Why it should be given to non inverting input????

[techguru]

Hi good evening all,
my doubt is when we take conventional voltage regulators we give the feedback in inverting terminal and series pass transistor is NMOs device but when go for PMOS device(LDO)as series pass device we give feedback to non inverting input? Why we have to give it in non inverting terminal?

                      I went through some books(most of the books does not contain LDO), they said that PMOS (Common source configuration)will give 180 degree phase shift and  we give connect to non inverting input. Why??

why we are worried about phase shift? we are working in DC and Why we are worried about transistor phase shift and Non inverting terminal in operational amplifier???

[rstofer]

In one case (N-channel) you need to raise the gate voltage above the source voltage to turn the device on.  For the P-channel, you need to lower the gate voltage relative to the source to turn it on.  So, the amplified error signal needs to work in the opposite direction.

[David Hess]

The transistor is inverting the output of the operational amplifier so feedback from the output of the transistor goes to the other input.  Any feedback from the output of the operational amplifier itself still goes to the inverting input.

[techguru]

Why i am worried about Phaseshift, when i am working in DC Quantities? What is the relationship between these?

[Audioguru]

Please learn about negative feedback and positive feedback. It can be AC or DC.

[David Hess]

Why i am worried about Phaseshift, when i am working in DC Quantities? What is the relationship between these?

Assuming that you are using the correct input terminal for the feedback, the amplifier still has gain at higher frequencies.  The negative feedback starts at -180 degrees.  In the simple case at high frequencies, the operational amplifier acts like an integrator which adds another -90 degrees.  If the external feedback then adds another -90 degrees yielding -360 degrees before the gain drops below 1, then there will be oscillation.

[techguru]

Still i am not clear. It is clear that opamp gives 180 phase shift(when we give feedback signal to inverting input) and integrating capacitor gives 90 degree. totally 270 degree. at high frequencies it may go up to 360 degree. But in regulated power supply we have only DC Voltages and No frequency components , Why we are worried about frequencies? When we work with DC Quantities when frequency is ZERO? why phase shift is introduced?

[danadak]

why we are worried about phase shift? we are working in DC and Why we are worried about transistor phase shift and Non inverting terminal in operational amplifier???

Systems have noise in them, such that if phase shift were 0 or 360 and there is loop gain
then instability occurs. Think 1 mV of noise at input, gained by 10, it becomes 10mV, now fed
back in phase, now its 10mV at input gained by 10 so output now 100 mV...... and so forth.


http://www.linear.com/solutions/4449


http://www.ecircuitcenter.com/Circuits/opfeedback1/opfeedback1.htm


Regards, Dana.

[David Hess]

Still i am not clear. It is clear that opamp gives 180 phase shift(when we give feedback signal to inverting input) and integrating capacitor gives 90 degree. totally 270 degree. at high frequencies it may go up to 360 degree. But in regulated power supply we have only DC Voltages and No frequency components , Why we are worried about frequencies? When we work with DC Quantities when frequency is ZERO? why phase shift is introduced?

The various parts within the feedback loop do not roll of their high frequency performance instantly.  The integrating characteristic of the operational amplifier is deliberate to provide unconditional stability at any closed loop gain and not all operational amplifiers are designed this way.  An externally compensated operational amplifier can be easily "overcompensated" to make it as slow as one would like.  Integrated power supply controllers use operational transconductance amplifiers which are compensated by an external capacitor to ground.

The frequency response can be further lowered by using a slower operational amplifier, increasing the compensation capacitance, increasing the external integration capacitance, or increasing the output capacitance.  Of course this has to be done while still leaving enough phase margin at a gain of 1 to prevent oscillation.

The cost of having a feedback loop which is slower than necessary is that the output will respond more slowly to load changes.