Hi Hans,
Yes you are correct. For a linear regulator, when we discuss noise we are referring to the noise that is generated by the regulator itself. This noise primarily comes from the reference of the regulator. Power supply rejection ratio (PSRR), also known as ripple rejection, is a measure of the LDO’s ability to filter ripple on the input rail. Noise sensitive applications require a “clean” output rail. As such they require low LDO noise and high PSRR.
As such in order to measure noise, you are exactly right that we would want to use the cleanest input supply possible so that we are only measuring the self generated noise. While a battery can sometimes be used as the supply, this is not always practical for the application conditions that you want to test. As such, we will often use a bench supply and then filter it before the Device Under Test (DUT) which would commonly be on an EVM as shown in the block diagram below. This filtering can be done using passive components or even using the PSRR of an additional LDO. You will want to be sure to measure Vin local to your DUT to account for any voltage drop in your filtering method.
While the above block diagram is from the white paper provided in my previous post, the same exact process would be used for a oscilloscope measurement.
Very Respectfully,
Ryan
