[Moved]: Curvature correction techniques

[kenambo]

Hi

while I am dealing with reference voltage and current circuits, I came across zero order,first order and superior order curvature correction.

What does these terms exactly mean? what is the order mentioned here?
thanks

 

[andre_luis]

It is not clear your need; "voltage reference" is an expression used at continous operation in general associated with a comparator, whereas a "n-th order filter" is used for signal filtering. Curvature correction techniques as far as I know are used to compensate the non-linearity in the ratio at input/output.

 

[kenambo]

Hi,

To be clear, I mentioned Bandgap reference circuits. In that circuit we employ both PTAT and CTAT voltages.

The temperature independence is obtained bycancelling the higher order terms which causes voltage drift with temperature. I understood this. But I want to be clear on what is exactly mentioned by zero, first or superior order curvature correction.?

Hope this explains my situation. I studied zero order curvature corrected voltage reference but i couldn't figure out the difference between zero and first order curvature correction techniques.

thanks

 

[andre_luis]

Oh sorry, now I got the question; Just to check, are you aware about polinomial interpolation techniques used to perform pattern matching ? As far as I can deduce, it make allusion to the behavior of the trnasfer function of the "curve" Voltage vs. temperature ( V=f(T) ) in order model it in an circuit that cancels the non-linear portion of the funcion.

 

[FvM]

I believe the question is well answered in literature and can be quickly found at Google.

Nevertheless, to give the basic idea. The order is the order of temperature compensation function.

Zero order: No temperature compensation at all, just trim the reference value. Leaves a temperature proportional error
First order: Linear temperature compensation, leaves a quadratic error

 

[andre_luis]

In short words, the order of the function increase circuit complexity but reduces the estimation error.