0% found this document useful (0 votes)
472 views38 pages

10 High Frequency Transistor Models

This document provides an outline and reading for a lecture on high frequency transistor models. It discusses MOSFET and BJT capacitances and develops high frequency models for each device. It introduces the concepts of transition frequency fT and Miller's theorem for analyzing amplifier frequency response. Key points covered include MOSFET gate, drain, and source capacitances; the dependence of fT on technology node for MOSFETs; elements of the hybrid-pi BJT model including base, collector, and emitter junction capacitances; and limitations of Miller's approximation at high frequencies.

Uploaded by

吳心
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
472 views38 pages

10 High Frequency Transistor Models

This document provides an outline and reading for a lecture on high frequency transistor models. It discusses MOSFET and BJT capacitances and develops high frequency models for each device. It introduces the concepts of transition frequency fT and Miller's theorem for analyzing amplifier frequency response. Key points covered include MOSFET gate, drain, and source capacitances; the dependence of fT on technology node for MOSFETs; elements of the hybrid-pi BJT model including base, collector, and emitter junction capacitances; and limitations of Miller's approximation at high frequencies.

Uploaded by

吳心
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 38

EEET 2404 Electronic Circuits

Lecture 10

High Frequency Transistor Models

Assoc. Prof. James Scott


James.Scott@rmit.edu.au

EEET2404: Electronic Circuits Page 10–1


Outline
•  MOSFET devices
–  capacitances, high frequency model
•  BJT devices
–  capacitances, high frequency model
•  Miller approximation
•  High frequency responses of amplifiers
–  CS/CE amplifiers
–  CG/CB and cascode amplifiers
–  Source/Emitter follower amplifiers

EEET2404: Electronic Circuits Page 10–2


Reading
•  Sedra/Smith, chapter 8, sections 8.1 to 8.7

EEET2404: Electronic Circuits Page 10–3


MOSFET Capacitances

Overlap capacitance consists of two


parts:
-overlap of the gate and drain/source
areas
-fringing capacitance between the gate
and the top of the S/D regions

EEET2404: Electronic Circuits Page 10–4


MOSFET Capacitances (2)
Gate Capacitances:
1
Cgs = Cgd = WLCox (triode region)
2
2 ! Note: these capacitances vary
Cgs = WLCox #
3 " (saturation region) with VGS and VDS
Cgd = Coverlap = WLovCox #
$
Lov is typically 0.05 to 0.5L
Cgs = Cgd = 0 !#
" (cutoff)
Cgb = WLCox #
$
Junction Capacitances:
Csb0
Csb = , where V0 is the built-in voltage (0.6V to 0.8V)
V
1+ SB
V0 Csb0 is the value of Csb at VSB = 0V
Cdb0 Cdb0 is the value of Cdb at VDB = 0V
Cdb =
V
1+ DB
V0
EEET2404: Electronic Circuits Page 10–5
High Frequency MOSFET Model

(a) High frequency MOSFET model, (b) device symbol with capacitances
shown explicitly.

If the base and the source are connected together, then CSB disappears and
CDB becomes CDS.

EEET2404: Electronic Circuits Page 10–6


MOSFET Transition Frequency
•  The transition frequency is the frequency
where the short circuit current gain drops to
1.
•  This is a figure of merit for the high
frequency operation of the transistor.

EEET2404: Electronic Circuits Page 10–7


MOSFET Transition Frequency (2)
Cgd Io

+
Is vgs Cgs gmvgs ro
-

Setting the current gain to


I o ≅ gm vgs unity:
Is gm
vgs = ωT =
jω (Cgs + Cgd ) (Cgs + Cgd )
Io gm gm
= fT =
I s jω (Cgs + Cgd ) 2π (Cgs + Cgd )

EEET2404: Electronic Circuits Page 10–8


MOSFET Transition Frequency (3)
•  Nano-scale MOSFETs can have ƒT values in
the GHz region
–  45nm Silicon-on-insulator (SOI) process can
deliver an ƒT of 200 GHz

