CMFB Loop Instabilty OR Design Issue

[ssuchitav]

Hi,

I am trying to simulate the circuit shown in the attachment. An ideal OPAMP model (dc gain=100dB UGB=1G) is used for CMFB. The input is of 100MHz.

The circuit is showing unexpected rail to rail swing instead of the expected across 1.25V. I have cheked the CMFB loop stabilty and it looks fine.

Please help me in debugging this behaviour.
Thanks in advance for your suggestions.

 

[FvM]

What's the actual loop bandwith? Can you show the loop gain plot?

 

[jimito13]

Just a remark,irrelevant with your issue..100dB for CMFB as standalone circuitry is too much,how can you realize such a large gain (which is probably larger than your amplifiers gain)?
In most cases,the biggest part of the CMFB loop-gain is located within the opamp,so the CMFB circuit as independent entity should have less gain (20-40dB let's say).

Now for your problem :
In parallel with what FvM asked you to provide can you tell us the amplitude of your input signal and also show us the output transient response for this input?

 

[ssuchitav]

Hi,

Please see the AC response and the setup for the same.

I have also attached the transient behaviour of the circuit.

I forgot to mention about the termination in my original post. This can be seen in the setup.

Best Regards

 

[FvM]

You apparently misinterpreted the loop gain diagram. There are 3 poles below respectively near the unity gain frequency and a negative phase margin. You build an effective oscillator. For present pole locations, you have to reduce the gain by 70 to 80 dB.

 

[jimito13]

This is not rail to rail swing,this is oscillations!Thus,you must investigate for the cause of instability.

 

[ssuchitav]

I'll try this out.

Btw, does a negative phase, not equal to -180, @gain=0dB i.e. UGB cause instability?

 

[jimito13]

Ofcourse.Remember the definition of Phase Margin and how it comes out from Nyquist's diagram.You measure the distanse of the phase at Gain=0dB from -180deg.PM must be always positive to ensure stability.Typically must be in effect PM>45deg.

 

[FvM]

You should try achieve 40 to 50 degree phase margin. Near zero or negative phase margin will cause oscillations in most cases.

 

[ssuchitav]

Thanks.
I'll rework on this.