Mixed Signal Circuit Design- Capcitive Feed through -Reduction methods
Oct 6, 2023
Module5
We have learned about the non-ideal effects of MOSFET when used as a switch.
We have learned about the charge injection in the first class.
Next is the clock feed-through or capacity feed-through.
In this capacity feed-through, the gate and drain of a MOSFET,
i.e. the gate and drain of a MOSFET,
is modelled as if the MOSFET is operating in the triode region.
The MOSFET would be operating in the triode region.
We are modelling the gate-switch capacitance and gate-drain capacitance.
The MOSFET would be operating in the triode region.
We are modelling the gate-switch capacitance and gate-drain capacitance.
The gate signal is a clock signal.
The clock signal is shown as phi.
When the clock signal phi is high,
the clock signal is fed through the gate-switch capacitance and gate-drain capacitance.
When the clock signal phi is high, the MOSFET would be switched on.
At that time, the V-in would be connected to the C-load capacitor.
The NMOS switch would be connected to the C-load capacitor.
When the phi is high, the V-in would be connected to the C-load capacitor.
At that time, the V-in would be connected to the C-load capacitor.
When the clock signal phi is low, the NMOS switch would be switched off.
When the NMOS switch is switched off, a capacitive voltage divider circuit would be formed across the gate, switch and drain.
When the clock signal transition is low, the N-channel MOSFET would be switched off.
A voltage divider circuit would be formed across the gate, switch and drain.
A portion of the clock signal would be connected to the C-load capacitor.
The voltage difference is given by delta Vload is equal to Coverlap into Vdd divided by Coverlap plus Cload.
Coverlap is the value of overlap capacitance.
LD stands for Length of Gate that Overlaps Drain Source.
Let's see how to reduce the charge injection and clock feed-through problem.
There are many methods to reduce the charge injection and clock feed-through problem.
The most common method is the Demi-switch method.
This figure shows the Demi-switch method.
Here, M2 is used as a MOSFET switch.
Its drain and source are shorted.
This Demi-switch is connected to the desired switch M1.
M1 is the actual switch.
The clock signals given to M1 and M2 are called Complement.
M2 is the complement of the clock signal given to M1.
These two clocks are applied in a slight delay.
Let's see how it works.
First, M1 turns off.
The clock signal is a low signal.
The size of M2 is half of the size of M1.
When M1 turns off, half of the channel charge is injected to the Demi-switch.
M2 is effectively shorted.
The drain and source are shorted.
A small voltage is applied to the gate to induce a channel.
So, the charge induced by M1 is essentially matched by the charge induced by M2
and the overall charge injection cancelled.
So, the overall charge injection cancelled.
Let's see when M2 turns off.
When M2 turns off, half of the channel charge is injected to both sides.
The drain and source of M2 is shorted.
M1 is on.
When M1 is on, all the charge from M2 is injected to the low impedance path and C load is charged.
So, the charge injection of M2 does not affect the voltage of C load.
This method is used to reduce the charge injection and clock feed through.
We are using Demi-switches.
M2 is the Demi-switch.
The drain and source of Demi-switch is shorted.
The clock feed of both switches is a complement.
Let's see when M1 is off.
When M1 is off, half of the charge is injected to M2.
The size of M2 is half of M1.
The drain and source of M2 is shorted.
So, we need to apply a voltage to induce a channel.
The drain and source of M2 is shorted.
So, there is a possibility of conduction.
The charge injected by M1 is essentially matched by the charge induced by M2.
And the overall injection is cancelled.
Let's see when M2 is turned off.
When M2 is turned off, half of the charge is injected to both directions.
The drain and source terminal is shorted.
M1 is turned on.
So, all the charges coming from M2 go through the low impedance path.
And C load is charged.
So, the charge injection of M2 does not affect the voltage of C load.
The next method is to reduce the charge injection and capacity feed through.
We are replacing the MOSFET switch using CMOS transmission gate.
The speciality of this is that we are using complementary signals in the transmission gate.
If there are any small changes in the complementary signals,
it will be cancelled and removed.
For this method, we need to control the clock signals precisely.
This is a requirement.
We need to control the complementary clock signals precisely.
Normally, non-ideal charge injection and clock feed through are taken as common mode signals of an amplifier.
The common mode rejection ratio of an amplifier reduces.
This figure is a full differential circuit to minimize the charge injection and clock feed through.
We all know the speciality of a differential circuit.
Common mode rejection ratio occurs in a differential circuit.
Non-ideal charge injection and clock feed through are common mode signals.
It is a requirement.
We can reduce it by this method.
The second order effects due to the amplitude dependence of the input signal affect the dynamic range of the circuit.
This is the method.
We are replacing the MOSFET switch using CMOS transmission gate.
What is the speciality of the transmission gate?
It operates in a complementary clock.
If there is noise in it, it will be cancelled.
The last figure is a full wave differential circuit.
It minimizes the charge injection and clock feed through.
The common mode rejection ratio reduces the non-ideal charge injection and clock feed through.

