LDO stability with package parasitic inductance

[vovan76]

I'm LDO designer ,and all stability simulation were with good PM ,but when i add package parasitcal inductance LDO's PM go crasy -LDO not stable at all
Did anybody meet this kind of problem ?

 

[kspalla]

Check if by increase in the bond wires or number of pads will help

 

[vovan76]

I don't have possibilty increase number of wire bonds

 

[LvW]

I'm LDO designer ,and all stability simulation were with good PM ,but when i add package parasitcal inductance LDO's PM go crasy -LDO not stable at all
Did anybody meet this kind of problem ?

This seems to be the classical case for the need of any over-compensation.

 

[vovan76]

"This seems to be the classical case for the need of any over-compensation."

Can you explain ,what do you mean?
I get very low phase margin

 

[LvW]

"This seems to be the classical case for the need of any over-compensation."

Over-compensation means to compensate the loop gain frequency response to a greater extent as derived from the classical BODE plot. Thus, you have "room" for some external unknown parasitic influences which are not covered during simulation.
In your case (if you cannot avoid the mentioned parasitics): Try to modify the (theoretical) design with the aim to have about 80..90 deg phase margin. Then, in reality (including inductive parasitics) you hopefully have enough stability margin.

I know, it sounds more simple as it will be in practice.

 

[dhasmana]

In your case , probably the bondwire inductance is introducing a zero well within the UGB of the LDO and which is enhancing the gain.

 

[jwfan]

What's the inductor value you are using for simulation? if you want accuracy, you may need some resistor in serias with this parasitic inductors. Maybe several mohms.