Designs › Fully differential › Op amps and OTA cores
Two-stage OTA, class-AB output
This differential amplifier uses an output stage that draws little current at rest but can deliver more during a large signal change. Complementary followers drive the two outputs, while a separate loop controls their average voltage. The extra circuitry requires careful biasing and compensation.
How it works
A low-power fully differential OTA with a class-AB output. A PMOS pair on diode loads is mirrored, with a gain of B, into the cascoded nodes d1 and d2. On each side a driver moves a stack in which two diodes hold the gates of complementary output followers a fixed voltage apart, so the followers idle at a current set by counts of devices, and either can carry far more when the output needs it. A common-mode loop trims the first-stage PMOS sources through a diode it feeds.
The bench draws Ib out of the PMOS diode MB, whose gate every PMOS source shares, so each current is a count of units. MB1 copies it seven times through RN1, RN2 into the NMOS diode MB2, which puts the NMOS cascode gate vcn a fixed drop above vgn; MB3 pulls the same through RP1, RP2 from MB4 for the PMOS cascode gate vcp. MTC copies MB2. The loop sets M31's current until the outputs sit at vocm, 0.6 V by default; at balance both of its pair's drains sit one diode drop below vdd at the same density.
Signal path
- Input pair (MT, M1, M2, M3, M4): MT's 4 Ib splits between the PMOS pair M1, M2, whose n-wells sit on their sources; the diodes M3 and M4 turn each half into a gate voltage.
- Mirrors and cascodes (M5, M6, M7, M8, M9, M10, M11, M12, M13, M14): M5 and M6 copy the diode currents B times through the NMOS cascodes M7, M8 from d1 and d2, which PMOS sources load under the cascodes M13, M14: M9, M10 fixed, M11, M12 set by the common-mode loop.
- Class-AB driver (M21, M15, M17, M19, CC1, RZ1): M21 amplifies d1 against the source M15; the diodes M17 and M19 in between keep gn1 and gp1 a fixed voltage apart; CC1 in series with RZ1 wraps M21 from gp1 back to d1. The other side, M22 with M16, M18, M20, matches.
- Output followers (M23, M25, M24, M26): Complementary source followers: M23 pushes current into outn and M25 pulls it out, each turning on harder as the stack moves its way; M24 and M26 serve outp.
- Common-mode detector (RS1, RS2, CS1, CS2): The resistors average outp and outn onto vsen; the capacitors across them keep the average from lagging.
- Common-mode amplifier (MTC, M27, M28, M29, M30, M31): M27, M28 compare vocm with vsen on the 2:1 mirror M29, M30; the current left, I28 − I27/2, runs into the diode M31, whose copies M11, M12 source more into d1, d2 as the outputs rise, which pulls the outputs back down.
Key relations
- Differential gain:
A_d ≈ B·g_m1·R_d·g_m21·R_gp. R_d is the cascoded resistance at d1 and R_gp the driver's output resistance at the stack; the followers pass the result at about one. - Unity-gain frequency:
ω_u ≈ B·g_m1/C_c. C_c is CC1, wrapped around the driver alone. - Class-AB bias:
V_GS23 + V_SG25 = V_GS17 + V_SG19. A translinear loop: the followers are A copies of the diodes and share their source voltage, so they idle at A times the stack current. - Follower pole and its zero:
1/(R_z·C_c) ≈ (g_m23 + g_m25)/C_L. R_z is RZ1: its zero lands near the followers' output pole, which the Miller capacitor does not enclose. - Stack headroom:
V_DD ≥ V_SD15 + V_GS17 + V_SG19 + V_DS21. Wide diodes and followers in weak inversion keep the middle two small; the driver near its zero-temperature-coefficient point keeps d1 high enough for the mirror and cascode beneath it.
Trade-offs
- Stack current S: more gives the driver the transconductance a small nulling resistor needs, and cleaner large steps, for more supply current.
- Level-shift and follower widths: wider lowers their gate-source voltages, which is what fits the stack in 1.2 V, and adds capacitance at gn1 and gp1.
- Driver width: narrower raises d1, the room the mirror and its cascode share, and raises the driver's own saturation voltage against gp1.
- Miller capacitor: larger adds margin and slows the slew; the nulling resistor must follow the followers' output pole.
Testbenches and limits
- Operating point: The amplifier as a unity-gain differential buffer, its inputs at their common mode: the current it draws, where its output common mode lands against the reference on vocm, and the differential offset.
- Open-loop gain: Differential gain and phase with the loop closed only at DC, through 1 TH inductors, so the operating point is the buffer's and every frequency sees the open loop.
