TPS54620: Add diode from Vo to BOOT to increase maximum duty cycle in low drop out mode

Wenhao Wu 

1.The maximum Duty cycle is 1-Fsw,min*Toff,min. I checked datasheet the minimum Toff is 0ns, so this mean TPS54620 can work in 100% duty cycle. Please confirm;

On a switching frequency cycle by cycle basis, the TPS54620 can operate at 100% duty cycle, but there is a critical piece of information in the datasheet "Test Conditions" 

There is a BOOT to PHASE (Switching node) UVLO circuit inside the TPS54620 that forces the high-side FET off and the low-side FET ON to recharge the boot-strap capacitor.  Adding a VOUT to BOOT diode with less drop-out and less forward resistance than the internal diode will help increase the time between these forced switching cycles, as long as the diode does not leak reverse current from BOOT to VOUT during 100% duty cycle conduction.

Note:  If a BOOT resistor is used, the diode should be connected to the BOOT Capacitor side of the resistor so that the resistor is not part of the Boot capacitor recharge path.

2.Is that ok to add a diode here even though it does not work for TPS54620? Customer wants to have bom to bom so if we are ok to add diode it will be good.

As long as the output voltage is less than 7V, it would be ok to add the diode from VOUT to BOOT.

3.What is other risk point here?

The primary risk of providing an external diode to recharge BOOT is over-charging BOOT, especially when the PH voltage is negative due to the body diode conduction of the low-side FET.  With a 5V output this should not be an issue.  I the diode was sourced from VIN instead of VOUT, the total BOOT to PH voltage might get too high during body diode conduction.

Feng Wu 

1) Each switching cycle there is a small period where both the high-side and low-side FETs are OFF and the inductor current continues to flow out of the switching node, during this time the body-diode of the low-side FET conducts current from its source (Ground) to its drain (PH) which forces the PH voltage negative with respect to ground.  If there is a diode from a voltage greater than 7V feeding BOOT when PH becomes negative, the BOOT to PH absolute maximum rating will be exceeded and that could pose a risk of damage to the TPS54620.

That is why I prefaced my answer with the qualifying statement that "as long as VOUT < 7V, this is ok"

2) It looks like the text from the datasheet did not post correctly in my original response.  The TPS54620 has an internal BOOT to PH UVLO circuit, when the boot-strap capacitor voltage drops below that UVLO voltage (2.1V typically) the high-side FET will turn-off and the low-side FET will turn on for the remaining of 1 switching cycle.  So the converter is running at 100% duty cycle for a large number of switching cycles, but periodically recharging the boot-strap capacitor.  The effective switching frequency, and the effective duty cycle, depend on how strongly the BOOT capacitor is recharge, which the external diode helps, and how much leakage and bias current is drawn from BOOT during the ON-time, which is why a high-quality diode with very low reverse leakage should be used to ensure the highest possible duty cycle.

Thanks Peter, fully understand them from your detailed reply.