A 10μW Hybrid OTA with Bulk-Biasing Driving 20nF Capacitive Load with 115 dB DC-gain and 1.68MHz GBW

A 10W Hybrid OTA with Bulk-Biasing Driving 20nF Capacitive Load with 115 dB DC-gain and 1.68MHz GBW

M.Manasa P.Prasad Rao

1. ech Scholar,

   Department of ECE

   HOD, Department of ECE

   Vaagdevi Engineering College (JNTUH) Vaagdevi Engineering College (JNTUH) Warangal,Bollikunta-506005 Warangal,Bollikunta506005

   AbstractFor better area and power efficient, single stage amplifiers are a better replacement of their multi-stage cmplifies ounterparts, particularly for display applications that imply massive buffer amplifier in their column drivers. This paper employs a signal-current enhancer technique for a single- stage amplifier to achieve significant enhancement of DC gain, gain-bandwidth product (GBW) and slew-rate. Ana bulk- biasing technique employed to minimize the internal node capacitances and area of OTA. Analytical treatments of the signal-current enhancer technique in terms of performance limits and robustness reveals that hybrid OTA can suppress the fundamental power-efficiency limit set by the basic differential- pair amplifier. The proposed hybrid OTA is implemented in a standard PTM 65 nm CMOS technology. Working under a 0.5V supply and proven by simulation results, the resultant OTA when driving an output capacitor of 20nF attained 115 dB DC gain, a GBW of 1.68 MHz and an Slewrate of 0.15 V/µs, consuming only 10µw of current and requiring no compensation capacitor.

   KeywordsAmplifier, frequency response, large capacitive loads, transient response, signal- current enhancement .

   1. INTRODUCTION

      For many modern applications require high-gain and fast- settling operational transconductance amplifiers (OTAs) driving off-chip loads in the order of hundreds (or even thousands) pico farads, such as high accuracy modulators, analog-to-digital converters, low-dropout regulators, headphone drivers and display applications like wide- dimension low-temperature polysilicon (LTPS) LCD panels

      [1] that involve thousands of buffer amplifiers in their column drivers, the area and power budgets of each buffer amplifier are extremely tight to meet the market pressure on cost and display quality [1]. The buffer amplifiers should drive a wide range of capacitive load (CL) up to tens of nano-farad (nF), while securing adequately large DC and output swing. Battery-operated portable systems integrate most of these blocks and, in addition, they work under tight power and area budgets.

      In this context the design of high-gain OTAs driving heavy capacitive loads is a challenging task, especially when nanometers technologies are adopted, as they suffer from a drastic reduction of the intrinsic gain that can be only partially mitigated by the adoption of non-minimum channel length transistors. In principle, stacked-device (i.e., cascode) topologies provide high DC gains but they reduce the output

      swing and have been progressively abandoned for this reason. Presently, the only viable solution to get DC gains in excess of 80 dB in scaled technologies is the adoption of multistage architectures, where simple gain stages are exploited. Since they provide very high DC gains without sacrificing output swings. Of course, the simplest gain stage available is a common source amplifier entailing only two complementary MOS transistors between the supply rails, though this choice limits the number of topologies available to the designer. However, the need for frequency compensation increases their design complexity, which also restricts their drivability of (range and size), area and power efficiencies.

      The design of multistage amplifiers is also complicated by the increased number of high impedance nodes (and, in turn, of low frequency poles and zeros in the loop-gain transfer function) which compromise stability, especially when the capacitive load is heavy. An OTA consuming power in µW- range generates tiny signal currents by input differential pair. By effectively enhancing these signal currents, both the GBW and gain can be significantly improved. This paper presents a signal-current enhancer to enhance the magnitude of the signal currents generated by the input differential pair. The signal- current enhancer is able to augment both the voltage gain and GBW even when non cascode structure is engaged.

   2. OPERATIONAL TRANSCONDUCTNACE AMPLFIER WITH SIGNAL-CURRENT ENAHCNER.

      In this section, the basic operation of the signal-current enhancer circuit with bulk biasing is described for OTA design.

