I have already set the bias point and all of them are in saturation. This amplifier has gain of around 35 dB.
My concern is: when the values of input ac signals are changed, the dc value at the output node is changing. I am afraid if it is possible for any of the transistors to exit saturation. If it is possible, should I bias the output node too? How can I do that?
Please note that I want to avoid resistors for biasing purpose.
(above two are PMOS. please ignore the position error of the sources in PMOS. sources are connected to Vdd)
Any suggestion, tutorial, link will be greatly appreciated.
The voltage at the output node would change with the input. That is how it would amplify. You cannot expect an amplifier to amplify if the output node doesn't change at all.
The problem you would face would be of the output range.
The maximum output would be when the PMOS_2 would be at the edge of saturation.
The minimum output would be when the NMOS_4 would be at the edge of saturation.
And this range divided by the gain would give the input range.
Try some basic Analog Design Books like Razavi or Grey/Meyer.
As long as the amplifier isn't overloaded, you have no reason to assume bias point shifting. A practical amplifier circuit will set the bias point by a feedback loop rather than using a precalculated input bias voltage. This is necessary to account for threshhold voltage exemplar variations and tempearture drifts.