TPS54386: TPS54386 EMC problems - 50 MHz harmonic

Hi Bruno,

Noise at this frequency suggests it is coming from the SW node rising/falling edges. Luckily you have included place holders for BOOT resistor in series with the BOOT capacitor to slow down the edges. Something you can try is to replace the 0 Ω resistor with a 1 Ω to 3 Ω resistor.

Since you are trying to keep it to a component change only, I suggest also trying different catch diodes. Different schottky diodes have different reverse recovery behavior and could contribute to more radiated noise from the buck converter.

Best Regards,
Anthony

Hi Anthony,

Thank you so much for your answer.

We made some test increasing RBoot and found a little decrease on EMI when we use 8.2 ohms. Could you tell me the consecuences of increasing this value? Do you think will affect the regulation? Device is battery powered from 12v lead-acid battery or 18v external power supply.

We will change the diode in order to improve recovery time.

Best regards,

Bruno

Hi Bruno,

For now I would be hesitant to go beyond the 3 Ω recommendation on the datasheet. I was not involved with putting this limit into the datasheet but I'm sure it's there for a reason. It will be hard for me to say exactly what the consequence of going higher will be but it would be something like the following. The larger BOOT resistor will cause a larger drop in the BOOT voltage at turn on of the high-side MOSFET. A large enough drop has potential to cause glitches in the control circuit which would result in unstable like behavior at the switching node.

Instead of increasing the BOOT resistance above 3 Ω, you could try increasing the strength of the SW node snubber. This is done by increasing the capacitance in the snubber from 470 pF. You may need to tune the resistor too to better optimize the snubber. You can either use the EMI measurement or a SW node measurement to judge which value works best.

There is another thing to try but this would result in a schematic and PCB change. I didn't make this recommendation initially because I know you would like to avoid this. If you add a lower value ceramic capacitor next to each PVDD pin it could better bypass the 50 MHz to help reduce the noise at this frequency. The ceramic capacitor should be chosen so it's resonance frequency is near the frequency of the noise you want to filter (50 MHz). I checked muRata's website and a 18nF-22nF capacitor should have a resonance frequency near 50 MHz. These capacitors should be in addition to what's already there.

Best Regards,
Anthony

Hi Bruno,

Did you end up finding a solution to this?

Anthony