      A. Concept of Signal-Current Enhancer

      The signal current enhancer circuit as shown in Fig. 1. In circuit IB indicates the DC bias current and is shows small signal current generated from input differential amplifier. This cell consumes a current of 2(k+1)IB. The OTA was designed for driving a large capacitive load by cascading of signal- current enhancers. The bulk terminal PMOS transistor can access directly without effecting other device performance. The minimum supply voltage require for each enhance circuit is VGS+VDS(sat) . With bulk biasing of PMOS gate-to- source voltage can be reduced for the same value of current. Which is useful to design amplifier with low-supply voltage (< 1 V).

      enhanced by the signal-current enhancer by (2k+1)6 times. The signal current is additionally scaled up by the transistor size ratio (i.e m) of the current mirrors to generate an output signal current io of m(2k+1)6gm1vin. The enhanced io will be used to charge/discharge CL according to the change of vin so that the GBW is also widened by m(2k+1)6 times to yield

      GBW m 2k 16 gm1

      (4)

      C

      C

      L

      m1

      m1

      The voltage gain likewise boosted by m(2k+1)6 time to give

      vout vin

      m 2k 16 g

      rNo2

      / /rPo2

      (5)

      Fig. 1 Signal-current enhancer with bulk biasing indicated with dc and small signal currents

      Where rNo2 and rPo2 are the rain resistance of MNo2 and MPo2, respectively. The total current IDD of the proposed OTA is hence

      Here, six signal-current enhanced cascaded to achieve high gain and to drive capacitive loads. The arrangement of cells are shown Fig. 2. The dc and signal current at the output of Cell 2 are IB and ±(2k+1)2is , respectively. Similarly, the dc and signal currents at the output of Cell 6 become IB and

      IDD 6 2k 1 m 4IB

      C. Largel-Signal Analysis of OTA with Signal-Current Enhancers

      (6)

      ±(2k+1)6is respectively. The dc ration, signal gain, and supply current (IDD) of the six cascade structures are, therefore, given by

      I I

      The large signal operation of the signal-current enhance is shown in Fig. 4. Two different cases are considered: having 2IB at one input and 0 current at other, and vice versa. It is noted that the output current of the enhancer are 0 and 2(k+1)IB, respectively.

      dc ratio= BOUT B 1

      (1)

      The large-signal responses [i.e., the positive slewing (SR+ )

      IBIN IB

      2k 16 i

      and negative slewing (SR )] of the proposed OTA are shown in Fig. 5. When a rising step input is applied to Vip + , the tail current goes entirely to M1 and so the input currents to Cell 1 are exactly the case shown in Fig. 4(a). The input currents 2IB

      signal gain

      s 2k 16

      is

      (2)

      and 0 are processed by Cell 1 to be the output currents 0 and

      2(k+1)IB. Based on the concepts presented in Fig. 4(a) and (b), after going through all the six cells, the output currents of the

      IDD

      6 2k 1 I

      B

      (3)

      six cascade signal-current enhancers are 2(k+1)6IB and 0, respectively. Since te input current of MN01 is 0, only MP01 and MPo2are turned ON, and the current of 2(k+1)6IB from Cell

      The proposed design uses less supply current, voltage and

      because of bulk bias less parasitic poles generated by signal nodes. Since for the bias current IB PMOS require small geometrics. In principle, the maximum achievable bandwidth of the proposed amplifier is higher due to less parasitic effects.

      B. Small-Signal Analysis of OTA with Signal-Current Enhancers

      In Fig. 3, the cascade signal-current enhancer is applied to a classical OTA by inserting it between the input differential

      6 is further scaled up by the m-ratio of MPo1 and MPo2. The

      output transient current is thus 2m(k+1)6IB, and the SR+ is also heightened by m(k+1)6 times as shown in Fig. 5(a), when comparing with a standard OTA. Similarly, as shown in Fig. 5(b), when a rising step input is applied to Vin the SR can also enjoy the same boosting factor. As such, the proposed circuit can enhance the transient current on its own, without SR enhancement circuit, to achieve

      2I

      pair and the output stage. When an input differential signal vin

      SR m k 16 B

      (7)

      is applied to the input stage is=0.5gm1vin where gm1 is the transconductance of M1 is generated, and is is further

      CL

      Fig. 2 Signal current enhancer (six cascade cells)- indicated both dc and signal currents

