Lecture 5 Review Current Source Active Load Modified Large / Small Signal Models Channel Length Modulation Text

Text sec 1.2 pp. 28-32; sec 3.2 pp. 128-129

Current source Ideal goal Small signal model: Open circuit RD=

Realizing current source: MOSFET Large signal nonideality: Compliance range Looks like current source only for VDS>Veff

MOSFET ID-VDS characteristic for fixed VGS Small-signal nonideality: slope in active region

Cause: Channel length modulation

Channel length modulation

Channel length modulation

Modify small-signal model: Finite rds current source Slope = I/V

SLOPE =

Ideal: zero slope rds =

Caution: rDS(on) vs. rds confusion!

rDS(on) Triode region Large signal True resistance (V-I through origin)

rds Active region Small signal Models nonideality of current source

Refined MOSFET Small Signal Model

Add rds in parallel with gmvgs current source at output SAME FOR N-ch, P-ch How to relate rds to DC operating point? Example: gm = 2ID/Veff

ID- VDS Characteristic for Different VGS

"Family" of curves

ID- VDS Characteristic for Different VGS

Extrapolate backward: intersect at VDS-axis

1/slope provides small signal resistance rds

Intersect at -1/l

Slope:

1 ID 1 = rds = rds 1 l lI D

MOSFET small signal model

2I D gm = Veff

rds =

1 lI D

Increasing Gain Typical gain (resistive load) Lab 4 example: |av | 2 Class example: |av | 11.1 How to increase av?

Transconductance gm Definition

dID gm = dVGS

Transconductance gm Definition ID from Square law: gm in terms of W/L, Veff

Summary of gm expressions All equivalent! choose whichever gives easier math Cant memorize? rederive from definition of gm
dID gm = dVGS
W2 gm = m nCox Veff L2 gm = 2I D Veff

W gm = 2mnCox ID L

Common source circuit (Lab 4)

Setting operating point: Adjusted function generator offset for DC output at midpoint of signal swing

Common source circuit (Lab 4) DC operating point Chosen for halfway between rails ID=1.25mA Veff 2.0V (depends on parameters)
VOUT = VDD - I D RD VOUT = +2.5V VDD - VOUT ID = RD 5V - 2.5V = = 1.25mA 2kW

Common source circuit (Lab 4) Small signal gain magnitude = 2.5 Not impressive!
gm = 2I D 2(1.25mA) = Veff 2.0V

gm = 1.25mA / V av = -g m RD = -(1.25mA/ V)(2 kW) av = -2.5

How to increase av? Look at gain expression:

av = gm RD

Increase RD New RD = 10k (5X old value) Problem: VOUT = VDD - I D RD DC operating point VOUT = 5V - (1.25mA)(10kW) 144 443 2 Violates condition for 12.5V active region: triode! V OUT = -7.5V ? DC operating point stuck at negative rail

Look at problem symbolically Use gm=2ID/Veff IDRD = DC drop on load Optimal bias at output: constrained to VDD / 2 ID, RD not involved!
2ID gm = Veff 2I R av = gm RD = D D Veff

VDD ID RD = 2 V av = DD Veff

Value of approximate symbolic approach vs. exact numerical results from simulation

How to increase gain (resistive load) increase VDD usually fixed by application, process decrease Veff does increase gm but ...

VDD av = Veff

Problems decreasing Veff Veff, W/L gm expression: W 2X increase in gm: gm = 2mnCox ID 4X increase in size L (cant increase IDRD) Increased area: cost penalty Increased capacitance speed penalty Veff < 200 mV: subthreshold region Not square law: gm expressions invalid

Increase av: Different approach Give up on resistive load ... What is highest resistance?

Increasing av: Different approach What is highest resistance? Infinite: open circuit Problem: no path for ID Any circuit element that: provides DC current, but is open circuit in small signal model?

Current source! Open in small signal model Realizing current source: MOSFET

Lab Circuit: MOSFET with active load Small signal model for M1 Thevenin equivalent looking into drain of M2 (see text sec. 3.1)

Small signal model

M1 common source

M2 Thevenin equivalent

Simplify small signal model

Combine rds1, rds2 in parallel

Small signal gain

vout = av = -gm1 (rds1 rds2 ) vin

Current source load: Large signal considerations Output swing limits Top: M2 crash into triode Bottom: M1 crash into triode

Common Source with Active Load DC Sweep Schematic

Active Load Simulation Result (DC Sweep)

Determining DC Operating Point

Set small signal vin = 0

Determining DC Operating Point Active region: Veff determines ID Correct VIN: M1, M2 agree Example: Veff1 = 1.0V ID = 100 A

If DC bias at input is wrong? Current source "disagreement" KCL crisis at output: 2A, nowhere to go What happens?

If DC bias at input is wrong? Capacitance at output node Vout 2A flows into cap, charges up VDS1 increases I1 increases VDS2 decreases I1 decreases Changes in VDS cause changes in ID until agreement is reached: ID1 = ID2

How much change in VDS? Changes in VDS cause changes in ID until agreement is reached: ID1 = ID2 BUT Active region: ID is a weak function of VDS Large change in VDS for small change in ID Output very sensitive to changes in ID: Small Veff at input Small ID Requires large VDS at output for ID agreement Good: high voltage gain Bad: tricky to get correct input bias point

Frequency Domain Considerations Ideal op-amp goals: Infinite gain Infinite bandwidth Active load helps gain What about bandwidth?

