Integrated Circuits and Systems group, IIT Madras

Amplifier compensation

  • Goals:
    • Realize high gain accuracy in a feedback amplifier
    • Preserve stability while doing so
  • Use a 6V supply for this experiment. You'll need the gm value measured in the previous experiment.

  • Determine the gain of the amplifier in (a) when the transconductance of the inverter tends to infinity. Build the amplifier in (a) with an ideal gain of 2 and R1 in 5kΩ to 10kΩ range. Use a 6V supply for this experiment. Measure the gain error. Connect inputs of unused inverters to ground
  • Apply a small(~100mVpp) squarewave and measure the step response. Is there an overshoot/ringing? Apply a small sinusoidal input, sweep its frequency, and determine the bandwidth-i.e. the frequency at which the gain is 1/sqrt(2) times the low frequency gain(“low frequency” should be above the cutoff frequency of the ac coupling network).

  • The gain error can be reduced by increasing the loop gain. Use another inverter in cascade (and a unity gain inverter for negative feedback) as shown above and measure the circuit. Does it result in a better amplifier? What do you see at the output?

  • Compensate the amplifier by connecting a capacitor C1 to ground as shown above. Adjust its value for 10% overshoot. While choosing the compensation capacitor, start from small values, of the order of 100pF. Determine the bandwidth.

  • Compensate the amplifier by connecting a capacitor C2(with a zero cancelling resistor in series) across the second stage as shown above. Adjust its value for 10% overshoot. Comment on the compensation capacitor values C1 and C2 in the two cases. While choosing the compensation capacitor, start from small values, of the order of 100pF. Determine the bandwidth.
    • Comment on the compensation capacitor value in the two cases.
    • Comment on the bandwidth in the three cases(single stage in feedback, and two compensated amplifiers).
  • Short out the zero cancelling resistor and test the step response.
  • Applications: A cascade of two transconductor stages(with perhaps an output buffer in some cases) is the most common opamp topology. Miller compensation is the most widely used method of frequency compensation for opamps. Although the transconductor topologies are different in an opamp, the principles of obtaining high gain and frequency compensation are exactly the same as what is done here. For instance see the schematics in the datasheets of LM324 of LF347 opamps which you'll be using later in this lab. You'll be using these steps in simulations/measurements of opamps and other feedback loops(such as the voltage regulator in the next experiment).