Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Strange problem about compensation

Status
Not open for further replies.

xiexi

Member level 4
Joined
Jan 14, 2005
Messages
79
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,286
Activity points
613
The miller cap makes A node as dominant pole, I expect that adding 10pF to A node will improve the dominant pole, but in fact the stablility is worse, why?
 

maybe the position of the pole shift. morever, the compensation method of 10pF is not recommented!
 

xiexi said:
The miller cap makes A node as dominant pole, I expect that adding 10pF to A node will improve the dominant pole, but in fact the stablility is worse, why?
hi,
i think the non-dominant pole is get smaller (reduce to about 1/(1+10/4) of the original)when 10pf is added.

good luck
jeff
 

You are probably introducing one more pole (by adding the 10 pF cap) to the system.
 

I think the Rc and Cc forms a LHP zero to cancel the 2nd pole (-gm5/C2) where C2 is the cap seeing from Vout node. If that is the case, original 3rd pole become the 2nd dominate pole (-1/Rc/C1) where C1 is the cap seeing from node A. Once you increase C1, the phase margain become small. Re-size Cc and Rc to make the time constant the same and increase Cc, decrease Rc.
 

Maybe because the 10pF is too large, and at a high frequency it act as "short" the gate of M5 to GND, and "shut off" M5. So the output resistance of the output stage became larger, and makes stability worse.



(BTW, the former message i post has misunderstood your meaning. You means add, but not replace)
 

Why"can not make any change to the non-dominate pole"?
Is the 2nd pole can't form Cout/gm2?
 

Miller compensation produces pole-splitting, then increasing phase marging.

The addition of the other capacitor only makes the two parasitic poles to move left without producing pole splitting. The first non-dominant pole is:
p2=-gmout.Cm/(C1.CL+(C1+CL).Cm)

The second non-dominant pole is:
p3=-1/(Rm.C1)

with Cm= miller capacitor, Rm miller resistor, C1 capacitor in parallel with gate of output transistor, and CL the capacitor connected at the output.
 

there is another non-dominant pole @ w=1/(Rc*C1) where Rc is the miller resistor and C1 is the capacitance at node A (gate M5), so adding a cap. at node A will increase the non-dominant pole and hence PM goes Bad, so either increase the Miller cap. or just use the cap. at node A but not both simultaneously
refer to gray for more details
 

xiexi said:
Why"can not make any change to the non-dominate pole"?
Is the 2nd pole can't form Cout/gm2?

It is actually gm5 / (Cout + CA).

Increasing CA is like increasing Cout.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top