

QUESTION 3 -- DO YOU HAVE ANY ADVICE, SUGGESTIONS, POSSIBLE SOLLUTIONS THAT WILL HELP ME FINIS PROJECT?

First, I believe that placing FQPF13N06L as S_H is not sufficient. Since the HV source is 60V, the IGBT will be operating very close to breakdown, and any noise or ringing on the switch node could damage the circuit. I recommend a component with breakdown voltage greater than or equal to 100V for S_H.

I also recommend selecting transistors which are optimized for lower current and lower gate charge, to decrease the gate capacitance and allow for rapid turn-on even with a large gate resistor. S_H requires a higher voltage since it is placed directly on the HV source, but S_L operates only up to 18V and could possibly be replaced with a smaller, less costly component.

If the input capacitance of S_H can be reduced to 500pF or less, you may find it advantageous to connect S_L drain to S_H gate directly, and to increase the value of R_SHG to 1kΩ or more. Since the duty cycle is less than 50% at 100kHz, the charge path for the bootstrap capacitor through S_H should have at least 5µs to activate during normal operation, so as long as the time constant of the input capacitance and gate resistance is kept small relative to this 5µs off time, the recharge circuit should behave equivalently. This will also substantially reduce your power supply current requirements, since you will no longer be shorting the 18V supply to ground through 13Ω every cycle. S_H will also be clamped off very quickly by connecting S_L drain to S_H gate directly, preventing a shoot-through path from forming through Q1 and S_H.

Regards,

Hi Eddie,

Most of the changes look reasonable. A few things stood out:

• A tantalum bootstrap capacitor may not be the best idea. The high-side supply will frequently see short noise spikes, especially if there is noise on the HS pin. Tantalum capacitors can be destroyed very quickly by overvoltage conditions, and their destruction can be quite violent compared to other types of capacitor. That said, if you are confident that the voltage rating of this capacitor will not be exceeded even momentarily by noise spikes, you should also double-check the ESR of the tantalum capacitor is low enough that the pulsed currents of the gate driver will not cause significant voltage drops.
• There is typically little harm in choosing a larger bootstrap capacitor value, especially considering the maximum duty cycle of this circuit is <50%. Unless you have a specific reason for keeping the capacitor at 100nF, I recommend stepping it up to 1µF.
• The output clamp diodes are listed as 600V diodes, which is definitely overkill for the HO output. There should be <20V between HO and HB/HS at all times. 600V diodes take up a lot of space, they are usually very slow due to their construction, and their reverse recovery characteristics are typically poor. I recommend these clamp diodes be replaced with the same diodes on the LO output.
• The 1000pF transformer capacitor will tend to pull the output structure up toward the HV supply. If HV sits near earth ground this may not pose any hazard. On the other hand, if the output common is connected to earth ground in some way (such as through an unsuspecting finger), the dv/dt at the transformer primary will cause a high current flow through the 1000pF capacitor which could be dangerous. Usually the ground points of the isolated circuits are connected through the capacitor, since they normally should not have any dv/dt between them.

Best of luck with the design.

Regards,