PDF 110
PDF 110
Abstract
The implementation of full-wave rectifiers using operational transconductance amplifiers (OTAs) as only main active
circuit elements is presented in this paper. The proposed scheme makes use of the characteristic of the differential amplifier
inside the OTA and avoids the use of diodes. The typical problems of the OTA circuit, the input voltage swing limitation and
the temperature dependence of the OTA transconductance gain, have been improved. Simulation and experimental results
demonstrate the performance of the proposed rectifier are included.
䉷 2006 Elsevier GmbH. All rights reserved.
reported rectifier circuits, diodes are mainly the important maximum common mode input voltage range, the output
key element to provide the rectifying action. It seems that current io becomes
none the exiting rectifiers has been built by using only OTAs.
The objective of this paper is to present a technique to real- io = IB . (2)
ize rectifiers using the characteristic of the differential am-
plifier inside the OTAs. The proposed full-wave rectifier are It should be noted that the current IB can only flow in the
therefore using OTAs as only the active circuit elements. direction that into the circuit, or positive direction. On the
A scheme to improve the input voltage swings restriction other hand, if vi is a constant DC voltage that much lower
and the temperature dependence of the OTA transconduc- than −2VT or vi > − 2VT , thus the output current io now
tance gain is also introduced. becomes
io = −IB . (3)
2. Basic principle
Eqs. (2) and (3) demonstrate that the direction or the po-
larity of the output current can be controlled by the DC volt-
Fig. 1(a) shows a monolithic OTA, where VCC , VEE de-
age that is applied at the vi . Therefore, we will propose the
note the positive and negative supply voltages, respectively.
use of the differential amplifier inside the OTA to realize
Consider an OTA typical circuit diagram as shown in the
rectifier circuits.
Fig. 1(b), its input stage consists of a differential amplifier
stage. Normally, IB is the bias current, and vi and io are the
input voltage and the output current, respectively. The out- 2.1. Basic half-wave rectifier
put current io can be given as
The basic of a half-wave rectifier circuit that using only
vi two OTAs is shown in Fig. 2. It consists of two OTAs and
io = ic1 − ic2 = IB tanh , (1)
2VT a constant voltage source VC . The OTA1 is functioned as a
voltage-to-current converter that converts an input voltage
where VT = KT /q. However, from (1), one can see that if
vin into a signal current i1 = gm1 vin , where gm1 = IB1 /2VT
IB is the input signal current and vi is a constant DC voltage
and IB1 is the bias current. The output current i1 of the OTA1
that much greater than 2VT or vi ?2VT , but lower than the
is fed as the bias current of the OTA2 . At the OTA2 and from
(1), when the constant voltage source VC ?2VT , the output
VCC current i2 can be expressed as
IB
g v for vin > 0,
vi io i2 = m1 in (4)
OTA 0 for vin < 0.
IB Q12 Q14 I B1
vin i1
Q4
OTA 1
I B2
Q3 Q5
VC i2
OTA 2
(b) VEE
2.2. Basic full-wave rectifier by controlling the polarity of VC . This advantage is usually
not available in the basic op-amp-based circuit.
As shown in Fig. 3, two of the half-wave rectifier circuits
are combined to form a full-wave rectifier circuit, where
(OTA1 and OTA2 ) and (OTA3 and OTA4 ) are the positive 3. The compensated OTA-based full-wave
and negative half-wave rectifiers, respectively. Ideally, let rectifier
the transconductance gains of the OTA1 and the OTA3 are
equal, such that gm1 = gm3 = gm , where IB1 = IB3 = IB and 3.1. Improve the limited input voltage swings and
gm = IB /2VT . For vin > 0, the current i1 and i3 of the OTA1
temperature dependence
and OTA3 can be, respectively, written as
i1 = gm1 vin = gm vin (6) It is clearly seen from (9) and (10) that the net result of
the output current is the full-wave rectification of the input
and signal. However, there are two major limitations that occur
from the use of the ordinary bipolar-based OTA in the circuit
i3 = −gm3 vin = −gm vin . (7) design. The first limitation is that the input stage of the OTA
The currents i1 and i3 are fed as the bias currents of the is a conventional differential pair, where the differential input
OTA2 and OTA4 , respectively. Noting that in the case of voltage is limited to be less than 50 mV for linear operation.