      Fig. 3 Hybrid OTA with bulk biasing current-enhancer

      Fig. 4 Large-signal responses of signal-current enhancer (a) Case 1: input currents are 2IB and 0. (b) Case 2: input current are 0 and 2IB

      The node resistances, node capacitances, and estimated locations of parasitic poles at node n0-n14 can be tabulated, as shown in Table I. In order to locate tall parasitic poles beyond

      the GBW to achieve sufficient phase margin (PM), the values of k and m cannot be too large. Therefore, k=2 is selected to provide an improvement factor of (2k+1)6 =15,625 for GBW and voltage gain, and to achieve an enhancement factor of (k+1)6=729 for the slewrate. These improvements are sufficient to have the OTA outperform the state-of-art designs.

      Table 1 Parasitic Poles in Hybrid OTA

      Nodes	Node resistance	Node capacitance	Parasitic Poles
      n1,n3,n5,n7,n9,and n11	1/gmp	(2k+2).Cgp	gmp/(2k+2).Cgp
      n13	1/gmp	(m+1). Cgp	gmp/(m+1).Cgp
      n0	1/gmp	2Cgp	gmp/2Cgp
      n2,n4,n6,n8,and n12	1/gmn	(2k+2).Cgn	gmp/(2k+2).Cgn
      n14	1/gmn	(m+1).Cgn	gmp/(m+1).Cgn

      Fig. 5 Large-signal responses of hybrid OTA (a) rising step input at Vip and (b) rising step input at Vin

      A summary on the performance of OTA with signal- current enhancer is tabulated in Table 2. The supply current increases linearly with k and a factor of 6 due to the six cascade cells used. Providentially, the GBW, voltage gain, and SR are all exponentially increased by a power of 6, which augment all significantly at only a slight increase in power.

      Table 2 Summary of Performances of Hybrid OTA

      Supply current	IDD={6[2(K+1)]+m+4}IB
      GBW product	GBW=m(2K+1)6 (gm1/Co)
      Voltage gain	VO/Vin=m(2k=1)6 gm1(rono2//roPO2)
      Slew rate	SR=m(k+1)6 (2IB/CO)

   3. SIMULATION RESULTS

      Amplifier in Fig. 3 implemented in 65 nm PTM CMOS technology. This CMOS technology has a nominal supply voltage of 0.5 V. The gate-to-source voltage of the NMOS and PMOS transistor are 0.25 and -0.25 V. The threshold voltage of PMOS transistor with a bulk bias voltage of 0.25 V is 0.32 V. The unit size of NMOS transistor is 2.5 µm/0.5 µm and that of PMOS transistor is 4 µm/0 .2 µm. The selected values of k and m are 2 and 3. The proposed OTA does not require compensation capacitor. The simulation results with 0.5V supply voltage and 20nF load capacitance CL are presented here.

      Fig. 6 shows the simulation frequency responses of bulk-

      slewrates are 0.15 V/µs for CL=20 nF, 0.075 V/µs for CL=40 nF, and 0.06 V/µs for CL=50 nF, respectively. The ovel all performance of realized amplifier is summarized in Table 3. The performance of the amplifier is compared with relevant design in Table 3.

      Fig. 6 Simulated open-loop frequency responses of bulk-biased Hybrid OTA

      Generally, the efficiency of a frequency compensation scheme is characterized by four commonly used figures of merit (FOM).

      biased hybrid OTA. B The obtained GBWs and PMs are 1.68 MHz and 530 for CL=20 nF, 0.877 MHz and 710 for CL=40 nF, and 0.7 MHz and 740 for CL=50 nF, respectively. Fig. 7 illustrate the frequency response of the OTA in unity-gain configuration. The low-impedance nodes (i.e., n0n14) in the proposed signal-current enhancer have minute effects on the stability of the proposed OTA since the parasitic poles are located at much higher frequencies than the GBW. Fig. 8 depicts the transient responses of the proposed OTA in unity- gain configuration with a step input of 0.5 V. The average