EEET2404: Electronic Circuits Page 10–9


Summary of MOSFET Model Elements

W Csb0
gm = µ nCox VOV Csb =
L VSB
1+
V0
VA Cdb0
ro = Cdb =
ID VDB
1+
V0
2 gm
Cgs = WLCox + WLovCox fT =
3 2π (Cgs + Cgd )
Cgd = WLovCox

EEET2404: Electronic Circuits Page 10–10


BJT Capacitances

EEET2404: Electronic Circuits Page 10–11


Base Charging or Diffusion Capacitance
In the active mode, minority carrier charge is stored
in the base region. For an NPN transistor:
Qn = τ F iC where τ F is the forward base-transit time (average time
taken to cross the base)
Since iC depends on vBE :
dQn
Cde ≡
dvBE
di
=τF C
dvBE
so
IC
Cde = τ F gm = τ F
VT

EEET2404: Electronic Circuits Page 10–12


Base-Emitter Junction Capacitance
•  A change in vBE also changes the charge
stored in the base-emitter depletion layer.
•  Since the base-emitter junction is forward
biased, the depletion layer capacitance is
approximately given by:

C je ≅ 2C je0
where C je0 is the value of C je at zero B-E junction
voltage.

EEET2404: Electronic Circuits Page 10–13


Collector-Base Junction Capacitance
•  In active-mode operation, the collector-base
junction is reverse biased.
•  Therefore the junction capacitance is a
depletion capacitance.
Cµ 0
Cµ = m
! VCB $
#1+ &
" V0c %
where
Cµ 0 is the value of Cµ at zero voltage
VCB is the magnitude of the reverse-bias voltage
V0c is the CB junction buit-in voltage (typically 0.75V)
m is its grading coefficient (typically 0.2 to 0.5)

EEET2404: Electronic Circuits Page 10–14


High Frequency Hybrid-π Model
rx Cµ
B B' C

+
rπ vπ Cπ gmvπ ro
-

E E

Cπ = Cde + Cje
Cπ is in the range of a few pF to a few tens of pF
Cµ is in the range of a fraction of a pF to a few pF

EEET2404: Electronic Circuits Page 10–15


BJT Transition Frequency
jωCµvπ
rx Cµ Ic
B B' C
+
Is rπ vπ Cπ gmvπ ro
-
E E
I c = ( gm − jωCµ ) vπ
Is
vπ = I s ( rπ || Cπ || Cµ ) = gm
1 rπ + jω (Cπ + Cµ ) ωT =
(Cπ + Cµ )
Ic gm − jωCµ
h fe = = gm
I s 1 rπ + jω (Cπ + Cµ ) fT =
2π (Cπ + Cµ )
gm rπ β0
≅ =
1+ jω (Cπ + Cµ ) rπ 1+ jω (Cπ + Cµ ) rπ

EEET2404: Electronic Circuits Page 10–16


Variation of ƒT with IC

ƒT is typically in the range of 100 MHz to tens of GHz.


The hybrid-π model characterises the BJT quite accurately up to a
frequency of about 0.2ƒT.

EEET2404: Electronic Circuits Page 10–17


Summary of BJT Model Elements

I Cµ 0
gm = C Cµ = m
VT ! VCB $
#1+ &
" V0c %
VA gm
ro = fT =
IC 2π (Cπ + Cµ )
Cπ = C je + Cde
= 2C je0 + τ F gm

EEET2404: Electronic Circuits Page 10–18


Miller’s Theorem

EEET2404: Electronic Circuits Page 10–19


Limitations of Miller’s Theorem
•  The floating impedance and the voltage gain
should be calculated at the same frequency
–  In practice, we often use the low-frequency
voltage gain to simplify calculations
–  This leads to some inaccuracies in the solution

EEET2404: Electronic Circuits Page 10–20


High Frequency Analysis Example
The following plot shows the errors in using
Miller’s approximation to analyse a common-
source amplifier.

