Using an Oscilloscope for Intermittent Faults — Catching Problems That Come and Go
Written by Anthony Calhoun, ASE Master Tech A1-A8
Intermittent faults are the ones that make techs question their sanity. The vehicle stalls in traffic, then starts right back up. The check engine light flashes for two seconds and disappears. The customer brings it in, you drive it for thirty minutes, and it never does a thing. You send it home clean, and the next morning the customer is back on the phone. Every tech has been there. The problem is not that you missed something on the inspection. The problem is that the tool you were using cannot see the fault when it is not happening — and even when it is happening, a fast enough glitch is invisible to that tool.
The oscilloscope changes that completely. This article covers how to use a scope specifically for intermittent fault diagnosis — trigger strategies, sensor monitoring, power and ground testing, CAN bus work, and long-term capture. These are real techniques used by working technicians every day. If you have a scope sitting in the tool cart because it feels complicated, by the end of this article you will have a clear path to putting it to work.
Why the Scope Catches What the Meter Misses
A digital multimeter is an averaging tool. When you put your meter on a voltage signal, it samples that signal and averages what it sees over a short window of time. That is fine for static DC voltage checks and resistance measurements. It fails completely when the fault is a fast transient event.
Consider this: a crankshaft position sensor signal dropout that lasts 2 milliseconds. Two milliseconds is 0.002 seconds. The ECM sees that dropout and the engine stumbles or stalls because it lost track of crank position. Your meter never shows it. The meter is updating its display several times per second, and it is averaging the signal, so that 2ms glitch is completely buried in the noise floor of the measurement. You can sit there with your meter connected all day and the reading will never move off 12 volts even while the fault is actively occurring.
The oscilloscope displays voltage over time. Every sample is plotted on a timeline, so you see the signal as a waveform moving across the screen. A 2ms dropout shows up as a clear notch in the waveform — unmistakable, time-stamped, and measurable. You can zoom in, measure the exact duration, measure the voltage drop, and have hard data to work from.
This is the fundamental advantage: the scope shows you the shape of the signal across time. Intermittent dropouts, voltage spikes, noise corruption, missing pulses, and amplitude changes are all visible on a scope because none of them are averaged away. What the meter hides, the scope reveals.
Scope Basics for Intermittent Work
Before getting into trigger strategies, you need to be solid on the basic controls that matter most for intermittent diagnosis.
Time Base
The time base controls how much time is displayed across the screen. On most scopes, this is expressed as time per division — for example, 5ms/div means each horizontal grid square represents 5 milliseconds. If your screen has 10 divisions, you are viewing 50ms of signal at once. For CKP and CMP signals on a running engine, you might use 10ms/div to see several pulses at once. For a CAN bus signal, you might drop to 20 microseconds per division to see individual data bits. For long-term power and ground monitoring while wiggling a harness, you might go to 200ms/div or even seconds per division to see a longer time window. Set the time base based on the speed of the signal you are watching and the speed of the fault you expect to catch.
Trigger Settings
The trigger is what tells the scope when to start capturing and drawing the waveform on screen. Without a trigger, the waveform scrolls continuously and you cannot read anything. With a trigger set correctly, the scope waits for a specific event, then captures and holds the screen so you can analyze what happened. For intermittent work, the trigger is the most critical setting you will make. It is the difference between catching the fault and missing it.
Single-Shot vs Auto Trigger
Auto trigger mode continuously updates the screen — it keeps refreshing as long as a trigger event occurs. This is useful when the signal is actively present. Single-shot mode arms the scope once, waits for a trigger event, captures it, and then holds the screen frozen. This is what you use for intermittent diagnosis. You arm the scope, you wait for the fault, and when it happens the scope freezes the capture for you to examine. The event is preserved on screen even if it only happened once.
Roll Mode and Record Mode
Roll mode scrolls the waveform continuously from right to left like a paper strip chart. There is no trigger — it just records everything. This is useful for slow events like temperature changes or long-duration electrical issues. Record mode (available on higher-end scopes like Pico) allows extended captures that can be played back later. You can record hours of data and then scrub through the timeline to find the fault event. This is a game-changer for intermittents that only happen after extended drive cycles.
