L #5: Mosfet C A: Sfsu - E 301 - E L AB Haracteristics and Pplications
L #5: Mosfet C A: Sfsu - E 301 - E L AB Haracteristics and Pplications
L #5: Mosfet C A: Sfsu - E 301 - E L AB Haracteristics and Pplications
n
= 0.05 V
-1
2002 Sergio Franco Engr 301 Lab #5 Page 4 of 14
The slope of the i
D
v
GS
curve at a particular operating point Q(I
D
, V
GS
) is denoted as g
m
, and is
called the transconductance, in A/V or also in 1/. (Beware not to confuse g
m
with k or k, whose units
are A/V
2
) Upon differentiating Eq. (2), we end up with three expressions,
( ) 2
0.5( )
D
m n GS tn n D
GS tn
I
g k V V k I
V V
= = =
(7)
Likewise, a pMOSFET has ( ) 2
m p GS tp p D
g k V V k I = + = . Compared to a BJT, for which g
m
= I
C
/V
T
=
I
C
/(26 mV), an nMOSFET has g
m
= I
D
/[0.5(V
GS
V
tn
)], where usually 0.5(V
GS
V
tn
) >> 26 mV. This
explains why MOSFETs in general have notoriously much lower g
m
s than their BJT counterparts. Indeed,
our homebrew nMOSFET, whose k
n
is typical of small-geometry devices, has g
m
(I
D
= 1 mA) = 0.632
mA/V = 1/(1580 ). Conversely, a BJT has g
m
(I
C
= 1 mA) = 38 mA/V = 1/(26 ).
It is interesting to note that if we plot the i-v characteristic of a diode-connected MOSFET on a
quadratic-versus-linear system of axes, the resulting curve becomes a straight line. Specifically, if we
plot Eq. (2) on an x-y plane with
x = v
GS
(= v
DS
) (8a)
and
1
D
n
i
y
x
=
+
(8b)
then the characteristic becomes / 2( )
n tn
y k x V = , that is, a straight line with slope dy/dx = / 2
n
k and
x-axis intercept at V
tn
. This is illustrated in Fig. 4 for our homebrew nMOSFET. We shall find this
feature quite convenient from an experimental standpoint.
An insightful alternative for illustrating nMOSFET operation is by plotting i
D
versus v
DS
for
different values of v
GS
. The PSpice circuit of Fig. 5 generates such a plot for incremental steps in V
GS
of
0.5 V each. The resulting family of curves, shown in Fig. 6, reveals three regions of operation:
Fig. 4 Linearization of the i-v
curve of a diode-connected
nMOSFET.
2002 Sergio Franco Engr 301 Lab #5 Page 5 of 14
For v
GS
V
tn
(= 1.5 V) we get i
D
= 0, indicating that the nMOSFET is operating in the cutoff (CO)
region. In this region both the GS and DS ports act as open circuits.
For v
GS
> V
tn
, the nMOSFET is on. The region corresponding to v
DS
> V
GS
V
tn
is called the pinchoff
(PO), or saturation, or active region. As we know, in this region Eq. (2) holds, which we repeat here
i
D
= ( ) ( )
2
1
2
n
GS tn n DS
k
v V v + for v
DS
> V
GS
V
tn
(9)
In the PO region, the i
D
versus v
DS
curves are almost horizontal, indicating current-source behavior
by the DS port there. Moreover, if we project the PO curves to the left, they all converge to the same
point on the v
DS
axis. This point is located at 1/
n
V. Note that when operated in the diode mode, the
MOSFET has v
DS
= v
GS
, which obviously satisfies v
DS
> V
GS
V
tn
, indicating PO operation when the
device is on.
Fig. 6 Illustrating the three regions of operation of an nMOSFET.
M
301nMOSFET
v DS
0Vdc
I
VGS
0Vdc
0
Fig. 5 - PSpice circuit to plot i
D
versus v
DS
for different values of V
GS
.
