# two-stage opamp design need help, pls

Status
Not open for further replies.

#### lhlbluesky

##### Banned
i want to design a two stage opamp with cascode compensation, just as the picture below shows. i use the capcitor Cc for compensation between node A and C, my question is, how to choose the value of Cc?

for ex:
resolution:10bit clock:50M swing:2V Vdd:3.3 DR:70dB PM>60deg CL=1PF

then how to design this opamp, what is the value of Cc?

if PM>60, then p2 > 2GBW, but what is the expression of GBW and p2?

i use GBW = gm2 / Cc, p2=gm13 / CL, but it doesn't comply with the simulation results, why?

there are three poles, A\B\C, which is the first pole, which is the second and third pole, why? how to anylysis in detail?

anyone can tell me how to get the expression of p2 and GBW exactly?

besides, if i use another Cc (Cc2) between node D and C, then what is the expression of GBW and p2? and what is the difference of the two case(one CC and two Cc), and which is better for stablity?

pls help me, any advice or help or papers will be appreciated, thanks all.

lhlbluesky said:
i want to design a two stage opamp with cascode compensation, just as the picture below shows. i use the capcitor Cc for compensation between node A and C, my question is, how to choose the value of Cc?

You can read Page305 of "design of analog CMOS integrated circuits" written by Razavi.
this is a more effective compensation technique compared with the cap between B and C.

for ex:
resolution:10bit clock:50M swing:2V Vdd:3.3 DR:70dB PM>60deg CL=1PF

then how to design this opamp, what is the value of Cc?

if PM>60, then p2 > 2GBW, but what is the expression of GBW and p2?

This results is based on the assumption that the zero is 10 times larger than the GMW.it may be wrong here.

i use GBW = gm2 / Cc, p2=gm13 / CL, but it doesn't comply with the simulation results, why?

there are three poles, A\B\C, which is the first pole, which is the second and third pole, why? how to anylysis in detail?

of course A is the first pole.it's the same expresstion with miller compensation.
anyone can tell me how to get the expression of p2 and GBW exactly?

besides, if i use another Cc (Cc2) between node D and C, then what is the expression of GBW and p2? and what is the difference of the two case(one CC and two Cc), and which is better for stablity?

pls help me, any advice or help or papers will be appreciated, thanks all.

if you calculated the results with the expression given by razavi, the results may derived from the simulation results because of the parasite capacitor at B or D.

any different ideas?

I woulded do a two stage, because of compensation and power. I would go for a singel stage folded cascode, you should get the swing of 2V, if needed gain boosting to regain Gain if you lose to much at the corners.

I mean wouldn't sorry

i know that gain-boost can provide large enough gain, but it cannot reach rail-to-rail, so two stage is needed, for two stage, cascode compensation can achieve more bandwidth and no need for compensation resisistor. therefore, i want to design a two-stage opamp with cascode compensation, nad have the above questions. pls help me.

refer the fig no 4-10 from the datasheet of mcp604(rail to rail op-amp).

Hope u should get some idea from it.

• hdqkwr

Points: 2