LDO strange frequency response

[PSG]

Hi,

I'm facing a problem with my LDO: I thought evthg was ok, but then as I ran corners the phase response started to look a bit strange as you can see on the attached plots. At 125C and down to about 60C the phase starts at 0, then quickly goes up to a decent value and then behaves normally (edasim1.gif). If I run at 60C the phase starts at 180deg, we're ok (edasim2.gif). Now as I reduce the temperature further to -40C, you can see the dip in phase response around 1kHz gets lower and lower, down to 30deg (edasim3.gif), not good! Despite that my phase margin is still about 60deg, am I worrying for nothing?
Can anyone explain to me those 2 behaviors.

Also attached are the schematics for the LDO (eda1-4).

Thanks.

 

[PSG]

I forgot to mention that those issues happen with no loading on the LDO output, or very small load (uA). As I increase the loading to 10mA the phase only dips to 45deg or more and that's not a problem anymore.

 

[gevy]

Sometimes Phase Frequency Responce is shifted by -180 deg. for low frequencies.
This is error of AC simulation. You can check stability by Transient siimulation.
Always it is needed additionally to check stability by Transient simulation!