How to simulate a SC-CMFB fully diffrential OTA with Hspice

[8k-rom]

ac analysis for sc with cmfb

I have designed a folded cascode fully diffrential OTA, and the SC-CMFB works well.
But when I use Hspice to do AC analysis, the open loop gain is so small
In the waveform,we can found that the Av(0) equal to -20db.

Someone told me that Hspice is not suitable for this kind of circuit,and they usually
do it by spectreRF .

I think when AC analysis, the circuit was not in suitable state.In other word,some transistors were not in saturation region when AC analysis....

How to do? Who can tell me??

Thanks in advance.

 

[jeff_zx]

hspice 2007 and sc

I have designed a folded cascode fully diffrential OTA, and the SC-CMFB works well.
But when I use Hspice to do AC analysis, the open loop gain is so small
In the waveform,we can found that the Av(0) equal to -20db.

Someone told me that Hspice is not suitable for this kind of circuit,and they usually
do it by spectreRF .

I think when AC analysis, the circuit was not in suitable state.In other word,some transistors were not in saturation region when AC analysis....

How to do? Who can tell me??

Thanks in advance.

Hsipe can not do ac analysis with SC. u can use ideal model to replace the sccmfb to do ac analysi.

 

[8k-rom]

Re: How to simulate a SC-CMFB fully diffrential OTA with Hsp

Hsipe can not do ac analysis with SC. u can use ideal model to replace the sccmfb to do ac analysi.

What kind of ideal model can be used? Where can i find some examples about it ?
thx~~!

 

[bbbb]

another way is to place voltage sources in series with Caps.Adjust these voltages for correct bias point. remove the switches and clocks.

 

[jonashat]

You can't do AC analysis in the presence of any switching activity. Simply because the AC analysis starts by solving the DC point and then linearizes the whole circuit around this point.

Since you have a lot of switching activity, the DC point is changing over time and hence the AC analysis has no meaning.

What I suggest you to do is to replace the SC circuit and add VCVS to do the function of the CMFB (average the output nodes, subtract from a reference and feedback to the OTA).

 

[8k-rom]

Re: How to simulate a SC-CMFB fully diffrential OTA with Hsp

You can't do AC analysis in the presence of any switching activity. Simply because the AC analysis starts by solving the DC point and then linearizes the whole circuit around this point.

Since you have a lot of switching activity, the DC point is changing over time and hence the AC analysis has no meaning.

What I suggest you to do is to replace the SC circuit and add VCVS to do the function of the CMFB (average the output nodes, subtract from a reference and feedback to the OTA).

thx a lot~~!!
I will try~

 

[walkingsun]

Re: How to simulate a SC-CMFB fully diffrential OTA with Hsp

it's very clear

 

[Hunter_BJUT]

you can also use PSS and PAC to do the job.

 

[hr_rezaee]

Message is unavailable.

 

[morix]

Re: How to simulate a SC-CMFB fully diffrential OTA with Hsp

the way i will do to analyze the ac simulation (gain only) with SC cmfb is to run the transient. u can give a sinewave input and observe the output response. see if the gain is correct.

 

[windgone0204]

Re: How to simulate a SC-CMFB fully diffrential OTA with Hsp

i do tran analysis firstly,then i can know when the transistors are working in the saturation region.finaly, i do op analysis at the time.

 

[williamy]

solve my problem. when i am designing opa with SC, i only simulate tran