Introduction

A PCB may appear to have a stable power supply when measured with a multimeter.

For example, you may measure:

VCC = 3.30 V

But this does not necessarily mean that the power rail is clean.

An oscilloscope may reveal:

• • Ripple

• • High-frequency noise

• • Voltage spikes

• • Ringing

• • Startup disturbances

• • Voltage drop during high-current operation

• • Switching noise

These power disturbances can cause problems such as:

• • MCU resets

• • Communication errors

• • ADC errors

• • Flash programming failures

• • Unstable digital signals

• • Unexpected system shutdowns

This is why power integrity is an important part of PCB design and troubleshooting.

A power distribution network should maintain a stable voltage over the relevant frequency range of the system’s switching currents.

1. What Is PCB Power Integrity?

Power integrity (PI) describes how effectively a PCB delivers clean and stable power to its electronic components.

A simplified system looks like this:

The objective is not simply to have the correct DC voltage.

For example:

3.3 V supply ≠ necessarily clean 3.3 V supply

You also need to consider:

• • DC voltage

• • Ripple

• • Noise

• • Transient response

• • Voltage drop

• • Ground voltage variation

• • Current demand

2. VCC Ripple

Ripple is a periodic or semi-periodic voltage variation riding on top of the DC voltage.

For example:

The DC voltage may still be approximately 3.3 V, but an oscilloscope could show a ripple waveform.

Ripple is commonly associated with:

• • Switching power supplies

• • DC/DC converters

• • Rectifier circuits

• • Load changes

• • Inadequate filtering

3. Ripple Measurement

Suppose the power rail is:  3.3 V DC

You connect an oscilloscope and observe:  3.3 V ± 25 mV  

The peak-to-peak ripple is approximately: 

           VRIPPLE(P−P)=50mV

For example:

Therefore: 

             VPP = 3.325-3.275VPP = 50mV

The important point is that the specification should come from the IC, regulator, or system requirements. There is no universal “acceptable ripple” value for every PCB.

4. Why a Multimeter Is Not Enough

A multimeter is excellent for measuring DC voltage.

For example:  VCC = 3.301 V

But it may not show a short 50 ns or 500 ns voltage spike.

An oscilloscope can show:

This is particularly important for digital systems.

A short power disturbance can occur when an MCU, processor, memory device, or communication circuit suddenly changes its current demand.

5. How to Measure VCC With an Oscilloscope

A basic setup is:

The probe tip connects to VCC.

The ground connection connects to the PCB ground.

However, the way you connect the probe matters greatly.

Keysight specifically notes that shorter ground connections produce more accurate ripple measurements.

6. Avoid the Long Ground Lead

This is one of the most common oscilloscope measurement mistakes.

Poor measurement

VCC ─────── Probe Tip

GND ─────── long wire ───────── Probe Ground

The long ground lead can act like an antenna and add unwanted inductance.

You may see:

and think:

“My PCB has terrible noise.”

But some of what you see may actually be measurement-induced noise.

Keysight recommends keeping probe and ground connections as short as possible to reduce noise and signal distortion.

7. Use a Ground Spring When Possible

For high-frequency ripple measurements, a short ground spring is often preferable to a long alligator ground lead.

The goal is to minimize the measurement loop.

This becomes increasingly important when measuring fast switching edges and high-frequency noise.

8. Oscilloscope Bandwidth Matters

Suppose your oscilloscope is set to:  20 MHz bandwidth limit

You may not see high-frequency components above that bandwidth.

With bandwidth limit OFF, you may observe:

Therefore, the measurement bandwidth should be selected intentionally.

For a general troubleshooting measurement, you can start with a reasonable bandwidth limit and then increase bandwidth if you need to investigate high-frequency spikes.

Modern power-rail measurements can involve very small signals riding on DC rails, so oscilloscope noise, probe noise, bandwidth, and probing technique all affect the result.

9. AC Coupling for Ripple Measurement

Suppose your VCC is:  5.0 V DC

but ripple is only:  20 mV

Displaying the full 5 V signal can make the ripple difficult to see.