Frequency Domain Analysis Start simple: Assume single CL at output (Ignore MOS capacitances for now ) Find transfer function vout/vin

Small signal model

Combine rds1||rds2 = rout

Simpler small signal model

Combine rout CL into impedance ZL

Simplified Small Signal Model

Small signal gain: vout/vin = av = -gmZL Frequency dependence of ZL provides frequency dependence of transfer function

Closer Look at ZL: Impedance is parallel combination of rout, 1/sCL

rout ZL = 1 + srout CL

Behavior of ZL over frequency: Let s=jw

Low frequency limit:mostly rout

High frequency limit: mostly CL

Transfer function Substitute in ZL

Magnitude
vout gm rout = vin 1 + (wrout CL )2

Bode Plot of Transfer Function Magnitude

Magnitude vs. Frequency (log-log plot) Bandwidth: w3dB

3-dB Frequency / Bandwidth Frequency at which magnitude is 3 dB down (reduced by factor 1/2)
vout MAX = gm rout AT w = 0 vin vout 1 THEN AT w3dB, = gm rout vin 2 1 gmrout 1 gmrout = w 3dB = 2 2 routC L 1+ (w 3dBroutCL )

Revisit Bode Plot: Gain, Bandwidth inversely related!

Unity Gain Frequency wT / Gain-Bandwidth Product wT : Frequency at which magnitude is 1 Use approximation wT >> 1/routCL
1= gm rout 1+ (wT rout CL )
2

gm rout

(wT rout CL )

gm wT = CL

Gain x Bandwidth Product

av = gm rout

w 3dB

1 = rout CL

1 gm gm rout = 1 3 rout CL 2 CL GAIN 13 2

Independent of rout! Poorly controlled rout OK

Summary: Active Load Active load DC considerations: Output swing limited by triode crash To voltage within Veff of rail Active load good news / bad news: Good news: high gain Bad news: very sensitive to input DC bias

Massage small signal gain result Small signal gain Look at parallel combination Substitute expression for rds
av = gm1( rds1 rds2 ) 1 rds1 rds 2 = 1 1 + rds1 rds2 1 rds1 rds 2 = l 1ID + l 2 ID 1 rds1 rds 2 = (l1 + l2 ) ID

Massage small signal gain result Small signal gain Substitute for gm, parallel rds ID drops out!
av = gm1( rds1 rds2 ) 2I 1 av = D Veff1 (l1 + l 2 )I D 4 4 { 1424 3
gm1 rds1 rds2

2 av = (l1 + l2 )Veff1

Only l, Veff to work with

Improve Gain Reduce Veff Minimum 200 to 300mV (subthreshold) May not want to go that low (W,L too big) Reduce l1, l2 How? Where does l1 come from?

Square law model with channel length mod

[ (

)]

I D -sat

ID-sat term (at pinchoff) + "extra"

Square law model with channel length modulation

mn Cox W 2 ID = (VGS - Vtn ) 1 + l VDS - Veff 2 L 1444 24443 4 4

[ (

)]

I D -sat

Fractional extra part is l(VDS-Veff) Meaning of l: Fractional change in current ID per volt change in VDS

What causes change? Where does l come from? Change in effective channel length L One way to reduce l: longer L Change L represents smaller fraction

After some semiconductor physics ... Definition of l Fractional change Semiconductor physics ... (see J&M p. 26)
l= DI D ID DVDS

DID DL = ID L l= 1 DL L DVDS 2KS e 0 1 qNSUB 2 VDS - Veff 2KS e 0 1 qNSUB 2 VDS - Veff

DL = DVDS l= 1 L

KS Silicon dielectric constant 11.8 NSUB Substrate doping units /cm3 Sanity check: 1E+14 to 1E+17 VDS from active-triode edge to large VDS Caution: consistent length units on L, NSUB, e0

Substrate doping NSUB parameter Needed for SPICE Extraction procedure: 1) Calculate slope from ID-VDS plot 2) rds = 1/slope (small signal model) 3) Calculate l 4) Calculate NSUB

Example VDS-ID data from Lab 5 for P-channel MOSFET:

1) Calculate slope from ID-VDS plot 2) rds = 1/slope (small signal model)

3) Calculate l

ID = 482 A

1 1 1 l= = = = 0.052V -1 I D rds ( 482mA)( 40.2kW) 19.4V

4) calculate NSUB

l=

1 2K Se0 1 L qN SUB 2 VDS - Veff

VDS = 4.48 V

0.052V

2(11.8)(8.85E -12F / m) 1 1 = (1E - 5m) (1.6E -19coul ) N SUB 2 4.48V - 0.84V

N SUB = 3.32E + 22 m -3 = 3.32E +16 cm -3

For CD4007, L = 10m = 1.0E-5m VDS, Veff for largest VDS data point

Simulation exercise Add NSUB to N-channel, P-channel models DC sweep for CS Amplifier with Active Load

Common Source with Active Load (DC) Sweep input over full range 0 to +5V

DC Sweep Around Operating Point