nonlinearity of cascode straucture

[xuedashun]

I got this interview question and am not sure what's the exact answer.
Could you give some suggestion? Thank you.

For the normal cascode CMOS structure, which contribute to more nonlinearity?
Does it have different solutions in analog design and RF design ?

 

[xuedashun]

Do anyone know how to answer this quesiont? Thanks.

 

[dick_freebird]

Well, as soon as your "guard" devices falls out of saturation
you will start to modulate the drain voltage on the "master"
and lose drain linearity. This could be DC or every cycle if
amplitude is high.

For RF, cascodes are often used not to defeat DC linearity,
but to stand off the majority of the output voltage AC so it
can't push back on the gate of the amplifier device. In this case
the cascode will be tied to a stiff rail (often vdd), where
for analog it's often a 2*VTN, low voltage (to maximize
the in-saturation drain voltage range).

 

[xuedashun]

dick_freebird, thanks for your reply.

So in analog design case, which one has larger non-linearity?

Well, as soon as your "guard" devices falls out of saturation
you will start to modulate the drain voltage on the "master"
and lose drain linearity. This could be DC or every cycle if
amplitude is high.

For RF, cascodes are often used not to defeat DC linearity,
but to stand off the majority of the output voltage AC so it
can't push back on the gate of the amplifier device. In this case
the cascode will be tied to a stiff rail (often vdd), where
for analog it's often a 2*VTN, low voltage (to maximize
the in-saturation drain voltage range).

 

[dick_freebird]

Which one of what? You seem to be asking us to distinguish
between cases you have yet to present.