2002 Sergio Franco Engr 301 Lab #5 Page 6 of 14
The region corresponding to v
DS
< V
GS
V
t
is called the ohmic (), or triode region. In this region
the channel behaves as a nonlinear resistor, and i
D
takes on the form
i
D
= ( ) ( )
2
1
2
DS
n GS tn DS n DS
v
k v V v v
(
+
(
for v
DS
< V
GS
V
tn
(10)
For small values of v
DS
, we can ignore v
DS
2
/2 as well as
n
v
DS
(recall that
n
<< 1), and approximate
1
D DS
DS
i v
r
(11a)
where
1
( )
DS
n GS tn
r
k V V
=
(11b)
represents the channel resistance in the limit v
DS
0. Such a resistance is controlled by V
GS
,
indicating voltage-controlled resistance behavior near the origin. As shown in Fig. 6, the borderline
between the PO and the regions is a locus of points forming a parabola.
As we know, it is convenient to express a total signal such as the drain current i
D
as the sum
i
D
= I
D
+ i
d
(12)
where
I
D
is the DC component, also known as large signal
i
d
is the AC component, also known as small signal
We work with DC signals when dealing with transistor biasing, and we work with AC signals when
dealing with amplification to find AC gain as well as in input and output resistances. To signify how a
FET relates DC voltages and currents (upper-case symbols with upper-case subscripts) we use large
signal models, while to signify how it relates AC voltages and currents (lower-case symbols with lower-
case subscripts) we use the small-signal model.
Large-Signal MOSFET Models:
With regard to the two regions of interest to us in this laboratory, namely the CO and the PO regions, we
make the following observations:
Fig. 7 Large-signal MOSFET models
2002 Sergio Franco Engr 301 Lab #5 Page 7 of 14
In the CO region a MOSFET draws only leakage currents, which for practical purposes can be neglected.
So, both the GS and the DS ports act essentially as open circuits.
In the PO region, the GS port still acts as an open circuit, but the DS port now acts as a dependent
current source, which in the case of the nMOSFET takes on the form
2
( / 2)( )
D n GS tn
I k V V , where
we are arbitrarily assuming
n
= 0 to simplify our DC calculations. Similar considerations hold for
the pMOSFET, and the two models are shown in Fig. 7.
Small-Signal MOSFET Model:
When used as an amplifier, a MOSFET is operated in the PO region where we model its way of relating
voltage and current variations via the small signal model. Due to its quadratic characteristic, the BJT is a
nonlinear device. However, if we stipulate to keep its signal variations sufficiently small (hence the
designation small-signal), then the model can be kept linear albeit approximate. For an nMOSFET, the
small signal constraint is
v
gs
<< 2(V
GS
V
tn
) (13)
while for a pMOSFET it is v
gs
<< 2(V
SG
+ V
tp
). Note that the large-signal models are different for the two
MOSFET types because their voltage polarities and current directions, are opposite. However, the small-
signal model is the same for the two types because it involves only variations. This common model is
shown in Fig. 8. Its parameters depend on the bias current I
D
as
D m
kI g 2 =
1
o
D
r
I
= (14)
At the same bias current, the r
o
of a MOSFET is typically on the same order of magnitude as that of a
BJT. However, the g
m
of a MOSFET is typically one to two orders of magnitude smaller. This is due to
the fact that the i-v characteristic of the MOSFET is quadratic, while that of the BJTis exponential and
thus generally much steeper. For the same reason, the MOSFET is not as pronouncedly nonlinear as the
BJT, so the small-signal constraint of Eq. (13) is generally far more relaxed than for a BJT, for which v
be
<< 52 mV.
The relationships presented in Lab #4 for the generalized ac equivalent of the BJT hold also for
the MOSFET, the only difference being that the MOSFET has i
g
= 0 and generally much lower g
m
under
Fig. 8 Small-signal MOSFET model, valid for v
gs
<< 2(V
GS
V
t
).
2002 Sergio Franco Engr 301 Lab #5 Page 8 of 14
similar operating conditions. With reference to Fig. 9, the relationships are as follows:
The input resistance seen looking into the gate is
g
R = (15)
indicating that we can bias the gate of a MOSFET with resistances in the M range, if needed.
The output resistance seen looking into the source is
1
s
m
R
g
= (16)
which is generally low, but not as low as in the case of a BJT operating under similar conditions. For
a low-power MOSFET operating at I
D
= 1 mA, R
s
is typically in the k range. As we know, a BJT
has r
e
(1 mA) = 26 .
