help for gm-c integrator

[baoer]

gm-c integrator

hi everyone, i got a problem with my integrator, my OTA has the 50db gain and 80 phase margin, but it doesn't work like matlab simulation.



The simulation should above the 0, but it decade down, why????
Does someone help me to solve this problem, plz.

 

[oxide]

cmfb+gm-c

Hi Baoer,
What is your time faktor ie 1/rc or . Have you tried to simulate longer time?
Regards

 

[baoer]

gm c integrator

fs=(gm/C)=0.4E9, input signal 0.3V, signal frequency 10MHz,
.tran 0~3.072u.


Thanks for your help , but i still have no ideal ,why???

 

[oxide]

are you using ideal biasing or does it have any time constant? Have you checked your common mode loop?

 

[baoer]

excuse , common mode loop mean CMFB???do i simulate only CMFB(not with the OTA) block to see the gain and phase???
i have the CT CMFB, but , it work bad, i don't know why?
all transistor is operate in saturation.


that's my Vocm=((vo+)+(vo-))/2, and my Vcm=1.65v
and my bias is not ideal.

thanks again oxide, plez help me.

 

[oxide]

Hi baoer,

I think you should simulate the cmfb loop together with the rest of ota. Try to find a good spot were to open the cmfb loop and analyze(loopgain, phasemrgin). Have you tried to decrease the input/output swing?

Regards,

 

[baoer]

Dear oxide:
this is my OTA with CMFB simulation,that is al i can find the best Vocm.



The following is my only OTA simulation.



The gain and phase margin is almost the same, the Vocm become better, but i can't loak it at the Vcm=1.65.why???but my all Vout=(vout+)-(vout-) shift toward to up of the graphic, so i think if i can solve this problem about cmfb , i will solve the integrator problem.

Why do you want me to decrease the in/output swing??In my knowledge, it is more widely more better, am i wrong???

thanks for your answer.

 

[oxide]

I was just thinking if this is a large signal phenomena. to have a the ota stable at smal signal doesn't necessary mean that you have a stable large signal ota. also i noticed that your input signal freq is quite high compared to unity freq. Try to lower your swing and also the freq??

 

[baoer]

Dear oxide:
try to decrease the signal frequency at 1Mhz , but it become worse,
theoretically the output shoud all above the '0', but it is not ,i also accpet your opinion, the stability in small signal does not mean large signal will stable.
i ask my friend, no one to see the large signal , everyone just tell me they just notice gain and phase, but their IC still work well , why???
now i want to solve the CMFB problem, do you have an ideal??

thanks for your answer, oxide.

 

[oxide]

1. I am afraid I do not understand that your signal "theoretically output should be all over 0"?
2. how do you generate you bias. are they ideal vdc or?
3. I have never used this kind of cmfb.
4. How does you vcc look like?
5. did you check your tb? initial conditions etc..

I am sorry that I could not help you ;-( .
Good luck!

 

[baoer]

i use the matlab to simulate the integrator.

that's why i said , output should over the '0'.

 

[jecyhale]

hi everyone, i got a problem with my integrator, my OTA has the 50db gain and 80 phase margin, but it doesn't work like matlab simulation.



The simulation should above the 0, but it decade down, why????
Does someone help me to solve this problem, plz.

Do you use a coupling capacitor?