Error amplifier puts pole in response when doing SMPS feedback compensation?

[treez]

Hello,
The opamp that gets used as an error amplifier in an SMPS has a limited gain bandwidth product.
Will the opamp's limtiations mean there is a pole at about 50KHz in all SMPS's that use standard opamps for error amplifiers?

 

[FvM]

Why 50 kHz?

 

[treez]

Because above that most opamps start to fall away in gain I thought?
I mean, I just wonder if all these feedback transfer functions that one sees in books, are they realistic?, would the actual opamp used in the error amplifier be high-performance enough to support the kind of gains needed?
I mean, I am thinking all this calculation is floundered...maybe we should just take out the AP300 frequency analyser and go by that, and that only.?

 

[FvM]

The actual closed loop bandwidth depends on amplifier gain bandwidth product (GBW) and feedback factor. It can be easily estimated or sketched in a bode diagram.

A designer will choose the OP for sufficient GBW to achieve the intended amplifier characteristic, or modify the circuit if necessary.

 

[mtwieg]

Actually when the circuit is based on an inverting amplifier (like most error amplifiers are) the finite GBP of the amp will manifest in an increase in gain above some point, since current can flow directly from input to output through the feedback components. The severity of this effect depends a lot on the component values used, and the load impedance on the op amp output, but it's something that needs attention because it can easily cause instability.