奇怪的问题！关于运放的相位补偿

大大山

原帖由 prof3 于 2009-3-8 15:38 发表

                               
登录/注册后可看大图

1. Please use .pole/zero analysis, and print all the poles and zeros postion before and after compensation.

2. Assume you are using Miller Compensation(neglect high frequency pole)
Before Compensa ...

To Miller compensation,after compensation,I think w2~=gm2/CL. Don't you think so.

prof3

Yes, but that is normal case. I am trying to explain why your P2 is not pushing to higher. That is why I still use gds2 not gm2.

In the case of your op design, if gds2(may be gds of current source or 1/RL, it depends) similar to gm2, then P2 is actually not pushing to high frequency.
In other word, your Miller compensation does not take effects(actually in that case, A is no longer larger than 1, P1 may not be pull to low either).

If your normal compenstation, P2 of course should be ~gm2/(Cc+CL), which is >> gds2/CL ( gm2 >> gds2 most of cases)

adampippo

原帖由 prof3 于 2009-3-8 15:38 发表

                               
登录/注册后可看大图

1. Please use .pole/zero analysis, and print all the poles and zeros postion before and after compensation.

2. Assume you are using Miller Compensation(neglect high frequency pole)
Before Compensa ...

这个说得有道理哦

大大山

原帖由 prof3 于 2009-3-9 01:15 发表

                               
登录/注册后可看大图

Yes, but that is normal case. I am trying to explain why your P2 is not pushing to higher. That is why I still use gds2 not gm2.

In the case of your op design, if gds2(may be gds of current source ...

You are all right,but it is not my answer i want.thank you all the same.

prof3

Assume all you did are correct, please tell me
(1) CL=?, C01=?  C02=?, Av1=? Av2=? Gm1=? Gm2=? Gds1=? Gds2=? Cc=? ( assume RL= inf. )
(2) any other parasitic pole, like at current mirror path ?
(3) Also if possible, show the opearation point

kool

原帖由 大大山 于 2009-3-2 23:06 发表

                               
登录/注册后可看大图

今天仿真出现了一个奇怪的问题。
在简单的二级运放（五管差分输入+共源级输出）中，为了达到更好地相位裕度，在第一级输出与第二级输出之间加了一个补偿电容，按理说，第一个主极点变小，第二个非主极点应该变大，（ ...

你只有了一个补偿电容吗？还是还串联了个电阻，串了电阻的话会引入另一个极点。

大大山

原帖由 prof3 于 2009-3-10 10:07 发表

                               
登录/注册后可看大图

Assume all you did are correct, please tell me
(1) CL=?, C01=?  C02=?, Av1=? Av2=? Gm1=? Gm2=? Gds1=? Gds2=? Cc=? ( assume RL= inf. )
(2) any other parasitic pole, like at current mirror path ?
(3) ...

thank you,the question has been solved.the question is lied on the process paper.

大大山

原帖由 kool 于 2009-3-10 11:01 发表

                               
登录/注册后可看大图

你只有了一个补偿电容吗？还是还串联了个电阻，串了电阻的话会引入另一个极点。

可以用密勒效应一起处理

lipenghbu

刚开始学，感觉补偿很难，读贴收 益良多啊

peter_fu

19# prof3
请问19#:Ahu ja Compensation是什么意思？