AC coupling removes the DC component from the displayed waveform:

DC + Ripple:

   5.02 V ─────────
   5.00 V ─────────
   4.98 V ─────────

With AC coupling:

   +20 mV ──────
   ~~~~~~~
   0 mV ──────
   ~~~~~~~
   -20 mV ──────

This allows you to zoom in on the ripple.

Keysight describes AC coupling as a method for eliminating the DC component when focusing on output ripple measurements.

Important: AC coupling is useful for viewing ripple, but don’t use it as your only measurement. You also need to check the actual DC level and low-frequency behavior.

10. Measuring Power Supply Noise

Noise is different from a clean periodic ripple waveform.

For example:

    

Noise may originate from:

• • Switching regulators

• • Digital ICs

• • Clock signals

• • DC/DC converters

• • Communication circuits

• • Motors

• • Relays

• • External EMI

• • Poor grounding

• • PCB coupling

A good troubleshooting approach is to compare the power rail under different operating conditions.

Test 1

System idle:

   VCC noise = 8 mV p-p

Test 2

MCU running:

   VCC noise = 15 mV p-p

Test 3

Communication active:

   VCC noise = 35 mV p-p

This suggests that the noise may be related to the changing system load or communication activity.

11. Measuring Voltage Drop

Voltage drop is another important power-integrity problem.

Suppose a regulator provides:  5.0 V

but the MCU receives:  4.7 V

under high load.

The voltage drop is: 

       VDROP = 5.0−4.7V_DROP  =  0.3V

or: 

       VDROP = 300mVV_DROP = 300mV

Possible causes include:

• • PCB trace resistance

• • Connector resistance

• • Cable resistance

• • Via resistance

• • Insufficient power supply capacity

• • High load current

• • Poor grounding

• • Battery internal resistance

12. Basic Voltage-Drop Calculation

The basic equation is: 

              VDROP = I × R

Suppose: 

              I =  2A

and the total resistance is: 

              R = 0.05 

Then: 

             VDROP = 0.1V

So a 5 V supply could become approximately: 

              5.0−0.1 = 4.9V

at the load.

13. PCB Trace Resistance

A PCB trace has resistance.

The basic relationship is:

             

where:

       

Therefore: Longer trace → higher resistance

and: Wider/thicker trace → lower resistance

This is one reason power traces often need to be wider than ordinary signal traces.

14. Voltage Drop Can Cause Digital Problems

Consider:

     

If the MCU’s required operating voltage is close to the actual supply voltage, a load transient can potentially cause a reset or malfunction.

For example:

    

The system may work normally most of the time but fail only during certain operations.

That makes this type of problem particularly difficult to troubleshoot.

15. Measure VCC at Two Locations

This is a very useful practical test.

Measure: 

Point A — Power supply output

    

Then measure the voltage at the MCU.

For example:

     Point A = 5.00 V

     Point B = 4.82 V

Therefore: 

     VDROP = 5.00-4.82VDROP = 0.18V

You have now identified a significant voltage difference along the power path.

16. Measure During Maximum Load

This is extremely important.

A PCB may show:   5.00 V at idle

but:  4.70 V during peak load

Therefore, don’t test only when the system is idle.

Test during:

• • MCU startup

• • Flash programming

• • Motor startup

• • Communication transmission

• • Relay operation

• • Maximum CPU activity

• • Vision-system operation

• • Maximum system load

The worst-case condition is often more informative than the idle measurement.

17. A Practical Test Example

Imagine a 24 V industrial controller.

At idle:  5.02 V

During communication:   4.95 V

During maximum load:  4.78 V

Now calculate: VDROP = 5.02-4.78VDROP = 0.24V

That’s a 240 mV change.

The next question is:    Where is the voltage being lost?

Measure progressively:

This allows you to locate the problem.

18. Check the Decoupling Capacitors

If the power rail has excessive high-frequency noise, check:

• • Capacitor value

• • Capacitor placement

• • ESR

• • Ground connection

• • VCC trace length

• • Ground-plane connection

• • Via placement

A capacitor can be electrically correct but physically poorly placed.