The output resistance seen looking into the drain is
( ) 1
d o m S
R r g R = + (17)
indicating that the presence of R
S
effectively raises the drain resistance from r
o
to R
d
.
The change in drain current i
d
stemming from a small-signal change in the gate voltage v
g
is
expressed as i
d
= G
m
v
g
, and we have
1
m
m
m S
g
G
g R
=
+
(18)
indicating that the presence of the source resistance R
S
introduces source degeneration.
We conclude by illustrating the use of PSpice to simulate a common-source (CS) amplifier. You
can simulate this circuit on your own by downloading its appropriate files from the Web. To this end, go
to http://online.sfsu.edu/~sfranco/CoursesAndLabs/Labs/301Labs.html, and once there,
click on PSpice Examples. Then, follow the instructions contained in the Readme file. Our PSpice
circuit is shown in Fig. 10a. After directing PSpice to perform the Bias Point Analysis, we obtain the
Fig. 9 Generalized AC equivalent.
2002 Sergio Franco Engr 301 Lab #5 Page 9 of 14
labeled schematic of Fig. 10b. Moreover, after directing PSpice to perform a one-point AC Analysis at f =
10 kHz, we find that the small-signal gain of the circuit is A
v
= v
o
/v
i
= 2.55 V/V. You will find it quite
instructive to confirm the above data (both bias and AC) via hand calculations!
Part II Experimental Part
We shall perform our experiments using the CD4007UB MOSFET Array. This array consists of
three nMOSFETs and three pMOSFETs, all of the enhancement type, interconnected in the manner
depicted in Fig. 11. Recall that the body of the nMOSFETs (pin 7) must always be held at the most
negative voltage (MNV) in the circuit; while the body of the pMOSFETs (pin 14) must always be held at
the most positive voltage (MPV) in the circuit. Failure to respect these constraints may invalidate all
measurements taken.
The CD4007 is a delicate device. To avoid damaging it, make sure that you always turn power
off before making any circuit changes, and that before reapplying power, each lab partner performs a
separate check that the circuit has been wired correctly. Also, refer to the Appendix for useful tips on
0V
0V
v s
1Vac
0Vdc
C1
0.1uF
RG
10Meg
C2
1uF
10.00V
0V 4.497V
VDD
10Vdc
M
301nMOSFET
0
RD
10k
4.497V
V
RL
47k
v s
1Vac
0Vdc
C1
0.1uF
RG
10Meg
C2
1uF
VDD
10Vdc
M
301nMOSFET
0
RD
10k
V
RL
47k
(a) (b)
Fig. 10 Common-source amplifier and its DC voltages
Fig. 11 The CD4007 MOSFET Array.
2002 Sergio Franco Engr 301 Lab #5 Page 10 of 14
how to wire proto-board circuits. The hand calculations and PSpice simulations that you will perform in
this and the following lab rely on the measurements taken for a particular CD4007 sample, so if you are
careless and end up damaging it, youll have to perform each measurement all over again!
You are urged to compare the data you are measuring on your particular sample against those
reported in the data sheets, which are typical in that they were obtained by averaging over a large number
of samples. To download the data sheets from the Web, go to http://www.google.com and search for
CD4007UB or variants thereof.
MOSFET Characterization:
MC1: We use the circuits of Fig. 12 to find
n
and
p
. Mark one of the CD4007 ICs (the other is a
spare). Turning first to the nMOSFET, assemble the circuit of Fig. 12a with power off (for R
D
use 2 10
k resistors in parallel). Using your multi-meter as a digital current meter (DCM), insert it in series
between R
D
and the drain D, as shown. Apply power and adjust the potentiometer for I
D
= 1 mA. With
reference to Fig. 13, this will bias M
1
at the operating point Q(I
D
, V
DS
) = Q(1 mA, 5 V). Now short out
R
D
with a wire (that is, close SW) to change the operating point from Q to Q with V
DS
= 5 V, and then
measure the corresponding change I
D
. Expect the latter to be small, so use as many digits as your
ammeter will allow. Finally, find r
o
= V
DS
/I
D
and
n
= 1/(r
o
I
D
V
DS
). Are your values typical?
