NMOS LDO Stability Problem

[refugee]

Hi,all
For NMOS LDO, we usually thought it intrinsic stability. And I'm sure for high load, the LDO is stability. But how about light load, or no load? For that condition, the output shows a low frequency pole. So, there are two low frequency poles, and it's hard to make good phase margin.
So, do you have any idea to make it stable at that condition?

 

[dick_freebird]

There are operating-point-adaptive decoupling schemes
and there is the option to add a current sink load to
enforce a minimum Zout (though this is unattractive in
battery powered equipment).