- Common-mode loop gain: Each common-mode loop opened at the break point the circuit brings out, by Middlebrook's double injection - a voltage in series, then a current into the controlled side - with the differential loop closed as in the buffer: the loop's gain, its crossover, and its phase margin.
- Common-mode step: A 50mV step on the output common-mode reference, vocm, with the amplifier a unity-gain differential buffer: the output common mode follows it, and rings if its loop is short of margin.
- Step response: A 200mV differential step into the unity-gain differential buffer: how fast the differential output follows, whether it rings, and how far the output common mode is pushed on the way.
- Supply rejection: A ripple on V_DD, and how much of it reaches the unity-gain buffer's differential output and its output common mode.
- Common-mode rejection: The same signal on both inputs, on the unity-gain buffer's input common mode: what reaches the differential output is the common-mode to differential gain over the differential one, and what reaches the output common mode is how firmly the common-mode loop holds it.
- Noise: Differential noise referred to the input of the unity-gain buffer: its density at 1 kHz and its total from 1 Hz to 1 MHz.
| Bench | Figure | Limit |
|---|---|---|
| Operating point | Supply current | ≤ 80 µA |
| Operating point | Output common mode less its reference | ≥ -15 mV, ≤ 15 mV |
| Open-loop gain | DC differential gain | ≥ 70 dB |
| Open-loop gain | Unity-gain frequency | ≥ 2 MHz |
| Open-loop gain | Phase margin | ≥ 60 deg |
| Common-mode loop gain | Common-mode loop: phase margin | ≥ 55 deg |
| Common-mode step | Common-mode settling, 1 % | ≤ 600 ns |
| Step response | Slew rate | ≥ 1 MV/s |
| Step response | Settling time, 1 % | ≤ 400 ns |
| Supply rejection | Supply rejection at 1 kHz, differential | ≥ 70 dB |
| Supply rejection | Supply rejection at 1 kHz, common mode | ≥ 45 dB |
| Common-mode rejection | Common-mode rejection at 1 kHz | ≥ 70 dB |
Design variables and defaults
| Variable | Default |
|---|---|
| Reference current | 1 µA |
| W input pair (M1, M2) | 40 µm |
| L input pair | 2 µm |
| W NMOS unit | 2 µm |
| L NMOS unit | 2 µm |
| W PMOS unit | 4 µm |
| L PMOS unit | 2 µm |
| W common-mode PMOS unit (M11, M12, M29-M31) | 2 µm |
| L common-mode PMOS unit | 1 µm |
| Mirror gain into d1, d2 | 2 |
| Stack current, in Ib | 8 |
| Followers, copies of the level-shift diodes | 1 |
| W NMOS level-shift unit (M17, M23) | 42 µm |
| W PMOS level-shift unit (M19, M25) | 200 µm |
| L level shifters and followers | 300 nm |
| W driver (M21, M22) | 700 nm |
| L driver | 500 nm |
| W common-mode pair (M27, M28) | 4 µm |
| L common-mode pair | 1 µm |
| PMOS cascode drop, each of two (RP1, RP2) | 19.5 kΩ |
| NMOS cascode drop, each of two (RN1, RN2) | 17 kΩ |
| Miller capacitor, per side | 3.2 pF |
| Nulling resistor, per side | 17 kΩ |
| Detector resistor, per side | 200 kΩ |
| Detector capacitor, per side | 250 fF |
| Load capacitance, per side | 2 pF |
| Input common mode | 400 mV |
| Output common mode asked for | 600 mV |
Ports
inpinputinninputoutpoutputoutnoutputvddsupplyvssgroundvocmbias: the output common mode asked foribbias: the reference current, drawn out of this pincmobias: the common-mode amplifier's output, the diode M31: the testbench joins it to cmg, and breaks the loop therecmgbias: the gates of the sources M11, M12 the loop sets
Reference
Y.-P. Hsu, Z. Liu, M. M. Hella, A 1.8 uW -65 dB THD ECG Acquisition Front-End IC Using a Bandpass Instrumentation Amplifier With Class-AB Output Configuration, 2018. IEEE Transactions on Circuits and Systems II: Express Briefs, 65(12):1859-1863.
The topology: a PMOS pair with diode loads mirrored with gain into cascoded nodes, a common-source driver moving a two-diode level shifter whose ends drive complementary source followers, PMOS devices of the class-AB stack in wells on their own sources, and a five-transistor error amplifier that sets the first-stage PMOS sources from the averaged outputs. The error amplifier's current output into a diode, the bias with its resistor-set cascode drops, the Miller capacitor across the driver, the sizing and the targets are this library's, for SG13G2 at 1.2 V.
IHP SG13G2 130 nm. Simulations run in your browser; open the workbench to run this design's benches and change its variables.