Now that we've summarized the many noise doorways of an ADC (see, for example, ADC Noise: A Second Look, Part 2), let's start focusing in on each one. First, we'll take a look at noise that enters the ADC through the power supply.
This can sometimes be one of the most frustrating to deal with, next to another doorway, the ground plane… but we'll get to that doorway later. There are a couple of terms to become familiar with when looking at power supply noise. These are PSRR-DC, PSRR-AC, and PSMR. PSRR is the power supply rejection ratio (we'll explore both DC and AC aspects). PSMR is the power supply modulation ratio.
In order to understand the power supply noise doorway, we need to understand these terms and what they mean to the ADC. Basically, these terms are telling us how far open the door is through the power supply. The less the rejection, the more open the door is for noise to enter the ADC through the power supply inputs.
First, we'll examine the PSRR-DC and PSRR-AC. Figure 1 below conceptually shows each. The PSRR-DC of the ADC is the ratio of change in the gain or offset error of the ADC to a change in the power supply voltage.
This is a static change and not a time-varying change as in Figure 2, which shows the PSRR-AC. The PSRR-AC is more widely used, as typically the supply voltages of the ADC are well regulated, and the PSRR-DC does not have a large impact on overall performance.
The PSRR-AC gives a much more insight into the ADC performance because high-frequency noise signals are typically much more difficult to filter from supply lines. These signals can couple onto the power supply planes and lines from many different mechanisms. Wherever they couple onto the ADC supply lines, the PSRR-AC will tell us how much the converter rejects these signals. The PSRR-AC gives the designer an idea of how much power supply noise will affect the input signal or add to it.
This can be measured as shown in Figure 3. For PSRR-AC a signal is injected onto the power supply input (AVDD shown, but could also be DVDD) while no input signal is present at the analog inputs of the ADC. The result will be a spur in the FFT at the output of the ADC that is at the frequency of the injected signal and is at some reduced amplitude with respect to the amplitude of the injected signal. The ratio of the spur to the input amplitude of the injected signal gives the PSRR-AC value.
In a manner similar to the PSRR-AC, the PSMR is measured with a signal injected onto the power supply input. However, for PSMR, an input signal is also applied to the analog inputs of the ADC. See Figure 4. In this case, the injected signal will mix with the analog input signal and result in a spur in the FFT at a product of the two signals (fINPUT ± fINJECTED ). The PSMR is again the ratio of the amplitude of the injected signal to the amplitude of the resultant spur in the FFT. In this case though, since there are two spurs in the FFT, typically the larger of the two spurs is used in the calculation.
There is so much that we could discuss, but unfortunately there is only so much space in a blog. For more information, I would recommend looking at Application note MS-2210.
App-note MS-2210 takes a little deeper dive into the specifics of PSRR and PSMR as well as more on how to measure each of these parameters. Please keep the great comments coming as we continue to look at the various noise doorways of an ADC.
- ADC Noise: A Second Look, Part 2
- ADC Noise: A Second Look, Part 1
- (Under)standing Your Ground
- The Ground Myth
- ADC Noise: Where Does It Come From?
- Interleaving Spurs: The Mathmatics of Timing Mismatch
- Interleaving Spurs: More Math Details for Gain Mismatch
- Interleaving Spurs: Let’s Look at the Math
- Interleaving Spurs: Bandwidth Mismatches
- Interleaving Spurs: Timing Mismatches
- Interleaving Spurs: Gain Mismatches
- Interleaving Spurs: Offset Mismatches
- More Thoughts on Interleaved ADCs
- Interleaved ADCs: The Basics
- LVDS Is Dead? Long Live LVDS & JESD204B
- Signal Chain Basics #82: Crosstalk Between Outputs in Clocking Devices