MC2: Repeat Step MC1 for the pMOSFET of Fig. 12b, except that you now find your parameters as r
o
=
V
SD
/I
D
, and
p
= 1/(r
o
I
D
V
SD
). Again, are your values typical? How do their magnitudes compare
with those of the nMOSFET of Step MC1?
M3: We use the circuits of Fig. 14 to find k
n
and V
tn
, and k
p
and V
tp
. Turning first to the nMOSFET,
assemble the circuit of Fig. 14a with the DCM in series with the drain D. Obtain v from one of your
variable power supplies, and measure and record i for v = 0 V, 1 V, 2 V, 3 V, 4 V, and 5 V.
(a) (b)
Fig. 12 Test circuits to find
n
and
p
.
2002 Sergio Franco Engr 301 Lab #5 Page 11 of 14
C4: Consider the data of Step M3 for which i > 0, and plot them on an x-y graph, with x = v and
/(1 )
n
y i x = + (use the value of
n
found in Step MC1.) Hence, find the best fit straight line, determine
its slope dy/dx, and obtain k
n
= 2(dy/dx)
2
. Moreover, obtain V
tn
as the value of x where your line
intercepts the x-axis. Are your values typical?
M5: Repeat Step M3 for the pMOSFET of Fig. 14b.
C6: Consider the data of Step M5 for which i > 0, and plot them on an x-y graph, with x = v and
/(1 )
p
y i x = + (use the value of
p
found in Step MC2.) Hence, find the best fit straight line, determine
its slope dy/dx, and obtain k
p
= 2(dy/dx)
2
. Moreover, obtain V
tp
as the negative of the value of x where
your line intercepts the x-axis. Are your values typical? How do they compare with those of the
nMOSFET?
(a) (b)
Fig. 14 Test circuits to find k
n
, V
tn
, k
p
, and V
tp
.
Fig. 13 Graphical illustration
for finding r
o
and hence
n
and
p
.
2002 Sergio Franco Engr 301 Lab #5 Page 12 of 14
Common-Source Amplifier:
C7: Using the large-signal nMOSFET model with the parameters just measured, predict the DC voltages
V
G
, V
D
, and V
S
at the gate, drain, and source terminals of the common-source amplifier of Fig. 15. Hence,
find the DC drain current I
D
, and predict the value of the small-signal gain A
v
= v
o
/v
i
, which in this case is
A
v
= g
m
(r
o
//R
D
) (19)
with g
m
given in Eq. (14).
M8: With power off, assemble the circuit of Fig. 15, but without connecting the waveform generator
yet. Apply power, and use your DVM to measure the DC voltages V
G
, V
D
, and V
S
. Next, connect the
waveform generator, and while monitoring v
i
with Ch.1 of the oscilloscope, adjust the generator so that v
i
is a 10-kHz sine-wave with 0-V DC and 1-V peak-to-peak amplitude. Finally, use Ch. 2 (set on AC) to
measure the peak-to-peak amplitude of v
o
, and then find the experimental gain A
v
= v
o
/v
i
of your amplifier.
S9: Simulate the circuit of Step M8 via PSpice (DC as well as AC analysis). For a realistic simulation
you need to create a PSpice model for your specific nMOSFET sample, using the above-measured values
of k
n
, V
tn
, and
n
. To this end, when in PSpice, click the device to select it, then click Edit PSpice
Model to change the values of its parameters to those found experimentally.
Note: From the above measurements you have no way to know the values of W
n
and L
n
. What
you can do is to arbitrarily assume W
n
= L
n
(say, 5- m each), and then set the model parameter Kp to
the value you have found for k
n
. By this simple trick, PSpice will end up using the correct value for k
n
.
C10: Compare the predicted values of Step C7 with the measured ones of Step M8 and the simulated
ones of Step S9. Account for possible discrepancies.
M11: Returning to the circuit of Fig. 15, switch Ch. 2 to the DC mode (make sure you know where your
0-V baseline is on the screen!), change the waveform generator from sine-wave to triangle-wave, and
increase its amplitude until v
o
distorts both at the top and at the bottom. Justify the two types of
distortion in terms of transistor operation.
Fig. 15 Common-source
(CS) amplifier.
2002 Sergio Franco Engr 301 Lab #5 Page 13 of 14
Common-Drain Amplifier:
C12: Assuming v
i
has a DC value of 0 V in Fig. 16, use the large-signal model to find the DC drain
current I
D
. Hence, predict the value of the small-signal gain A
v
= v
o
/v
s
, which in this case is
1
1
1
( // )
v
m S o
A
g R r
=
+
(20)
Also, predict the value of the output resistance seen by the load, which in this case is R
o
= (1/g
m
)//R
S
.
M13: With power off, assemble the circuit of Fig. 16 (dont connect R
L
yet.) Keep leads short, and
mount the 0.1-F power-supply bypass capacitors in close proximity to your IC. Next, apply power,
adjust the waveform generator so that v
i
is a 10-kHz sine-wave with 0-V DC and a peak-to-peak
amplitude of 2 V, and measure the gain A
v
= v
o
/v
i
. How does it compare with the predicted value of Step
C12? What happens if you connect the load R
L
to your circuit? Is loading noticeable? Explain!
CMS14: Simulate the circuit of Fig. 16 via PSpice (DC as well as AC analysis), compare with predicted
and measured values, and account for any possible discrepancies.
CMOS Amplifier:
A widely used MOSFET amplifier configuration is the complementary MOS (CMOS) cell of Fig. 17,
consisting of an nMOSFET and a pMOSFET with the gates tied together to form the input node, and the
drains tied together to form the output node. Biasing is accomplished by a plain feedback resistance R
F
,
which can be made relatively large as the gate terminals draw zero current. Using the MOSFETs large
signal models, one can prove that with this resistor in place, the output and input nodes will automatically
bias at the common DC value V
O
= V
I
= V
BIAS
, where
/ ( )
1 /
tn p n DD tp
BIAS
p n
V k k V V
V
k k
+
=
+
(21)
Fig. 16 - Common-drain
(CD) amplifier.
2002 Sergio Franco Engr 301 Lab #5 Page 14 of 14
Moreover, using the small-signal models, one can prove that the ac gain around this bias point is
( 1/ ) ( // )
1 ( // ) /
mn mp F on op
o
v
i on op F
g g R r r
v
A
v r r R
+
= =
+
(22)
where g
mn
, g
mp
, r
on
, and r
op
are calculated via Eq. (14). To find I
D
, we use Eq. (9) with v
GS
= V
I
.
As you assemble (with power off!) the circuit of Fig. 17, keep leads short and bypass the power
supply with the usual 0.1-F capacitor. Anticipating a relatively high gain, we need to keep the input v
i
suitably small, so we interpose the voltage divider R
1
and R
2
between the input source and the amplifier to
suitably scale down the source. With the resistance values shown, we have v
i
v
s
/100.
C15: Using the MOSFET parameter values found experimentally, predict the values of V
BIAS
, I
D
, and A
v
for the CMOS amplifier of Fig. 17.
M16: Apply power to the circuit of Fig. 17, but without connecting the waveform generator yet, and use
the DVM to measure the output DC voltage V
O
. Next, connect the waveform generator, and set it for a
10-kHz, 0-V DC sine-wave. While monitoring v
s
with Ch. 1 and v
o
with Ch. 2 of the oscilloscope, adjust
the waveform generator until that v
o
has a peak-to-peak amplitude of 2 V. Measure the corresponding
peak-to-peak amplitude of v
s
, find v
i
= v
s
/100, and finally determine the voltage gain A
v
= v
o
/v
i
of your
amplifier.
S17: Simulate the circuit of Step M16 via PSpice (DC as well as AC analysis). For a realistic simulation
you need to create a PSpice model both for your nMOSFET and pMOSFET samples, using the above-
measured values of k
n
, V
tn
, and
n
, as well as k
p
, V
tp
, and
p
.
C18: Compare the predicted values of Step C15 with the measured ones of Step M16 and the simulated
ones of Step S17. Account for possible discrepancies.
Fig. 17 CMOS amplifier.