SBOA590 November   2024 OPA186 , OPA206 , OPA328 , OPA391 , OPA928

 

  1.   1
  2.   Abstract
  3.   Trademarks
  4. 1Input Offset Voltage (VOS) Definition
    1. 1.1 Input Offset Voltage Drift (dVOS/dT) Definition
    2. 1.2 VOS and VOS Temperature Drift Inside the Amplifier
    3. 1.3 Laser Trim to Adjust Performance
    4. 1.4 Package Trim (e-Trim™) to Adjust Performance
  5. 2Input bias current (IB) definition
    1. 2.1 Input Bias Current (IB) and IB Temperature Drift Inside the Amplifier
    2. 2.2 Derivation of IB Conversion to VOS
    3. 2.3 Internal Bias Current Cancelation
    4. 2.4 Super Beta Input Transistors
  6. 3Other Factors Influencing Offset
    1. 3.1 Finite Open Loop Gain (AOL)
    2. 3.2 Common Mode Rejection Ratio (CMRR)
    3. 3.3 Power Supply Rejection Ratio (PSRR)
    4. 3.4 AOL, CMRR, and PSRR Over Frequency
    5. 3.5 Electromagnetic Interference Ratio (EMIRR)
    6. 3.6 Mechanical Stress Induced Offset Shift
    7. 3.7 Parasitic Thermocouples
    8. 3.8 Flux Residue and Cleanliness
  7. 4Zero-drift Amplifiers to Minimize VOS and VOS Drift
  8. 5Calibration of VOS, IB, and Gain Error
  9. 6References
  10. 7Revision History

AOL, CMRR, and PSRR Over Frequency

Open-loop gain, common mode rejection ratio, and power supply rejection all decrease over frequency. Generally, this frequency response is a first-order response. That is, the response is flat at low frequency then begin to roll-off at 20 dB/decade afterwards. This roll-off over frequency means that the offset introduced by these parameters increases at higher frequencies. Furthermore, these bandwidth limitations are small signal responses. A small signal generally means that the signal needs to be less than 10 mV, but the requirement can be different depending on the design or technology. For larger signals, the response may differ from the specified small-signal response because of slew-rate limitation of the amplifier. These slew-rate limitations are sometimes referred to as full-power bandwidth limits.

Figure 3-8 is a simulation example that adds an AC signal on the positive supply. This circuit condition simulates a noise signal on the amplifiers power supply. The AC signal causes variations in the power supply over frequency, so PSRR is being tested. However, because the power supply change is non-symmetrical, CMRR limitations is also being exercised. Finally, because the output signal is not held constant, AOL limitations also affect the result. In addition to showing how AOL, CMRR, and PSRR change over frequency, this simulation also illustrates how simulation can be used to solve a complex issue with many factors interacting. Figure 3-9 compares the PSRR curve from the data sheet to the PSRR simulation of Figure 3-8.

In general, bode plots such as the AOL, CMRR, and PSRR curves apply to sinusoidal signals. Often, noise signals applied to the power supply are not sinusoidal. To understand how the rejection curves function on non-sinusoidal curves it is helpful to use the Fourier theorem. This theorem states that any non-sinusoidal signal can be built with a series of sinusoidal signals. For example, a square wave at 100 kHz is composed of a sine wav at 100 kHz, 300 kHz, 500 kHz and other odd multiples of 100 kHz. The various multiples of the square wave frequency are called harmonics. All of the sinusoidal components can be applied to the bode plot to see how each component is affected. The SPICE simulation does this automatically, but it is useful to think of the math behind the simulator because the response to the non-sinusoidal signals may not be intuitive otherwise.

OPA206 OPA210 with Noise Signal on
                    Positive Power Supply Figure 3-8 OPA210 with Noise Signal on Positive Power Supply
OPA206 VOS vs Frequency
                    for Noise Signal on Power Supply Figure 3-9 VOS vs Frequency for Noise Signal on Power Supply