LMZ10501 pre-biased output behavior

Hello Gilad, 

Yes, the LMZ10501 will sink current during startup. Here are some more details.

The LMZ10501 will try to equalize the feedback voltage to the VCONxGAIN voltage while VCON is ramping up and will turn the low side FET on to discharge VOUT. This discharge is current limited through the low side FET. The startup is relatively fast (usually less than 100us). In several 10s of microseconds the VCON voltage will be already high enough and the partially discharged output will match VCONx2.5V/V.  At that point the VOUT voltage will continue its normal ramp up. You can see this behavior on an EVM if you toggle the enable pin ON, OFF, and back ON with no load at the output to discharge the output capacitance. When you re-enable the regulator, you will see a dip in the output voltage for a few microseconds. 

Regards, 

Denislav

Hello Gilad, 

Thinking about this a bit more, you can make the LMZ10501 to not discharge the prebiased output if you speed up the VCON voltage ramp up time. Here is an example:

This is the startup with 1000pF on VCON. There is a period of time when the output is discharged, note the negative inductor current (CH4) and the dip in VOUT (CH1):

Here is the startup with 470pF on VCON, showing no discharge on VOUT.

The idea is to make the time constant on the VCON voltage lower than ~50us so that VCON is up to its final value once the startup sequence is complete. 

Regards, 

Denislav

I'm thinking of enabling all the various power rails together. That way, if the voltages rise more or less together, there won't be much sinking.

However, I'd like to avoid this since I'm not sure if the power rails rise time closely match one another (they are not all based on LMZ10501).

Your idea seems better, but won't reducing the capacitor on the VCON change the transient current response of the device or harm the PSRR response?