Lab 5
The 5T OTA
Build Chapter 10's five-transistor OTA as a subcircuit and verify its hand design: gain, bandwidth, phase margin, slew rate, input range, offset and power.
Contents
| Lab 5 at a glance | |
|---|---|
| Goal | Build Chapter 10’s five-transistor OTA as a subcircuit and verify its hand design: gain, bandwidth, phase margin, slew rate, input range, offset and power. |
| Reproduces | Chapter 10 (Sections 10.5–10.8, Tables 10.1 and 10.2, Figures 10.6 and 10.7) |
| Time | 2–3 hours |
| Files | labs/lab5: ota5t.sch and ota5t.sym (the OTA), ota5t_ac_tb, ota5t_tran_tb, ota5t_icmr_tb (.sch) per PDK |
Hand calculation
From Table 10.1: \(g_{m1}\) = 320 µS and \(I_{tail}\) = 40 µA into \(C_L\) = 2 pF. Predict the GBW, \(g_{m1}/(2\pi C_L)\), and the slew rate, \(I_{tail}/C_L\). With the gate voltages of Table 10.1, evaluate Eq. (10.12) for the input common-mode range.
Steps
- Open the OTA.
ota5t.schis the OTA of Table 10.1 with six ports (vdd,vss,inp,inn,out,ibias), andota5t.symis its symbol. In the GUI, select the symbol in a testbench and presseto descend into the schematic,Ctrl+eto come back. Every testbench of this lab instantiates it asx1, so its internal nodes arex1.n1,x1.tailand its devicesx1.xm1and so on. - Gain, bandwidth, phase margin. Run
../../sim.sh ota5t_ac_tb.sch. As in Section 10.8, a 1 GH inductor closes the unity-gain loop at DC only and a 1 F capacitor grounds the inverting input for AC, so the AC analysis shows the open-loop gain at the buffer’s operating point. The testbench also prints the systematic offset and the power, including the 10 µA reference. - Slew rate. Run
../../sim.sh ota5t_tran_tb.sch: two buffers, one with a 300 mV step and one with a 20 mV step around \(V_{ic}\). It prints the steepest rising and falling slopes of the large step and the final value of the small step as a fraction of the step.plot v(in1) v(out1)shows Figure 10.7(a). - Input range. Run
../../sim.sh ota5t_icmr_tb.sch. It sweeps the buffer input and computes the saturation margins \(V_{DS} - V_{DS,sat}\) of M1 and M5 from their operating points; the ICMR is where both are positive.plot m1 m5shows Figure 10.6(a).
Checkpoint
Table 18.6. Lab 5 results (unity-gain buffer at Vic = 0.70 V and 1.10 V, CL = 2 pF, tt, 27 °C).
| Quantity | SG13G2 | SKY130 |
|---|---|---|
| DC gain, GBW, phase margin | 26.2 dB, 23.8 MHz, 87° | 41.7 dB, 24.5 MHz, 85° |
| Systematic offset, power | −1.07 mV, 59.7 µW | −4.31 mV, 87.9 µW |
| Slew rate, rising / falling | 20.4 / 17.7 V/µs | 19.4 / 21.9 V/µs |
| 20 mV step: final value / step | 0.951 | 0.987 |
| Input common-mode range | 0.49–0.91 V | 0.87–1.30 V |
Questions
- The final value of the small step is 0.951 in SG13G2. Show that this is \(A_0/(1 + A_0)\) for your DC gain. What does it say about using this OTA as a buffer?
- Why do the rising and falling slew rates differ, and why in opposite directions in the two PDKs?
- The SKY130 OTA has 50 % more supply but barely more input range. Where does the extra 0.6 V go?
Extension
Change the length of M3 and M4 in ota5t.sch to 2 µm (keep the width) and rerun ota5t_ac_tb: the gain should rise by less than 1 dB while the phase margin falls (Section 10.8, step 4). Then move \(V_{ic}\) to mid-supply in ota5t_ac_tb and find out from the operating point (print @n.x1.xm1.nsg13_lv_nmos[vdss] in SG13G2, @m.x1.xm1.msky130_fd_pr__nfet_01v8[vdsat] in SKY130, and the node voltages) which transistor leaves saturation first.