Sample Rate
Sample rate is how many data points per second the scope captures. Higher sample rate means more detail, especially for fast signals. A CAN bus signal runs at 500 kilobits per second on most modern vehicles — you need a scope with a fast enough sample rate to resolve the individual bits. Most professional automotive scopes (Pico, ATS, Snap-on Zeus) have more than enough sample rate for any automotive signal. The key point: do not use a cheap scope with a low sample rate for intermittent work. You may miss the exact event you are trying to catch because the scope was not sampling fast enough to see it.
Trigger Setup for Intermittent Capture
Setting the trigger correctly is the skill that separates techs who catch intermittents on the scope from techs who use the scope for ten minutes, give up, and go back to the meter. The strategy is always the same: set the trigger to catch the abnormal event, not the normal signal.
Edge Trigger
Edge trigger fires when the signal crosses a voltage threshold, either rising (low to high) or falling (high to low). For a normal sensor signal, you set the threshold at the normal voltage level and use auto trigger to confirm the signal looks right. Then, to catch a dropout, you flip to a falling edge trigger with the threshold set at a voltage level the signal should never normally reach — for example, if a 5-volt reference signal should never drop below 4 volts, set the trigger threshold at 3.5 volts, falling edge. Now the scope arms and waits. The moment the signal drops below 3.5 volts, the scope fires and captures the event. You were not looking. You did not have to be. The scope caught it automatically.
Pulse Width Trigger
Pulse width trigger captures when the scope detects a pulse that is shorter or longer than a programmed time limit. This is extremely useful for CKP sensor monitoring. The CKP signal produces pulses at a consistent interval based on engine speed. If one pulse is missing or dramatically shorter than the rest, the pulse width trigger fires. You set it up to capture any pulse shorter than, say, 80 percent of the expected pulse width at idle, and the scope will capture the exact moment a tooth is missing or a dropout occurs. This is a more sophisticated trigger mode but it is available on most professional scopes and it is worth learning.
The Triggering Mindset
The key principle is this: you are not trying to capture the normal signal. You are setting a trap for the abnormal event. Think about what the signal looks like when the fault is present, and set your trigger to fire on that condition. If you are hunting a dropout, trigger on the dropout voltage. If you are hunting a spike, trigger on the spike voltage. If you are hunting a missing pulse, use pulse width trigger. This mindset makes intermittent scope work systematic instead of lucky.
Monitoring Sensor Signals
Crankshaft and Camshaft Position Sensors
CKP and CMP signals are responsible for some of the most frustrating intermittents in the field — random stalls, random misfires, hard starts that happen once a month. Put the scope on the CKP signal wire with the engine running. On a Hall effect CKP, you should see a clean digital square wave with consistent pulse width across the entire capture. The amplitude should be consistent — full reference voltage high and clean ground low. A dropout shows as either a missing pulse (the line stays low for an extra period) or a partial amplitude drop (the pulse does not reach full voltage). Both are obvious on a scope and both are invisible on a meter. Set a single-shot trigger using pulse width trigger mode or set a low-voltage edge trigger and let the scope arm and wait. Even a single 2ms dropout will be captured and frozen on screen for you to analyze and photograph.
Throttle Position Sensor
TPS intermittents are classic — hesitation, surging, erratic idle, or codes for TPS out of range that set once and never come back. Put the scope on the TPS signal wire and slowly sweep the throttle from closed to wide open by hand, taking at least 10 full seconds to move through the full range. A good TPS produces a perfectly smooth voltage ramp with no discontinuities. An intermittent dead spot or wiper wear shows up as a spike, a dropout, or a flat spot in the middle of the ramp. Set the time base wide enough to capture the full sweep — 2 to 5 seconds per division. You will see the fault clearly. On a meter, the needle might jitter for a fraction of a second and you might miss it entirely if you are not watching at exactly the right moment. On the scope, the waveform is a permanent record of everything that happened during the sweep.
MAP and MAF Sensors
MAP and MAF intermittents often produce noise and spikes on the signal wire that the ECM interprets as real airflow or pressure changes. This causes rich or lean conditions, surging, or random misfires. Put the scope on the MAP or MAF signal wire with the engine at idle. The signal should be a steady voltage with very little noise — a clean, relatively flat line. Internal sensor failure often shows up as random voltage spikes or bursts of noise superimposed on an otherwise normal signal. You might see this at idle, or it might only appear under certain load conditions. Use the scope to monitor the signal under the conditions that produce the symptom. The noise pattern will be visible, and you will know the sensor is failing internally before it ever sets a code.
