how to simulate the settling time of AMP

hspice manual

Hi all:
how to simualte the settling time of a large gain(100db) 's AMP.My AMP's gain is 104db,i want to simuate it's settling time,but i don't know how to simualte.i guess the AMP must be simulated in the closed-loop,who can tell me,the schematic of the circuit is prefered.
thanks

I think that the gain you mentioned is the open loop gain of the chip. if so, you can use open loop simulation (with DC feedback for proper biasing of the input stage) to see the phase margin which in turn can be translated to settling time. otherwise, you can simulate it with feed back and a pulse input, remember that the simulation must be time domain (transient).

U can not accurately translate Phase margin and Open loop bode plot to settling response of AMP..
It is always recommended to do transient analysis after getting satisfactory ac response..

To perform settling time analysis;
1. put the AMP in a closed loop (choose close loop gain according to ur application)
2. apply the input (diff or single ended) signal (Pulse with certain width: depends on how much settling time u r expecting) to ur AMP so that u will get full output swing.
3. check the output

plz add if i missed something..

yes,you are right.but i want to know what is the really meaning of the output range,all the transitors must be saturated in the range?maybe i can't express my problem clearly in english.now,i deisnged a AMP with continous-time CMFB.and i found the CMFB network can't work properly in all the output range.so i wan to know how i can found the output range in which the CMFB can work properly.and if i use the switch capacitor CMFB replace with it,then i how to simuate the AMP.thanks

opamp741 said:

U can not accurately translate Phase margin and Open loop bode plot to settling response of AMP..
It is always recommended to do transient analysis after getting satisfactory ac response..

To perform settling time analysis;
1. put the AMP in a closed loop (choose close loop gain according to ur application)
2. apply the input (diff or single ended) signal (Pulse with certain width: depends on how much settling time u r expecting) to ur AMP so that u will get full output swing.
3. check the output

plz add if i missed something..

Click to expand...

ya, output range means the voltage range at output that supports all the transitors to be in saturation.
now wht do u mean by "CMFB network can't work properly in all the output range".. is it a speed issue of wht..bcaz switched cap CMFB dont hav any such issue (as far as i know)..
but if u r using any continuous time CMFB then try simulate is seperately to know abt its allowed i/p range..

winsonpku said:

yes,you are right.but i want to know what is the really meaning of the output range,all the transitors must be saturated in the range?maybe i can't express my problem clearly in english.now,i deisnged a AMP with continous-time CMFB.and i found the CMFB network can't work properly in all the output range.so i wan to know how i can found the output range in which the CMFB can work properly.and if i use the switch capacitor CMFB replace with it,then i how to simuate the AMP.thanks

Click to expand...

in my opinion,the output range is the meaning:all the transisotrs work in staturation in the range.but,i want to know how i simulate the output range.especially how to simulate use the hspice.the CMFB now i used is continuous time,it refered in the razavi book.

opamp741 said:

ya, output range means the voltage range at output that supports all the transitors to be in saturation.
now wht do u mean by "CMFB network can't work properly in all the output range".. is it a speed issue of wht..bcaz switched cap CMFB dont hav any such issue (as far as i know)..
but if u r using any continuous time CMFB then try simulate is seperately to know abt its allowed i/p range..

Click to expand...

If you are using a difference differential amplifier as the continuous time CMFB network, then the output range of the amplifier itself is more likely to be limited by the input range of the CMFB (i.e. the range of voltage for all the transistors in CMFB network to be in saturation, more likely to be determined by the source-coupled differential pair). On the other hand, no such problems occur with the switched-cap CMFB network.

the continuous time CMFB now i used is refered in the razavi book,chapter 9,section6.use the deep triode transistor CMFB,for the theroy,the output range will not be limited by the CMFB.i know, the switch capacitor CMFB network is a good chonice for it dosen't limit the output range,but if i want to simulate the amp,then how i do,for the switch-cap is descrete time circuit.

pseudockb said:

If you are using a difference differential amplifier as the continuous time CMFB network, then the output range of the amplifier itself is more likely to be limited by the input range of the CMFB (i.e. the range of voltage for all the transistors in CMFB network to be in saturation, more likely to be determined by the source-coupled differential pair). On the other hand, no such problems occur with the switched-cap CMFB network.

Click to expand...

If you want to simulate the freq response of the opamp with switched-cap CMFB network, you will have to use spectreRF. Please refer to https://www.designers-guide.org/Analysis/sc-filters.pdf for more details.

it maybe takes me much time to study how to use the spectreRF, and the fatal thing is thatut company dosen't has such software.
so i want to know,can you tell me the principle of the switch-cap CMFB. i found so many papers,they all said so:the cs capacitor is used to sample the difference between the ideal common mode output voltage and the vref voltage which is used to bias the tail current source,and capacitor Cc is used to generate the average of the output voltages.but the charge on the capacitors are how distributed on the two phase. how to caculate the CMFB loop's bandwidth and the relationship between the CMFB loop and the main AMP dosen't not express clearly.
so can you tell me the details or give me some papers/books that can describe such things in details.thanks agian. in the attachment,ccm=cc cm=cs.

pseudockb said:

If you want to simulate the freq response of the opamp with switched-cap CMFB network, you will have to use spectreRF. Please refer to h**p://www.designers-guide.org/Analysis/sc-filters.pdf for more details.

Click to expand...