UCC21521: gate driver getting failed during start-up condition

Hi GP,

Can you please use the UCC21550 instead? We resolved a lot of issues based on UCC21520 feedback in the updated design.

The output stage of the UCC21520 is too fast and it can interrupt Igate fast enough to cause Vdd >30V abs max, due to the Vgs loop inductance.

That is the root cause of a lot of failures. It is not possible in the UCC21551, which has two parallel protection circuits against this issue. Also, more quality controls are added in the the wafer production process of UCC21551 (Over Voltage Stress Test) to eliminate units with weak spots in their large oxide layers.

Is it possible to use our updated device instead? 

Best regards,

Sean 

Hi Sean,

Thanks for your response.

Does UCC21550 form, fit, function remains same as UCC21521??. if yes, i can test it in parallel.

Meanwhile, Does any modification in the gate driver schematic will help to counter the current issue??

Introducing optimum gate resistor to slow down Igate ramp up or soft start capacitor to slow down the dv/dt rise ??

If we add Protection circuits externally, does it help ??

Kindly advise..

Hi sean,

Thanks for your advise.

We tried TVS diode SMBJ15A-E3/52 (Stand off Voltage 15V) across Mosfet driver UCC21521 power supply voltage (VDD15V) to clamp the transient voltage that goes beyond 25V.

As expected, TVS diode clamps fast transients if its crosses 25V and above. But, we still unable to avoid UCC21521 failure which happens on random manner. 

We replaced UCC21521 with UCC21551 as potential frontrunner. But, Still failure persists on Low side power supply section(short condition exists b/w VDDB and VSSB) even with TVS diode across power supply lines.

  once again, Need your inputs on the same for trouble shooting the issue.

Hi GP,

The UCC21551 failed with the TVS diode? This is a very robust part. There might be something wrong with the circuit itself. Usually these issues come down to high voltage switch node ringing, and the solution is better high-voltage decoupling.

Does the supply short failure occur on the low-side of the same phase as before? Do you have a measurement of the supply and Vout at the same time? Can you also measure the 150V switch node?

Best regards,

Sean

Hi sean,

As you mentioned, High voltage switch node ringing at mosfet turnoff could be the potential cause.

Does adding Passive Snubber circuit (Series RC network)  across Mosfet drain source will help to bringdown switch node ringing frequency to accepted margin ??

I would like to brief the Mosfet driver failure scenario to narrow down the potential cause.

1. BLDC Motor operation is 2-Quadrant (Forward & reverse) with PI close-loop control 

    V_Drive: 100V to 150V &  I_Drive: 75A (Max)

For Motor forward (Clock-wise) motion control, following are the inputs to Mosfet driver

> Mosfet driver INA voltage : 0 V--> 5 V (Low to high transition with 100% duty) for High side driver output drive with Vdd=15.

                         INB voltage : 0 V--> 5 V (Low to high transition with 50%~100% duty @20Khz) for Low side driver output drive with Vdd=15.

                         Dead time : 3.6uS

For Reverse (Counter clock-wise) motion control,        

> Mosfet driver INA voltage : 0 V--> 5 V (Low to high transition with 50%~100% duty @20Khz) for High side driver output drive with Vdd=15.

                         INB voltage : 0 V--> 5 V (Low to high transition with 100% duty) for Low side driver output drive with Vdd=15.

when Control system is online, BLDC Motor can be operated either forward or reverse direction w.r.to direction input (5V --> Clockwise & 0V--> counter clockwise) 

Consider Mosfet driver actuates Motor in Forward direction w.r.t direction input (5V --> Clockwise), Mosfet driver low side failure occurs if we change motor direction from forward to reverse by Changing direction input.

Kindly advise your feedback on the same..