UCC5390: high frequency noise at the switching on and off

Hi Mohammad, 

Thanks for your interest in our driver, my name is Mamadou Diallo, AE from the High Power Drivers' team.

Can you please share your waveforms as well? Are you observing the overshoot on the both high and low-side FETs. Can you clarify where you are probing in your circuit. Also ensure that you're using differential probes to capture the waveforms from the high-side FETs gate and switch node. Additionally, please ensure you're using "tip and barrel" method to probe the signals.

Few comments on your schematic:

-We do not recommend electrolytic decoupling capacitors on the supply pin of the driver as they tend to have wide capacitance shift over supply range and temperature. For our drivers, please use 2 parallel low-ESR/ESL, surface-mount, multi-layer ceramic capacitors (MLCC) with sufficient voltage ratings, temperature coefficients, and capacitance tolerances for adequate filtering. Generally 0.1uF and 10uF are adequate for best performance from our drivers. These decoupling capacitors must be placed directly across the driver's VCC2 pin.

-Please increase C16 to ~10uF or 20uF at most and remove C21 as 180uF is unnecessarily high. Ensure that C16 and C22 are directly the driver's pin. 

-Please remove VD2 and VD4, these are redundant as the FETs have body diodes already incorporated.

-The FETs should be placed as closed as possible to the driver. In your schematic and layout, I notice you're using R13, R14, R15, R16 and other protection components which is causing a longer than usual turn-on and turn-off for the peak current to the gate. You could reduce that resistor network to a single resistor and similarly reduce R23 R24 into a single resistor to reduce your components count move closer to the driver's OUT pin. Remember, the more PCB trace between the driver's output pin and the FETs, the more stray inductance on the current path to the gate. This causes overshoot at the gate, ringing, and reduction in efficiency of your design.

I have attached an app note below discussing layout guidelines for our gate drivers. Please follow these guidelines, remove and/or adjust the components highlighted above and share the waveforms taken directly at the driver's pins of the following signals: IN+, VCC2, OUT, and GATE for the low-side.

http://www.ti.com/lit/an/slua964/slua964.pdf

Regards,

-Mamadou

how about gate resistance value i was using https://www.infineon.com/dgdl/Infineon-IPX50R190CE-DS-v02_02-EN.pdf?fileId=db3a304339d29c450139d43facbe02fb mosfet.

is 4.7 ohm too low for this? 

for 10 kHz, I am getting a very good response.
but things going noisy when I am increasing frequency.

this one is 10 kHz. with 90-degree phase shift. we can see its ideal and wavefrom looks very clean.

gate signal vs VGS at 10 kHz.

now at 50 kHz. the noise going crazy.

i also changed my two gate turn-on resistor from 4.7 ohms to 22 ohms.
now equivalent gate turn-on resistance is 14 ohm and the waveform looks like.

i have used Aluminum Polymer Capacitor Radial of 180 uF.
i have other board as well were as you suggest 

1. I will remove these and put 10 uF Multilayer Ceramic Capacitors in place of 1 uF.
2. i will reduce the Polymer Capacitor to around 33μF 25V dc Aluminium Polymer Capacitor.

the ESR of these capcaitors  is around 15-20 mOhm. 

Hi Mohammad,

Is this full-bridge still running without an output filter/no load?

Do you have any updates on the performance after implementing the changes listed above?

Thanks,

John

Thanks for your kind reply,

i have connected only 100 ohms and it does not have any filter now.

My application is a Dual Active bridge.

so before working on dab, I was validating my H bridge as a full bridge.

i didn't change the listed above. I am in the process of making other PCB with above changes. 

Hi Mohammad,

I also noticed that you are switching very low voltage across your bridge, using 900V SiC FETs.

These FETs have a non-linear Coss curve, where Coss is very high at lower voltage. This causes your circuit to oscillate more when switching lower voltages. I would recommend working your way up to your desired bus voltage.

Please let me know when you have more results.

Thanks and best regards,

John

I got the capacitance variation with temperature and high-frequency operation from the manufacturer.  the deviation is around -5-10 percent in the whole range.