CMOS Circuits



The abbreviation CMOS stands for complementary MOS. This technology integrates both n-channel and p-channel MOSFETs on the same chip, resulting in extremely low-power devices with wide ranges of supply voltage.



1. Technology
2. Function
3. Application

1. Technology  

The components of integrated CMOS circuits are field effect transistors of the enhancement MOS type with both, p- and n-channel conductivity. These transistors are most often formed on a low doped (low conductivity) n-type silicon substrate. Into this substrate a p-type tub or well is implanted for n-channel devices and an n-type well or tub for p-channel devices. Into these areas the source and drain regions are implanted with the opposite conduction type (i.e. n-type implantation into the p-type well to form an n-channel MOSFET).
   The gate contact made of polycrystalline silicon is formed over a thin isolating oxide layer. drain, gate and source are isolated from each other by thick silicon dioxide barriers as are different transistors against each other.
   CDue to the absence of biasing resistors the CMOS technology is scalable, consumes low power and the transistors are quite fast. Thus this technology has become the major technology for verly large scale integrated circuits (VLSI), for example modern microprocessors with more than 10 million transistors per 1cm x 1cm chip area. The scaling is measured against the smallest feature size which normally is the length of the gate electrode seen in the direction from source to drain (see: MOSFET).

Cross section through a CMOS device showing both n- and p-channel MOSFETs.

Schematic of a single stage CMOS inverter.

2. Function  

The principle of a CMOS circuit can be seen from its schematic. The threshould voltages of the n-MOS and p-MOS transistors are chosen to a value that allows for the p-MOS transitors to form their channel on a gate voltage Vg<0.5Vdd while the channel in the n-MOS transistors is formed for Vg>0.5Vdd. Thus an inverter consists of just one n-MOSFET and one p-MOSFET series connected between the positive supply Vdd and the negative supply Vss with both gates connected to the input.
   Because the subthreshould current through a MOSFET is low, there is almost no current going from the positive to the negative supply as either the p-MOSFETs or the n-MOSFETs are in the non-conducting state. This is the reason for the low static power consumption of CMOS ciruits. However during transients, the charge on the gate electrodes has to be moved and for a very narrow voltage interval of the changing input signal, both transistors can actually become conductive. This leads to an almost linear increase of power consumption with switching frequency for CMOS devices.

 

Schematic of a 2 input CMOS NAND gate (left hand side) and NOR gate (right hand side).

 

 

Logic response and power consumption of a 2 input CMOS NAND gate (left hand side) and NOR gate (right hand side). (pspice simulation)

 

Transfer function of a CMOS inverter. The output voltage is ploted as a function of the input voltage. (pspice simulation)

3. Application  

The early CMOS devices of the 40xx family were very sensitive to electrostatic discharges damaging the thin gate oxide and destroying the chip. Most of today's CMOS devices are protected against damage by a set of diodes connecting the inputs to the supply voltage rails which in turn are connected to the drains, sources and the substrate. This first series of CMOS devices had a much wider supply voltage range than the TTL family but they were also significantly slower especially if used at a +5V supply voltage. In order to use CMOS and TTL in the same circuit, the CMOS devices usually had a +12V supply, making level conversion between CMOS and TTL necessary. Even if the same supply voltage is used for both CMOS and TTL chips a level conversion is advised due to the different definition of the logic levels and the limited drive capability of standard CMOS output stages.

familylogic "0"logic "1"
TTL< 0.8V> 2.0V
CMOS
   Vdd=+5V
< 30% Vdd
< 1.5V
> 70% Vdd
> 3.5V

   The functional range of the 40xx family was less than that of the 74xx TTL family and a CMOS version of the 74xx family the 74Cxx was introduced. However they suffered from their low frequency range and in the early 1980s the high-speed CMOS version 74HC(T)xx was introduced. The devices are 100% pin and function compatible with the original 74xx family. In the 1990s a variety of function and pin compatible families emerges, mainly for lower supply voltage application as the supply voltage for VLSI circuits (processors and chipsets) went down from 5V to less than 2V (in 2000).

Familyoperating
voltage
temperature
range
power
per gate
delay time
per gate
logic levelsoutput currentinput current
 
40xxB+3 to +15V-40 to +85°CDC: 0.001mW
1MHz: 1mW
40nsL: <30%Vdd
H: >70%Vdd
L: 1mA
H: 1mA
L: <1µA
H: <1µA
74Cxx+3 to +15V-40 to +85°CDC: 0.01µW
1MHz: 1mW
80nsL: <30%Vdd
H: >70%Vdd
L: 1mA(5V)/8mA(10V)
H: 1mA(5V)/8mA(10V)
L: 0.005µA
H:  0.005µA
74HCxx+2 to +6V-40 to +85°CDC: 0.0025µW
1MHz: 0.75mW
8nsL: <30%Vdd
H: >70%Vdd
L: 25mA
H: 25mA
L: <1µA
H:  <1µA
74HCTxx+4.5 to +5.5V-40 to +85°CDC: 0.0025µW
1MHz: 0.75mW
8nsL: <0.8V
H: >2V
L: 25mA
H: 25mA
L: <1µA
H:  <1µA
74ACxx+2 to +6V-40 to +85°CDC: 0.0025µW
1MHz: 0.75mW
4nsL: <30%Vdd
H: >70%Vdd
L: 50mA
H: 50mA
L: <1µA
H:  <1µA
74ACTxx+4.5 to +5.5V-40 to +85°CDC: 0.0025µW
1MHz: 0.75mW
6nsL: <0.8V
H: >2V
L: 50mA
H: 50mA
L: <1µA
H:  <1µA
74FCTxx+4.5 to +5.5V-40 to +85°CDC: 0.05mW
1MHz: 0.75mW
1.5nsL: <0.8V
H: >2V
L: 50mA
H: 50mA
L: <5µA
H:  <5µA
74LCXxx+2.3 to +3.6V-40 to +85°CDC: 0.0025µW
1MHz: 0.75mW
<6nsL: <30%Vdd
H: >70%Vdd
L: <25mA
H: <25mA
L: <5µA
H:  <5µA
74LVxx+1.0 to +5.5V-40 to +85°CDC: 0.0025µW
1MHz: 0.75mW
6nsL: <30%Vdd
H: >70%Vdd
L: 25mA
H: 25mA
L: <1µA
H:  <1µA
74LVXxx+2.0 to +3.6V-40 to +85°CDC: 0.0025µW
1MHz: 0.75mW
<11nsL: <30%Vdd
H: >70%Vdd
L: 25mA
H: 25mA
L: <1µA
H:  <1µA

 


 

Responsible for these pages: U. Zimmermann