TPS54821: BOOT-PH UVLO

Hi Mitchell,

What is the input voltage/battery when triggering this issue? According to the waveform, CH3 is PH which high level is only ~5V. In order to output 5V, the device try to operate at 100% duty cycle. But the boot capacitor voltage can't support 100% duty cycle since BOOT-PH UVLO is triggered so that PH pin should be pulled low to recharge the boot capacitor. With the input voltage/battery continue to drop, output will follow to drop.

Below is the description on page 10 of datasheet, 

The device reduces the external component count by integrating the boot recharge circuit. The bias voltage for the integrated high-side MOSFET is supplied by a capacitor between the BOOT and PH pins. The boot capacitor voltage is monitored by a BOOT to PH UVLO (BOOT-PH UVLO) circuit allowing PH pin to be pulled low to recharge the boot capacitor. The device can operate at 100% duty cycle as long as the boot capacitor voltage is higher than the preset BOOT-PH UVLO threshold which is typically 2.1V. The output voltage can be stepped down to as low as the 0.6V voltage reference (Vref).

Thanks,

Lishuang

Hi Lishuang,

Yes, the input voltage is getting very close to 5V.

To clarify the request, what we want to understand is the minimum dropout we can reliably operate our design at so that we can work around it. I had already read in the data sheet, and what is not clear under what conditions the BOOT-PH voltage will drop below that 2.1V threshold.

Thanks,
Mitchell

Hi Mitchell,

For what conditions the BOOT-PH voltage will drop below threshold, honestly speaking, it is really hard to get the answer. When high side fet is on, the current comes from BOOT-PH. With heavy output loading, the more current will be needed. Then BOOT-PH drop will be large. Besides this current, some leakage current will also contribute BOOT-PH drop.

If we can make the duty cycle within 90%, I think it should be okay. So the input voltage should be higher than 5/0.9=5.5V at least. But I still need to double check on bench with full loading 8A situation. Will get back to you no later than COB of next Monday.

Thanks,

Lishuang

Hi Mitchell,

I tested on bench, for 5Vout/8A, we need to guarantee the input voltage higher than 7V which should be more safe.

Thanks,

Lishuang

Hi Lishuang,

Thanks. That’s a very high dropout. What is driving this?

Is more dropout required at higher load?

Are there more details you can share to help us guide how this will impact my customer's design?

Hi Mitchell,

What is the frequency setting for customer?

Actually, within 1MHz, under 8A, once Vin<6.2V, it will trigger BST UV. BST (CH1-Blue)-SW (CH3-Pink) voltage is below 2V which is caused by large on time. the waveform is as below.

Since I only tested one board with one chip and with the comparison RC according to my calculation. So I recommended higher margin here and 7V will be more safe.

Thanks,

Lishuang

Hi Lishuang,

They are using a 71.5k resistor to ground on the RT/CK pin, which I believe gives us a 600kHz switching frequency.

Thanks,
Mitchell