Look at this frequency response

here is the openloop gain of a system. I am a little worried about the point A. at this point, the phase is around -160 degree. Does this have any impact on the performance or stability of the system? the system is using 100% feedback.

It is the phase at the zero dB axis crossing that counts. Your circuit has an amplitude slope of one pole at the zero crossing and frequencies nearby so it will be stable.

The root locus drawing will also show stability. There are two poles on the negative real axis then a zero (and a zero at infinity). As the feedback is increased the two poles will go towards each other, break off into a circle to the left and hit the real axis, then each goes to one of the zeros.

What causes instability is poles in the right half plane. Not the 180 degree phase with a gain greater than one. These two items are related most of the time in simple circuits so people get the habit of using the second because it is easier to do.

The Barkhausen criteria is used improperly by many people. It comes from the servo world and states that if there are poles in the right half plane, the system will oscillate at the frequency where the phase shift is 180 degrees.

From this you can see that many people get the cause and effect reversed.

I would be careful of the point with low phase with gain. The phase margin and gain margin look good upfront, but if something adds additional phase change (more poles at higher frequency, for example, or a neglected feedforward zero, or movement of the poles and zeros with temperature), it can become positive feedback with gain, and thus, an oscillator.

You will also have to be careful with your transient response with this type of compensation. Since there are multiple poles and zeros at multiple frequencies, the transient response may have undesirable characteristics.

This does not mean that such a response can not be used, it just means that you need to be more careful to be sure that it will not cause a problem in your system.

fantaci said:

here is the openloop gain of a system. I am a little worried about the point A. at this point, the phase is around -160 degree. Does this have any impact on the performance or stability of the system? the system is using 100% feedback.

Click to expand...

Normally rule of thumb is: feedback=very possible chance for instability.
What is this system? are desciption?

try to get at least 45 degree phase margin. 20 is not enough

fantaci said:

here is the openloop gain of a system. I am a little worried about the point A. at this point, the phase is around -160 degree. Does this have any impact on the performance or stability of the system? the system is using 100% feedback.

Click to expand...

hey fantaci, in your mag response, where is 0 dB in Y axis? Is it the X axis?
I think you d better point it out first. For the cross point of 0 dB and your response
will be your loop bandwidth.

djalli said:

fantaci said:

here is the openloop gain of a system. I am a little worried about the point A. at this point, the phase is around -160 degree. Does this have any impact on the performance or stability of the system? the system is using 100% feedback.

Click to expand...

Normally rule of thumb is: feedback=very possible chance for instability.
What is this system? are desciption?

Click to expand...

My system is a DCDC modulator which works only in DCM mode, so there is only one pole(the second pole) for this modulator. The dominant pole is the compensated pole, and the zero is compensated zero which is used to boost the phase around 0dB bandwith.

Added after 2 minutes:

fehler said:

fantaci said:

here is the openloop gain of a system. I am a little worried about the point A. at this point, the phase is around -160 degree. Does this have any impact on the performance or stability of the system? the system is using 100% feedback.

Click to expand...

hey fantaci, in your mag response, where is 0 dB in Y axis? Is it the X axis?
I think you d better point it out first. For the cross point of 0 dB and your response
will be your loop bandwidth.

Click to expand...

I am sorry, please refer to the following fig.

as to your fig, i give a new one, with root locus, openloop and closed loop
frequency response. i think whether it is stable depends on the phase margin
measured at 0 dB. So, i think your design could be stable.

i think if the phase margin is small when the gain is large, the system is still dangerous. then the negative feedback become positive feedback and the loop gain is large than unity, the system will oscillate.

Your system will exhibit a lightly damped oscillation at a frequency slightly lower than the frequency at which the phase margin is lowest (Point A). The system is, however stable, so the oscillation eventually damps out.
Regards,
Kral