      FOMs

      FOML

      GBW.CL

      Power

      SR.CL ,

      Power

      , IFOMs

      IFOML

      GBW .CL

      IDD

      SR.CL

      IDD

      Table 3 Performance Summary and Comparison

      Parameters	IAC [2]	CFCC[3]	CBMC[4]	CLIA[5]	NCM[6]	Hybrid OTA[1]	This work
      Technology(µm)	0.35	0.065	0.35	0.13	0.18	0.13	0.065
      CL (pF)	150	500	15 000	680	500	10 000	20 000
      VDD (V)	1.5	1.2	2	1.2	1.2	0.7	0.5
      IDD (µA)	20	17	72	10.5	3	24	10µw
      Dc gain (dB)	110	>100	>100	>100	84	~100	115
      GBW(MHz)	4.4	2	0.95	3.37	0.396	1.99	1.68
      SR (V/µs)	1.8	0.65	0.22	0.67	0.0115	0.7	0.15
      PM (o)	57	52	52.3	45	81.4	47	53
      On-chip capacitance(pF)	1.6	1.15	2.6	0.587	0	0	0
      FOMS (MHz.pF/mW)	22 000	49 020	98 958	181 873	55 000	1 184 524	3 360 000
      FOML (V/µs.pF/mW)	9000	15 931	22 917	36 159	1597	416 667	300 000
      IFOMS (MHz.pF/mA)	33 000	58 823	197 916	218 247	66 000	829 166	1 680 000
      IFOML (V/µs.pF/mA)	p>13 500	19 118	45 834	43 391	1916	291 667	150 000

      Fig. 7 Simulated frequency responses of bulk-biased Hybrid OTA in unity- gain configuration.

      FOMs are actually inadequate to qualitatively evaluate a frequency compensation scheme, because the stability is also a critical criterion to impact the quality of a frequency compensation performance and should be taken into account for frequency compensation design.

   4. CONCLUSION

The operational transconductance amplifier is designed to drive a large capacitive loads. To operate the OTA under low-supply voltage bulk-biasing technique employed. A signal-current enhancers is employed in this paper. The cascade structure of six signal current enhancers is applied to a standard OTA to improve its GBW, SR, and voltage gain through the enhancements of the small-signal and transient output currents from the enhancers, The GBW and SR are improved by a factor of a m(2K+1)6 and m(K+1)6 , respectively, the simulation results have fully verified performances scheme.

Fig. 8 Step responses of Hybrid OTA

FOMS and FOML evaluate the compensation performance on a system level design point of view, while IFOMS and IFOML address an inspection on the transistor level, since both GBW and SR depend only on the bias current of a transistor. FOMS and IFOMS are the indications of the small-signal behavior of the amplifier, and FOML and IFOML illustrate the large- signal behavior of the amplifier. Nonetheless, these four

REFERENCES

1. K. H. Mak et al., "A 0.7-V 24 µA Hybrid OTA Driving 15 nF Capacitive Load With 1.46 MHz GBW," in IEEE Journal of Solid-State Circuits, vol. 50, no. 11, pp. 2750-2757, Nov. 2015.

2. X. Peng, W. Sansen, L. Hou, J. Wang, and W. Wu, Impedance adapting compensation for low-power multistage amplifiers,IEEE J. Solid State Circuits, vol. 46, no. 2, pp. 445451, Feb. 2011.

3. S. S. Chong and P. K. Chan, Cross feedforward cascode compensation for low-power three-stage amplifier with large capacitive load,IEEE J. Solid State Circuits, vol. 47, no. 9, pp. 22272234, Sep. 2012.

4. Z. Yan, P.-I. Mak, M.-K. Law, and R. P. Martins, A0.016-mm2 144-

   µW three-stage amplifier capable of driving 1-to-15 nF capacitive load with 0.95-MHz GBW,IEEE J. Solid State Circuits, vol. 48, no. 2, pp. 527 540, Feb. 2013.

5. M. Tan and W.-H. Ki, A cascode Miller-compensated three-stage amplifier with local impedance attenuation for optimized complex-pole control, IEEE J. Solid State Circuits, vol. 50, no. 2, pp. 440449, Feb. 2015.

6. Z. Yan, P.-I. Mak, M.-K. Law, R. P. Martins, and F. Maloberti, A 0.0013mm2 3.6µW nested-current-mirror single-stage amplifier driving 0.15-to-15 nF capacitive loads with >62 phase margin, inProc.IEEE Int. Solid State Circuits Conf. (ISSCC) Dig. Tech. Pap., Feb. 2014,pp. 288289.