Monitoring Power and Ground Circuits
Intermittent power and ground issues are some of the hardest faults to catch with a meter because the fault only appears during harness movement, vibration, or thermal stress. The scope is significantly more sensitive for this work.
For intermittent power supply monitoring: connect the scope channel to the power wire at the component. Set the time base to something wide — 100ms to 500ms per division — so you have a long view of the signal. Start wiggling the harness, connectors, and wiring at different points while watching the scope. A loose connection or damaged wire will produce a momentary voltage dip when the connection breaks. On a meter, you might see the voltage number flicker once and assume it is a meter artifact. On the scope, you see a clear voltage dip — exactly how deep, exactly how long, exactly how many times it happened during your wiggle test. This is hard evidence.
For ground circuit monitoring: connect the scope channel to the ground wire at the component with the other lead on a known good chassis ground. A good ground connection should show nearly zero volts — just a few millivolts at most. When you wiggle the harness and a poor ground connection opens momentarily, you will see a voltage spike on the scope. Even a brief 0.3 volt spike on a ground circuit is enough to cause problems for sensitive electronics, and it is clearly visible on the scope. On a meter, that 0.3 volt spike lasts a few milliseconds and the meter never moves off zero. The scope does not miss it.
This technique is particularly effective for headlight dimming complaints, radio reset issues, module communication faults, and any symptom that is caused by vibration or connector flex. Combine scope monitoring with a systematic harness wiggle from the component back to the battery, and you will find the fault location.
CAN Bus Monitoring
Modern vehicles communicate between modules on a CAN bus — a two-wire differential signal called CAN-H and CAN-L. When that communication is intermittently corrupted, you get U-codes, modules going offline, instrument cluster flickering, random transmission shifts, and other symptoms that seem completely unrelated but share a common cause. The scan tool often cannot catch intermittent CAN bus corruption because the corruption clears before the scan tool can record it. The scope can.
Connect one scope channel to CAN-H and a second channel to CAN-L. At idle with the bus active, you should see a clean differential signal — CAN-H swinging above 2.5 volts and CAN-L swinging below 2.5 volts during data transmission, with both lines resting at 2.5 volts during recessive states. The two waveforms should be mirror images of each other. Any corruption — noise, a module loading the bus abnormally, a short between the wires — will show as a disruption in this clean differential pattern.
When an aftermarket module or a failing OEM module intermittently loads the bus, you will see a burst of noise or a waveform collapse during the moment the module wakes up or transmits incorrectly. By monitoring at different splice points along the bus, you can narrow down which section of the wiring is affected and which module is the source of the corruption. This level of diagnosis is not possible with a scan tool alone — the scope is the only tool that gives you this visibility.
Long-Term Recording
Some intermittents are so rare that they cannot be caught during a typical shop drive cycle. The vehicle might sit for an hour and then stall once when returning from lunch. Or a fault only occurs after an extended highway drive. For these situations, long-term scope recording is the answer.
Higher-end scopes like the Pico 4425A can record signal data for extended periods — hours in some cases — and store it to a laptop hard drive. You set up the scope, connect your channels to the circuits you want to monitor, set the record mode, and send the vehicle out for a drive or leave it running overnight. When the vehicle comes back, you open the recording and scrub through the timeline. When you reach the point where the fault occurred — identified by the customer describing the time and conditions — you zoom in and see exactly what the signal was doing.
For this to work, you need the scope connected to the circuits most likely to be involved. Use your diagnostic reasoning first: what symptom is it, what circuits control that system, what signals need to be monitored to catch the root cause. Connect those channels, arm the record mode, and let the scope do the work. Long-term recording turns a multi-week intermittent into a solved case within one vehicle loan.
The Pico Scope Advantage
The Pico scope has become the industry standard for advanced automotive diagnostics, and there are specific reasons why it has pulled ahead of other options for intermittent fault work.
The PicoDiagnostics software includes guided tests for common automotive applications — compression testing, relative compression, ignition waveforms, injector patterns, and more — with on-screen guidance that walks you through the test. For techs who are still building scope skills, this removes the barrier of figuring out the setup from scratch every time.
Math channels allow you to perform calculations on live signals — adding two channels together, subtracting one from another, or calculating the difference between CAN-H and CAN-L to display a true differential signal. This is a powerful capability for CAN bus analysis and for current ramp testing on actuators.
The automatic analysis features can measure frequency, duty cycle, pulse width, amplitude, and timing relationships automatically and display the numbers on screen without manual measurement. This speeds up analysis significantly, especially during long recording playback when you are scrubbing through data.
