How to Test Noisy Power Supply Outputs

Figure 1: 3.3 V output of a DC-DC converter.
The waveform shows the nominal DC level,
ripple and high frequency noise bursts.

Did you ever acquire the output of a power supply with your oscilloscope and find an unexpectedly high level of noise? Did you try adding filter capacitors only to find the noise level was not changed? 

In this post, we'll discuss how the choice of probe affects the noise present in power measurements, as well as how oscilloscope settings such as termination impedance, bandwidth and coupling can be adjusted to lessen noise and improve measurement results.

Figure 1 shows a typical DC-DC converter output measurement. The mean value of the waveform is 3.294 V.  Ripple appears at the switching frequency of 1.2 MHz, and noise in the form of high frequency bursts and baseline thickening is visible throughout.

Waveforms like this can be acquired with a 10:1 high impedance probe, a 1:1 coaxial cable connection, or a 1.2:1 rail probe using either DC or AC coupling, as available.  Figure 2 summarizes how each oscilloscope/probe configuration affects the measurement.

Figure 2: How oscilloscope/probe configurations affect various measurement qualities.

The 10 MOhm passive probe is good for measuring low frequency phenomena under 20 MHz.  It would also do to characterize the ripple at 1.2 MHz.  However, if a long ground lead is used, it will add noise to the measurement due to inductive pickup, so it is recommended to use a short ground spring clip when using a passive probe.  Also note that these probes attenuate the input signal by a factor of 10:1 (20 dB), but measurement related noise is not similarly attenuated, so the result is that the probe reduces the signal-to-noise ratio (SNR) by 20 dB.  

A coaxial connection working into the oscilloscope’s 1 MOhm termination produces a unity gain (1:1) connection. This provides good sensitivity but limits the offset range to what is available in the oscilloscope.  Generally, high sensitivity ranges have limited DC offset range.  So, if the offset range is limited, it may not be possible to DC couple the input and keep the trace onscreen. 

AC coupling can be used to keep the trace on screen, but then the DC measurement is lost.  Measuring both power supply output level and ripple/noise may require two acquisitions.  Another issue is that if high frequency signals or fast edges are involved, the mis-terminated cable may cause reflections that will result in measurement errors. 

Reflections can be eliminated by adding a series resistor with a resistance equal to the characteristic impedance of the cable in series with source. It is also possible to terminate the cable using the oscilloscope’s 50 Ohm input. This prevents reflections and provides a measurement bandwidth equal to the oscilloscope’s bandwidth. However, 50 Ohm termination is limited to a maximum input voltage of 5 V and results in a high load on the measured circuit.  Consider the 3.3 V signal we showed earlier. The 50 Ohm termination would draw 66 mA from that circuit.

Figure 3:  3.3V DC-DC converter output acquired
using a 2 GHz rail probe bandwidth limited
to 200 MHz and DC coupled.  

The rail probe has a maximum bandwidth of 4 GHz. It splits the measurement path into an AC path and a DC path.  The DC path has an input impedance of 50 kOhms, while the AC path maintains a 50 Ohm termination. The two paths are combined to achieve a probe gain of 1.2:1 with a ±30 V offset range. 

Figure 3 shows the DC-DC converter output measured using a rail probe with the bandwidth limited to 200 MHz. The parameters show that the DC-DC converter output has a nominal or mean output level of 3.295 V (P1) and a peak-to-peak ripple plus noise level of 31 mV (P2). The ratio of those two values (P3) is 0.095 or 9.5%.

Figure 4: Reducing rail probe bandwidth to
20 MHz eliminates the high frequency noise.

Reducing the measurement bandwidth to 20 MHz suppresses the high frequency noise, which helps to measure only the 1.2 MHz ripple (Figure 4). The mean value of the DC output remains a 3.295 V, but the noise bursts have been eliminated and the peak-to-peak amplitude is reduced to 13 mV.

In short: DC coupling retains the power supply’s mean or nominal output level, while controlling the probe bandwidth provides a technique for separating ripple from high frequency noise.

Watch Steve Murphy explain this in the on-demand webinar, "Testing Noisy Power Supply Outputs."

See also:

Five Tips to Improve Dynamic Range

How to Use Memory Properly

How to Use Measurement Statistics to Set Up Triggers

How to "Layer" Measurement Tools