Achieving the lowest noise in a signal chain is vital as industry trends push the boundaries of resolution and precision. And when pushing these boundaries, it’s important to consider not just the noise of signal-chain components such as analog-to-digital converters (ADCs) and amplifiers, but also power products such as switching and low-dropout regulators (LDOs). Advances in silicon technologies have reduced the trade-offs when attempting to achieve low noise and high precision in power topologies.
Recent trends in 24-bit delta-sigma ADCs have increased sampling speeds and lowered power consumption. New low-noise power supplies and low-noise voltage references can take advantage of these trends and help ADCs achieve high-resolution measurements in low-power applications.
To achieve the lowest noise, let’s review the sources of noise in the signal chain and power architecture. Figure 1 shows a typical signal-chain application centered around an ADC that requires an external voltage reference, clock and signal-conditioning circuit. Every component in Figure 1 contributes to system noise and requires optimization.
Marcoo Zamora
System Engineer
Linear Power
1 Defining noise and precision in a power architecture | Noise is often application-specific, but in the context of this paper, noise
is any unwanted signal that originates from thermal noise, 1/f noise and
low-frequency oscillations, up to approximately 100 kHz. |
2 Innovations in low-noise and low-power voltage references | Reducing noise in power architectures helps increase an analog-to-digital
converter's resolution and precision but creates design challenges with
power consumption, printed circuit board (PCB) size, manufacturing flow and
cost. |
3 Innovations in precision battery monitoring | Having creative solutions in silicon technologies enables designers to
optimize their power architectures and battery systems. |
Noise in an ADC can cause errors in precise voltage measurements. You must consider the total contribution of noise in the signal chain from internal and external sources. Total noise is often a combination of ADC thermal noise, ADC-quantization noise, amplifier noise, voltage-reference noise and power-supply noise.
Equation 1 depicts the total referred noise at the input of the ADC (at full-scale voltage) as it measures the sensor based on Figure 1. The main design challenge is to optimize all noise sources to achieve the noise target that the application requires. In Equation 1, the ADC’s power-supply rejection ratio (PSRR) reduces the power-supply noise, which is plotted out to 1 MHz:
Given the existence of uncorrelated noise sources, the total noise is the root sum square of all sources, which heavily favors the largest noise source. One noisy component can heavily skew the measurement. For example, if a voltage reference contributes more noise than an ADC and power supply, reducing noise on the voltage reference will be the best way to lower the system noise, as shown in Figure 2 and Figure 3. In addition, ADC noise types vary with resolution: quantization noise is significant for a 16-bit ADC, but you can ignore it for a 24-bit ADC.