Design and Implementation of Operational Amplifiers With Programmable Characteristics in A 90nm CMOS Process
Design and Implementation of Operational Amplifiers With Programmable Characteristics in A 90nm CMOS Process
Abstract—Operational amplifiers (op-amps) serve as the basic like power management, circuit integrity, rule-based physical
building blocks in almost every analog and mixed-signal verification, and parasitic extraction may not just be enough.
electronic circuit. However, one of the most common problems in A designer’s strategy must shift from basic and fundamental
op-amp design is the variation in the op-amp's performance design techniques to new and optimal design methods which
caused by process variations during fabrication. As such, it is of may require changing the design itself.
utmost importance that provisions be made on the op-amp to
compensate for possible deviations in performance after chip During chip fabrication, there are unavoidable permanent
fabrication. In this paper, a procedure on integrating variations in the parameters of the design. Because of this,
programmable bias networks into an operational amplifier is solutions are being developed on how to minimize the effects
developed. The programmable network driven by digital input of the variations. One possible solution is to incorporate
words makes the output bias currents and voltages variable, thus programmable blocks into the chip. This makes certain
making the op-amp tunable for proper operation even after chip
parameters variable even after chip fabrication. In this paper,
fabrication. Furthermore, the programmability of other op-amp
digital input words (i.e. logic high or logic low) control the
parameters such as gain, slew rate, and compensation are
explored to increase the capability and flexibility of the op-amp. switches of the programmable blocks, making these parameter
The programmability schemes are employed on two base op-amp variations in discrete levels only.
topologies namely, the Two-stage Miller and the Folded Cascode.
The project is implemented in a standard 90nm CMOS process. II.BASE OP-AMPS
Part of this project is the development of two op-amps
I.INTRODUCTION which are used as base amplifiers for the programmable
speed and bandwidth. But this also makes a single chip more
intricate because of the increased complexity in the layout and
Fig. 1. Two-stage Miller op-amp schematic diagram
fabrication rules, making chip manufacturability a great
challenge. Minor process variations and parameter drifts
become major problems due to the chip’s high sensitivity to Figure 2 shows the schematic diagram used for the
implementation of the Folded Cascode op-amp.
physical defects. As a result, old design techniques are no
longer applicable to the present technology. Fundamentals
M9
M10 M16
(PCRC), and Programmable Current Source (PCS).
Programmable Current Attenuator (PCA)[1]
Fig. 2. Folded Cascode op-amp schematic diagram Figure 4 shows a 4-bit digitally-programmed current
attenuator circuit. It is based from the conceptual scheme of
Snapshots of the layout of the two base op-amps are shown an R-2R ladder network. The value of the current passing
in Figure 3. through the legs of the circuit is determined by two major
factors: the division factor, D, introduced by the scheme itself,
5.88 μm 5.84 μm and the input digital word at the gates of the transistor pairs.
Two complementary outputs, I1 and I2 are produced at the
output node which is mainly dependent on the reference
current, Iref.
I1
I2
B1 B1 B2 B2 B3 B3 B4 B4
M1 M2 M8 M7 M12 M11 M14 M15
18.75 μm 24.41 μm
Iref/2 Iref/4 Iref/8 Iref/8
M3 M6 M10 M13
Iref/2 Iref/4 Iref/8
Iref
M4 M5 M9
PM 47.4o 7.721° R0
M24 M18
R2
210
Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on January 25,2025 at 12:14:29 UTC from IEEE Xplore. Restrictions apply.
in series or in parallel to achieve the needed current, Iout. This Programmable Gain Amplifier (PGA)[5]
type of transistor connection effectively increases or decreases
the total W/L ratio of each leg: increasing the number of Figure 9 shows a 4-bit digitally-programmed gain amplifier.
transistors in serial connection decreases the current flowing The PGA consists of a negative feedback gm-boosted source
through it while parallel connection increases the leg current. degenerated differential pair with resistive loads, RLOAD =
The architecture delivers pre-programmed binary weighted 50kΩ. The main concept of this approach is to change the
current levels at every current leg which are all summed up at value of the resistance R so that the gain of the amplifier will
the Iout node. The number of bits on logic high contributes to also change. The PGA has a fixed dominant pole at
the additive nature of the current Iout. thus having a constant GBWP at all sweeps of
toM9
the degenerated resistance. It is a stand-alone amplifier;
B0 B1 B2 B3
therefore no integration with the base op-amps was done.
M1 M2 M6 M14 M30
M3 M7 M15 M31
Vdd
M4 M16 M32
M1 M2 M3
M5 M33
M12
M60 M7
M8 M6 M9
M61
M3 M2 M1 M0 To op-amp
Fig. 8. Obtained results for the integration of PCA, PCRC, and PCS with the
Figure 12 shows a graph of the obtained results for the
Folded Cascode op-amp integration with the Two-stage Miller op-amp and Figure 13
for the integration with the Folded Cascode op-amp.
IV.OTHER PROGRAMMABLE TOPOLOGIES
To further increase the programmability and capability of
our op-amps, other programmable topologies are implemented
such as Programmable Gain Amplifier, Programmable Slew
Rate, and Programmable Compensation.
211
Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on January 25,2025 at 12:14:29 UTC from IEEE Xplore. Restrictions apply.
Fig. 12. Obtained results for the integration of PSR with the Two-stage Miller Fig. 16. Obtained results for the integration of PC with the Folded Cascode
op-amp op-amp
Fig. 15. Obtained results for the integration of PC with the Two-stage Miller
op-amp
212
Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on January 25,2025 at 12:14:29 UTC from IEEE Xplore. Restrictions apply.