Pico also maintains a free waveform library with known-good captures from a wide variety of vehicles and sensors. This is invaluable for intermittent diagnosis — before you can identify an abnormal signal, you need to know what normal looks like. The Pico waveform library gives you a reference point for signals you have never captured before.
Practical Examples From Real Repairs
Intermittent CKP Signal Causing Random Stall
The complaint was a random stall with no warning, no codes, and no pattern. The vehicle would restart immediately. Classic CKP dropout behavior. The tech connected a Pico channel to the CKP signal wire and set a single-shot trigger using falling edge at 1 volt — well below the normal 5-volt digital signal. With the scope armed and the vehicle at idle, the tech waited. After approximately eight minutes, the stall occurred. The scope captured the event: a clean 2-millisecond dropout on the CKP signal, the voltage collapsing to near zero and recovering. Duration: 2ms. Depth: complete signal loss. The meter connected to the same wire at the same time showed no change. The scope captured the exact fault. The CKP sensor was replaced, the stall did not return, and the repair was documented with a saved scope capture.
Intermittent Ground Causing Dim Headlights
The complaint was headlights that occasionally dimmed briefly, especially when going over bumps. No codes, no obvious issues on visual inspection. The tech connected a scope channel to the headlight ground wire at the headlight assembly with the reference lead on a known chassis ground point. Time base set at 200ms per division. With the headlights on, the tech flexed the harness at the engine-to-body ground strap connection. The scope captured a 0.3-volt spike on the ground circuit — brief, sharp, and clearly associated with harness flex at that location. The ground strap was corroded at the body attachment point. Cleaning and tightening the connection eliminated the spike on the scope and the customer complaint. Total diagnosis time: under twenty minutes with the scope. Without it, this fault could have taken hours of chasing.
Intermittent CAN Bus Corruption from Aftermarket Module
The complaint was random U-codes, occasional gauge cluster freezes, and a transmission that occasionally dropped into a failsafe mode with no warning. The vehicle had an aftermarket remote start module installed by a previous owner. The tech connected two scope channels to CAN-H and CAN-L at the OBD-II port. With the scope in roll mode, the tech observed the CAN bus signal during normal operation — clean, consistent differential waveform. Then the tech used the remote start key fob to activate the module. At the moment the aftermarket module transmitted, the scope showed a burst of noise on both CAN lines — approximately 400 microseconds of signal corruption. This was enough to cause the receiving modules to reject a data frame and log a communication fault. The aftermarket module was removed, the CAN bus returned to a clean signal, and all symptoms resolved. The scope capture was shown to the customer as documentation for why the module had to come out.
Building Scope Skills
The single most effective habit for building scope diagnostic skills is capturing known-good signals on every vehicle you work on, even when the vehicle is in for routine maintenance. Pull up the CKP signal on a vehicle you just changed oil on and the engine is running perfectly. Save that capture. Pull up the TPS sweep on a vehicle with no throttle complaints. Save that capture. Do the same for MAP, MAF, CAN bus, and oxygen sensor signals. Build a personal library of what normal looks like across different makes, models, and engine configurations.
When a fault is eventually present, you have a reference to compare against. The difference between normal and abnormal becomes obvious when you have a side-by-side comparison. Without that reference, you are trying to judge whether a signal looks right from memory — and memory is unreliable, especially for signals you do not see every day.
Start with easy circuits when you are learning. A battery voltage signal or a simple on-off relay circuit is a good starting point for understanding how the trigger works and how to read the time base. Move to sensor signals once you are comfortable with the basic controls. Save every useful capture and organize them by signal type and vehicle. Over six months of deliberate practice, you will build a mental library of waveform patterns that makes fault identification fast and confident.
Two free resources that accelerate scope learning significantly: the Pico waveform library at the Pico Technology website, which has hundreds of documented automotive waveforms with explanations, and the ScannerDanner YouTube channel, which has extensive scope demonstration videos on real vehicles with real faults. Both are worth studying systematically, not just watching casually.
The oscilloscope is not a replacement for strong diagnostic reasoning — it is the tool that executes the diagnosis once you have identified what circuit to test and what fault to catch. Combine solid systems knowledge with proper trigger setup and you have the most powerful intermittent diagnostic capability available in any shop. The faults that used to leave vehicles sitting for weeks waiting to act up are now catchable in a single session. That is the difference the scope makes.