overshoot undershoot during load transient analysis

[Shilpa poddar]

I have a question on the LDO load transient (P-Pass LDO, internally compensated). According to most of the literature, the voltage undershoot/overshoot with a fast load transient can be calculated as V=(I/(C*BW)), where I is the load variation, C the output capacitance and BW the unity gain frequency of the loop gain. However in my simulation, the voltage droop from no load to full load is much less than what is calculated.i don't understand why it is not coming same according to calculation

 

[deep_sea]

Can you put a link to a reference stating that? Is your dominant pole at the output?

 

[Shilpa poddar]

Can you put a link to a reference stating that? Is your dominant pole at the output?

Dominant pole is at output of error amplifier. I am using miller compensation cap at the output of error amplifier to make it dominant pole for stability.

 

[deep_sea]

The dominant pole (capacitor size) affects mainly the time the LDO voltage to settle but the instantaneous voltage spike is affected mainly by the current peak and the output "resistive part).
Please see the following document.
Toshiba application note

I have simulated with different compensation capacitor values and I can see that the output droop depends to some extent on the capacitance but not heavily. The reason for this dependency I do not know but the settling time is affected directly by the capacitance (large capacitor means longer settling or less GBW).

 

[sutapanaki]

Dominant pole is at output of error amplifier. I am using miller compensation cap at the output of error amplifier to make it dominant pole for stability.

As you see from your attachment, it talks about the recovery of the voltage drop once the feedback of the LDO comes into picture. And the formula does not include only the BW but 3/BW, which is 3 time constants of the loop. So, basically, when load current kicks in, it drops the voltage and the loop is kind of blind for that sudden change. Once the loop wakes up it recovers the voltage and here they say it happens in about 3 tau.