TI documentation gives examples where very close placement of a small decoupling capacitor is specified to reduce noise and ripple, with a larger bulk capacitor located nearby.

19. Check the Ground

Sometimes the problem isn’t actually the VCC trace.

It may be the ground path.

Consider:

      

If significant current flows through a shared ground path:

        V = IR

the ground potential can change.

This can create what engineers commonly call ground bounce or ground voltage variation.

Therefore, measure:   

          VCC relative to the actual local MCU ground.

Don’t assume that every ground point on a PCB is at exactly the same voltage under dynamic conditions.

20. Compare Normal and Abnormal Boards

If you have a working PCB and a failing PCB, this can be one of the fastest troubleshooting methods.

Measure the same points:

Test Point	Good Board	Problem Board
DC/DC output	5.01 V	5.01 V
MCU VCC	4.98 V	4.72 V
Ripple	15 mV p-p	85 mV p-p
Reset	Stable	Disturbed

This can quickly narrow the problem to the power-distribution or reset circuitry.

21. Power Integrity and Your Real-World Controller Experience

This is particularly relevant to your previous controller-development experience.

You experienced an intermittent startup problem where an unstable power-reset signal could prevent the controller from starting.

A useful troubleshooting sequence for that type of problem is:

     

Instead of immediately assuming:

“The MCU is defective.”

measure the power and reset signals first.

This is a practical engineering lesson:

A digital system cannot be more stable than the power and reset conditions that support it.

22. Troubleshooting Flow

When you encounter an intermittent PCB failure, try this sequence:

Step 1 — Measure DC voltage

Use a multimeter.

    VCC = ?

Step 2 — Check voltage under load

    Idle → Load → Maximum Load

Step 3 — Measure ripple

Use an oscilloscope.

    VCC → AC Coupling → Vpp

Step 4 — Check high-frequency noise

    Reduce the time scale and examine the waveform.

Step 5 — Check startup

    Trigger the oscilloscope during power-up.

Step 6 — Check reset

    Compare:

    VCC vs RESET

Step 7 — Check ground

    Measure local ground differences if appropriate.

Step 8 — Check decoupling

    Inspect capacitor value, placement, routing and grounding.

Step 9 — Check PCB power path

    Look at:

• • Trace width

• • Trace length

• • Vias

• • Connectors

• • Ground paths

Step 10 — Compare with a known-good board

    This can dramatically reduce troubleshooting time.

23. What Should You Look for on the Oscilloscope?

When analyzing VCC, don’t look only for ripple amplitude.

Look for:

     Ripple

     

     Spike

     

     Ringing

     

    Voltage sag

     

     Startup instability

     

Each waveform can suggest a different type of problem.

24. Measurement Checklist

Before accepting your PCB power design:

• • Measure DC VCC

• • Measure VCC at the load

• • Measure VCC under maximum load

• • Measure ripple

• • Check high-frequency noise

• • Check startup waveform

• • Check voltage drop

• • Check reset signal

• • Check ground

• • Check decoupling capacitors

• • Check regulator input/output capacitors

• • Check power trace width

• • Check connectors and vias

• • Compare good vs. failed boards

• • Verify with an oscilloscope

Conclusion

PCB power integrity is not simply about measuring whether the power supply is 3.3 V, 5 V, or 12 V.

A reliable PCB requires you to understand what happens to that voltage during real operation.

The three fundamental measurements are:

1. DC Voltage

Is the voltage correct?

2. Ripple and Noise

Is the voltage clean?

3. Voltage Drop

Does the voltage remain stable when current increases?

A simple measurement strategy is:

     

For accurate low-level ripple measurements, the measurement setup itself must not add significant noise or loading. Probe selection, short ground connections, bandwidth, oscilloscope noise floor, and measurement technique can all affect the result.

PowerTechGuide takeaway:

  Measure the power where the IC actually operates—not just where the power supply generates it.