EEET2404: Electronic Circuits Page 10–21


High–Frequency Response of Amplifiers
•  To analyse the frequency response of an amplifier:
–  Replace the transistors with their high frequency
equivalent circuits
–  Use Miller’s approximation or analyse the resulting
circuit directly
•  The device capacitances cause a high frequency
roll-off in the response
•  Any DC blocking capacitors (which are much
larger) will cause a low frequency roll off in the
response

EEET2404: Electronic Circuits Page 10–22


Common-Source Amplifier

EEET2404: Electronic Circuits Page 10–23


Common-Source Amplifier (2)

The node equation at the drain is


V
sCgd (Vgs − Vo ) = gmVgs + o' + sCLVo
RL
which can be expressed as
'
−Vo 1+ s (CL + Cgd ) RL
Vgs =
gm RL' 1− sCgd gm

EEET2404: Electronic Circuits Page 10–24


Common-Source Amplifier (3)
A loop equation at the input gives
Vsig' = I i Rsig
'
+ Vgs
and we can derive an expression for I i at the G node
I i = sCgsVgs + sCgd (Vgs − Vo )
so
Vsig' = Vgs "#1+ s (Cgs + Cgd ) Rsig
' $ '
% − sCgd RsigVo
Therefore, we obtain
Vo − ( gm RL' )"#1− s (Cgd gm )$%
=
Vsig' 1+ s "Cgs + Cgd (1+ gm RL' )$ Rsig
{ '
+ (CL + Cgd ) RL' + s 2 "#(CL + Cgd ) Cgs + CLCgd $% Rsig
'
RL' }
# %
The amplifier has two poles:
1
ω P1 ≅
"C + C (1+ g R' )$ R' + (C + C ) R'
# gs gd m L % sig L gd L DC gain is
"C + C (1+ g R' )$ R' + (C + C ) R'
ωP2 ≅
# gs gd m L % sig L gd L A0 = −gm RL'
"(CL + Cgd ) Cgs + CLCgd $ Rsig
' '
# % RL
where ω P1 is the dominant pole.

EEET2404: Electronic Circuits Page 10–25


Common-Source Amplifier (4)
Since the gain expression has a numerator of the first order,
it has one of the two transmission zeroes at infinite frequency.
The second zero is at
g
s = sZ = m
Cgd
Since this is on the positive real axis of the s-plane and has a
frequency
g
ωZ = m
Cgd
Since gm is usually large and Cgd is usually small, fZ is normally
a very high frequency and has negligible effect on the -3dB frequency,
fH = fP1

EEET2404: Electronic Circuits Page 10–26


Common-Emitter Amplifier

Similar circuit
to the CS
amplifier!

EEET2404: Electronic Circuits Page 10–27


Common-Emitter Amplifier (2)
First we will relate Vsig' and Rsig
'
to Vsig and Rsig and the other
equivalent circuit parameters:
! rπ $
Vsig' = ## &&Vsig
" Rsig + rx + rπ %
'
Rsig = rπ || ( Rsig + rx )
The DC gain becomes

A0 = −
Rsig + rx + rπ
( gm RL' )

The zero frequency is


1 gm
fZ =
2π Cµ
The poles are
1 1
fP1 ≅
2π )*Cπ + Cµ (1+ gm RL' )+, Rsig
'
+ (CL + Cµ ) RL'
) ' + '
1 *Cπ + Cµ (1+ gm RL ), Rsig + (CL + Cµ ) RL
'

fP 2 ≅
2π )Cπ (CL + Cµ ) + CLCµ + Rsig
' '
* , RL
For fZ , fP 2 >> fP1,
fH ≅ fP1

EEET2404: Electronic Circuits Page 10–28


Effect of Rsig low

If we assume that Rsig is zero, then for the CS amplifier

Vo ( −g R )"#1− s (C
m
'
L gd gm )$%
=
Vsig 1+ s (CL + Cgd ) RL'
so the high frequency response is now controlled by a
pole formed by CL + Cgd together with RL' .
The -3dB frequency is now
1
fH =
2π (CL + Cgd ) RL'
i.e. it is now controlled by the output circuit.