VC ?2VT , the OTA2 and OTA4 allow only positive current The second limitation is that the OTA transconductance gain
flow through it. This means that for vin > 0, the currents gm is inversely proportional to temperature. Therefore, based
i2 = i1 and i4 = 0 or on the compensation scheme reported by Surakampontorn
et al. [15], the temperature-insensitive wide-dynamic range
i2 = i1 = gm vin and i4 = 0 for vin > 0. (8a) OTA-based full-wave rectifier is shown in Fig. 4.
The OTA5 , which is formed as an electronically tunable
During vin < 0, the current i1 is negative and the current i2 resistor, and a resistor R1 function as a voltage divider cir-
is positive. Therefore, i2 = 0 and i4 = i3 or cuit. This makes the vin − vX voltage across the OTA5 small
i2 = 0 and i4 = i3 = gm vin for vin < 0. (8b) and therefore, increases the circuit dynamics range. The
voltage across the OTA5 is then used as the input voltages
Since the output current iout is the summation of the cur- for the OTA1 and OTA2 . From routine circuit analysis of
rent i2 and i4 , we can write the relations between the input Fig. 4, voltage at node X or vX can be written as
voltage vin and the output current iout as
gm5 R1
IB vX = vin (11)
iout = i2 + i4 = gm vin = |vin |. (9) 1 + gm5 R1
2VT
and the currents i1 and i3 can be expressed as
We will call this as a positive full-wave rectifier. Alterna-
tively, if VC > − 2VT , a negative full-wave rectifier can be i1 = −i3 = gm (vin − vX ), (12)
realized, where the output current can be expressed as
where gm1 = gm3 = gm . From (11) and (12), the currents i1
IB and i3 become
iout = −gm |vin | = − |vin |. (10)
2VT gm
i1 = −i3 = vin . (13)
Note that (9) and (10) indicate that the circuit of Fig. 3 1 + gm5 R1
can be operated as a positive or a negative full-wave rectifier
I B1
I B1 i1
vin i1 OTA 1
OTA1 I B5 I B2
I B2 vin i2
i2 VC
VC OTA5 OTA 2
OTA2
I B3 I B3
vX i3 i out
i3 i out i5 vout
OTA3
OTA 3 R1 I B4
I B4
i4 i4 R2
OTA4 OTA4
Fig. 3. Basic OTA-based full-wave rectifier. Fig. 4. The compensated OTA-based full-wave rectifier.
198 C. Jongkunstidchai et al. / Int. J. Electron. Commun. (AEÜ) 61 (2007) 195 – 201
If by design, we let gm5 R1 ?1, the currents i2 and i4 can for VC ?2VT and
be approximately given by
1 4 R2
Vout = − 1 − + 2 vin (19b)
gm IB vin 1 + gm5 R1 R1
i2 = −i4 = vin = . (14)
gm5 R1 IB5 R1
for VC > − 2VT .
Thus, the output voltage vout is equal to Thus, for = 100, gm5 = 0.0096S, R1 = 10 k, and R2 =
10 k, then the error of less than 2% is expected. The third
I B R2 factor is the temperature dependence of transconductance
vout = |vin | for VC ?2VT (15a)
IB5 R1 gains gm that affects the output current. The percentage error
of the output current for the compensated voltage-to-current
and
base on OTA [15] can be approximated as
I B R2
vout = − |vin | for VC > − 2VT . (15b) iout −2VT T
IB5 R1 = × 100%. (20)
iout 2VT + IB5 R1 T
Eqs. (15a) and (15b) show that the temperature de- From (20), it is clearly seen that the temperature influence
pendence of the transconductance gains gm1 and gm3 are can be reduced by increasing the value of IB5 R1 , where T is
compensated by gm5 . In addition, the input voltage swing the room temperature (300 K or 27 ◦ C) and T is the change
limitation has been improved from 2VT to approximately of temperature from room temperature. If VT = 26 mV at
IB5 R1 . The output voltage vout can be electronically tuned room temperature, IB5 = 500 A and R1 = 10 k and tem-
by the current ratio IB /IB5 . perature change at 50 and 70 ◦ C, then the percentage error
of the output current iout are about 0.0789% and 0.1475%,
3.2. Performance analysis respectively. The fourth factor is the dc output offset cur-
rent that arise from the non-ideal characteristic of the OTAs,
From the previous description, the OTAs were assumed such as nonzero-input bias and input-offset currents errors.
to be ideal. In this section, the major factors that contribute In this case, OTAs with low offset currents should be used,
to the inaccuracy of the circuit performance will be consid- in particular for the OTA1 and OTA3 . In addition, small DC
ered. The first factor is due to the transfer error from the currents should be injected to compensate the offset current,
conversion of the input voltage vin into the small signal cur- at the outputs of the OTA1 and OTA3 , since the DC currents
rent i5 by the OTA5 and R1 . The relationship of the input that presented at these points produce the distortion of the
vin and the current i5 can be accurately expressed as rectified signal.
gm
i5 = (1 − v ) vin , (16)
gm5 R1 4. Simulation and experimental results
where the percentage error of the voltage-to-current conver-
sion error (v ) can be given by In order to test the developed ideas and to verify the theo-
retical prediction, the performance of the circuit in Fig. 4 was
1
v = × 100%. (17) 5
1 + gm5 R1
200
v
2
(mV)
2
c = × 100%, (18) 100
2 + 2 + 2
out
v
1
where is the common emitter current gain of the bipolar 0
-200 -100 0 100 200
transistor [16]. As an example, if = 100, the error is about v (mV)
in
0.02%. With this approximation, we can write the relations 0
-6 -4 -2 0 2 4 6
between the input voltage vin and the output voltages vout as vin (V)
1 4 R2
Vout = 1 − + 2 vin (19a) Fig. 5. DC transfer characteristic of the proposed OTA-based
1 + gm5 R1 R1 full-wave rectifier.
C. Jongkunstidchai et al. / Int. J. Electron. Commun. (AEÜ) 61 (2007) 195 – 201 199
60
0
50 Vin = 500 mV
Vout (mV)
-200
Vin = 5 mV
-400
40
op-amp-based
-600
Basic circuit in Fig 3.
µA)
Vout (mV)
-200
20 Vin = 200 mV
-400
Vin = 2 mV
-600
10
0 0.5 1.0 1.5 2.0
(a) Time (ms)
0
0 10 20 30 40 50 60 70 80 90 100
Temperature (°C) 0
Vout (mV)
Fig. 6. Variation of output currents iout against temperature. -200
-400
600
op-amp-based
Vout (mV)
-600
400 OTA-based for Vc = -1V 0
200
Vout (mV)
-200
0 op-amp-based -400
600
OTA-based for Vc = 1V
-600
Vout (mV)
400 vin less than ±50 mV. This is due to the OTA5 and R1 that
200 formed as a voltage divider. For small vin the voltage across
the OTA5 that used as the input voltages for the OTA1 and
0 OTA2 is very small and cause the error at the output of the
rectifier.
0 50 100 150 200
Fig. 6 shows the comparison of the simulated output cur-
(b) µ s)
Time (µ
rents iout against temperature for the case of the basic and
Fig. 7. Simulation results of the positive full-wave rectifier at the the compensated circuits, for the input voltages vin of 2,
frequency: (a) 1 kHz; (b) 10 kHz. 5, 200 and 500 mV, respectively, where the nominal output
currents were set to 18 and 48 A, respectively. It is readily
seen that the temperature performance of the compensated
simulated using the PSPICE simulation program. We use circuit is much better than the basic circuit.
the conventional OTA LM13600 models from the PSPICE Figs. 7 and 8 show the positive and negative output volt-
library. The supply voltage VCC = −VEE were set to 15 V. ages of the conventional op-amps-based rectifier circuits is
200 C. Jongkunstidchai et al. / Int. J. Electron. Commun. (AEÜ) 61 (2007) 195 – 201
1
vin R D1
vout
OP1 OP2
D2 op-amp-based
R RL
2
(a)
R R R
OTA-based for VC = -1V
vin R D1 (a)
vout
OP1 OP2 CH1-200mV CH2-200mV 200us /div
D2 1
R RL
(b)
the rectifier.