designing an integrator

[AtticusWang]

I wanna design an active RC integrator with a fully differential opamp. The -3dB cutoff frequency is 2MHz. Now I need to determine the DC gain and GBW of the Opamp.

My point: I use an ideal opamp model -VCCS parralle with a output resistor and capacitor - by aultering the gain and GBW of the ideal opamp, I get different frequency responses of the integrator. and I can sweep a minimum GBW under which condition the -3dB frequency of the intergrator is 2MHz.

However, according to mathmatical analysis, the GBW of the opamp is just the second pole of the integrator(the dominant pole is 2M). If GBW>4.1*2M, the -3dB cutoff frequency is very close to 2M(5% error); the result is quite different from the sweep simulation result.

could u plz tell me what is the trick is ? or any good suggestions about designing a integrator?
Many thanks.

 

[LvW]

Your description is a bit confusing. Normally, an integrator is specified by its time constant (reverse of the cross-over frequency with gain of unity).

Quote: I wanna design an active RC integrator with a fully differential opamp. The -3dB cutoff frequency is 2MHz. Now I need to determine the DC gain and GBW of the Opamp.

I suppose you speak about an active RC low pass with a 3-dB-frequency of 2 MHz, which will be used (far) above this frequency as an integrator, right?

Quote: My point: I use an ideal opamp model -VCCS parralle with a output resistor and capacitor

Usually, an opamp is called VCVS (and not VCCS). Therefore the question: Is your active device a voltage or a current source?

 

[alfredop]

Atticus:

Connect a VCVS (set its gain to 1) at the output of the macromodel you are describing (VCCS + R||C) so that it acts as a buffer between this stage and the load introduced by the feedback network of the integrator. This may solve your problem, assuming that you are only checking the frequency response of your circuit in AC simulations.

You mention that this is a FD integrator, so you will need to set the common-mode level of your macromodel if you need to check its transient response.

 

[AtticusWang]

LvW:
Excuse me, maybe you are confused by my mentioning of integrator and low pass filter. I want to design a low pass filter with -3dB frequency of 2MHz, and this is the major specification.
I use a VCCS in paralle with R||C + VCVS(as a buffer) as a macro opamp model. Now my question is, how can I get the specifications of the Opamp I need to design? should I just sweep the GBW of the macromodel? Or caculate and analysis according to the single pole model opamp?

Your description is a bit confusing. Normally, an integrator is specified by its time constant (reverse of the cross-over frequency with gain of unity).

Quote: I wanna design an active RC integrator with a fully differential opamp. The -3dB cutoff frequency is 2MHz. Now I need to determine the DC gain and GBW of the Opamp.

I suppose you speak about an active RC low pass with a 3-dB-frequency of 2 MHz, which will be used (far) above this frequency as an integrator, right?

Quote: My point: I use an ideal opamp model -VCCS parralle with a output resistor and capacitor

Usually, an opamp is called VCVS (and not VCCS). Therefore the question: Is your active device a voltage or a current source?

Added after 3 minutes:

Thank you.I will have a try.
A CMFB is needed. And what I want to do now is to get all the specifications of the FD according to those specifications of the low pass filter.

Atticus:

Connect a VCVS (set its gain to 1) at the output of the macromodel you are describing (VCCS + R||C) so that it acts as a buffer between this stage and the load introduced by the feedback network of the integrator. This may solve your problem, assuming that you are only checking the frequency response of your circuit in AC simulations.

You mention that this is a FD integrator, so you will need to set the common-mode level of your macromodel if you need to check its transient response.

 

[LvW]

LvW:
Excuse me, maybe you are confused by my mentioning of integrator and low pass filter. I want to design a low pass filter with -3dB frequency of 2MHz, and this is the major specification.
I use a VCCS in paralle with R||C + VCVS(as a buffer) as a macro opamp model. Now my question is, how can I get the specifications of the Opamp I need to design? should I just sweep the GBW of the macromodel? Or caculate and analysis according to the single pole model opamp?

OK, now your specification seems to be clear.
It is your (first) task to find the minimum GBW of an opamp which will be used to build a low pass with a 3-dB-corner frequency of 2 MHz, right?
However, as your simplified opamp model (consisting of a VCCS with R||C as a load and an output buffer) is only of first order, I do not expect any difference between simulating/sweeping the opamp GBW and "caculate and analysis according to the single pole model opamp".
When there is any difference, I suppose your calculation (with pencil and paper) is not correct. It is a second order transfer function which has to be solved for the first pole. But I think, it is really not necessary to duplicate simulation results.