Adam Cooman

Previous

Distortion Contribution Analysis
with the Best Linear Approximation

Next

Example 1: BLA of a Miller op-amp in feedback

Before we use the BLA in a DCA, let us illustrate how the BLA is used to describe the behaviour of an op-amp placed in a negative feedback configuration as shown in Figure E1.1.

opamp under test

Figure E1.1 The op-amp under test is a two-stage Miller op-amp designed in a commercial 0.18μm0.18\mathrm{\mu m} technology with a Gain-Bandwidth product of 10MHz10\mathrm{MHz} for a load capacitance of 10pF10\mathrm{pF}.

The reference signals are lowpass random-odd RPM (2.3) with f0=fmin=0.1kHzf_{0}=f_{\mathrm{min}}=0.1\mathrm{kHz} and fmax=100kHzf_{\mathrm{max}}=100\mathrm{kHz} (Figure E1.2). The amplitude of the multisines is chosen flat as a function of frequency and such that the RMS voltage equals 50mV50\mathrm{mV}.

opamp under test

Figure E1.2 Spectrum of the reference signal.

The input and output voltages obtained with a Harmonic Balance (HB) simulation clearly contain non-linear distortion (Figure E1.3), as there is energy appearing at non-excited frequency lines. Even-order distortion (blue) and odd-order distortion (red) are separated by using the odd multisines. The in-band odd non-linear distortion is visible on the detection lines. The BLA obtained with 77 different-phase multisines is shown in Figure E1.4. A compression of 0.1dB0.1\mathrm{dB} is observed with respect to the results obtained with an AC simulation.

opamp under test

Figure E1.3 Spectra of the signals at the input and output of the op-amp.

opamp under test

Figure E1.4 Obtained BLA and its uncertainty σBLA\sigma_{BLA}. Due to compression, the BLA differs from the linearised response GACG_{AC} obtained with an AC simulation.

Previous
1  2  3  5  6  7  8  9  10  11  12  13  14  15  16  17  
Next