EEET2404: Electronic Circuits Page 10–29


Common-Gate Amplifier

1
fH =
2π !"Cgs Rgs + (Cgd + CL ) Rgd #$
where
Rgs = Rsig || Rin
Rgd = RL || Ro
ro + RL
Rin =
1+ gm ro
Ro = ro + Rsig + ( gm ro ) Rsig

EEET2404: Electronic Circuits Page 10–30


MOS Cascode Amplifier

Note that Cdb1 and Cgs2 are


in parallel, and CL and
Cgd2 are also in parallel.

EEET2404: Electronic Circuits Page 10–31


MOS Cascode Amplifier (2)
Using the open-circuit time constants method:
1. Capacitance Cgs1 sees a resistance Rsig
2. Capacitance Cgd1 sees a resistance Rgd1
Rgd1 = (1+ gm1Rd1 ) Rsig + Rd1
where Rd1, the total resistance at D1 is
ro2 + RL
Rd1 = ro1 || Rin2 = ro1 ||
gm2 ro2
3. Capacitance (Cdb1 + Cgs2 ) sees a resistance Rd1
4. Capacitance (CL + Cgd 2 ) sees a resistance ( RL || Ro ) where
Ro = ro2 + ro1 + ( gm2 ro2 ) ro1
So the effective time constant τ H is
τ H = Cgs1Rsig + Cgd1 !"(1+ gm1Rd1 ) Rsig + Rd1 #$
+ (Cdb1 + Cgs2 ) Rd1 + (CL + Cgd 2 ) ( RL || Ro )

EEET2404: Electronic Circuits Page 10–32


MOS Cascode Amplifier (3)
So the -3dB frequency is
1
fH ≅
2πτ H
Re-writing τ H
τ H = Rsig "#Cgs1 + Cgd1 (1+ gm1Rd1 )$% + Rd1 (Cgd1 + Cdb1 + Cgs2 )
+ ( RL || Ro ) (CL + Cgd 2 )

If Rsig is large, the first term can dominate, especially if (1+gm1Rd1) is


large. This happens when RL is large. So to extend the bandwidth, we
have to lower RL.
If Rsig is small, a large value of RL makes the dc gain large. Therefore,
the third term dominates.

EEET2404: Electronic Circuits Page 10–33


Comparison of CS and Cascode stages

EEET2404: Electronic Circuits Page 10–34


BJT Cascode Amplifier

EEET2404: Electronic Circuits Page 10–35


Source Follower Amplifier

EEET2404: Electronic Circuits Page 10–36


Source Follower Amplifier (2)
Low frequency gain:

AM =
( RL || ro )
( RL || ro ) + (1 gm )
1
Ro = || ro
gm
We can combine RL and ro into RL'
Transmission zeros occur at infinite frequency and at
gm
ωZ =
Cgs
This is very close to fT since Cgd << Cgs
It can be shown that
1 1
fH = =
2πτ H 2π (Cgd Rsig + Cgs Rgs + CL RCL )
where
Rsig + RL'
Rgs =
1+ gm RL'
1
RCL = RL || ro ||
gm

EEET2404: Electronic Circuits Page 10–37


Emitter Follower Amplifier
'
We can combine Rsig , rx and rπ into Rsig
'
Rsig = Rsig + rx
and
RL' = RL || ro
Transmission zeros occur at infinite frequency and at
1
ωZ =
Cπ ro
This is very close to fT .
It can be shown that
1 1
fH = =
2πτ H 2π (Cµ Rµ + Cπ Rπ ) ƒH is usually very high
where
'
Rµ = Rsig || !"rπ + ( β +1) RL' #$
'
Rsig + RL'
Rπ = '
Rsig R'
1+ + L
rπ ro

EEET2404: Electronic Circuits Page 10–38

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy