Measuring Power Supply Noise Correctly
When using an oscilloscope to measure the noise output of a power supply, one must utilise the correct return probe attachment. Usually oscilloscope probes come with the crocodile clip return attachment already fitted, which is handy and flexible for most measurements, however for noise measurements it results in an unfavourable or misleading result.
The crocodile return attachment of the probe creates an inductive hoop that induces current from RF (like an aerial) which is then superimposed on the voltage signal that is seen on the oscilloscope.
Below is the oscilloscope image of a power supply when using the crocodile clip attachment
Below is the oscilloscope image of the same power supply under the same load, but using the spring return probe attachment
We can see some subtle differences in the two graphs, which would mislead you in to thinking that the power supply was noisy.
In addition to this it is always important to see how the noise specification is defined on the product datasheet. Often the specification detail states that the oscilloscope must have a 20 MHz bandwidth filter on the probe and is measured with some capacitor across the lines. This may be in the small print notes. On all FiDUS POWER datasheets we state if a bandwidth filter or capacitor(s) are used in defining the specification.