Practical Technique
Intermittent Fault Capture
How to catch intermittent automotive faults with a lab scope: glitch and pulse-width triggers, roll mode, segmented memory, long recordings and unattended capture strategies for problems that only happen once a week.
The intermittent problem
Intermittents are the hardest, most profitable repairs in driveability work. Scan-tool freeze frames help, but they only tell you what happened after the fault. A scope, set up the right way, captures the fault itself — the dropout, the noise spike, the missing pulse — so you stop guessing.
Step 1 — Pick the right trigger weapon
Glitch / pulse-width trigger
The single most powerful tool for intermittents. Tell the scope "trigger when the high pulse is shorter than 500 µs" or "trigger when the low pulse is longer than 50 ms" and it watches silently, ignoring every normal pulse. When the bad event happens, the screen freezes around it. Combine with single-shot modeso the scope stops and waits for you to come look.
Window / level trigger
Fires when the signal leaves a voltage window. Great for a 5 V reference that briefly drops to 4 V, or a 12 V feed that occasionally dips to 8 V — neither extreme would fire a normal edge trigger.
Trigger holdoff
If your signal has lots of legitimate edges (CKP, CAN), trigger holdoff tells the scope "after firing, ignore the next N milliseconds." That stops it from re-triggering inside the same event and missing the next one.
Step 2 — Choose the right capture mode
- Single-shot: arm once, capture one event, freeze. Pair with glitch trigger for unattended capture.
- Normal: continuously waits for a trigger, redraws each time. Good when the fault repeats every few minutes.
- Roll mode: samples scroll across the screen in real time, no trigger needed. Use for very slow drifts (O2 bias, parasitic draw, MAP at idle).
- Segmented memory: if your scope supports it, this captures dozens of short events back-to-back, ignoring the dead time between them. Brilliant for catching every glitch in a 10-minute road test.
Step 3 — Stretch the recording window
The longer the scope can record at useful resolution, the better your odds. Two ways to get there:
- Deep memory: a scope with 1 MS memory at 1 MS/s gives 1 second of capture; at 10 kS/s it gives 100 seconds. Match memory to the event speed.
- Datalogger mode: many automotive scopes have a separate logger that captures at 1–100 Hz for hours. Use it for slow signals — O2, fuel trim, battery voltage during a road test.
Step 4 — Provoke the fault
While the scope is armed, force the conditions:
- Heat: heat gun on suspect modules, connectors, sensors. Most "warm restart no-start" faults are heat-soaked PCMs or crank sensors.
- Cold: freeze spray on the same suspects when the failure is cold-only.
- Vibration: tap connectors and grounds with a screwdriver handle. A bad pin will trigger the scope on the very next tap.
- Load: turn on every accessory (headlights, blower, defrost, rear window) to load the charging system and shake out marginal grounds.
- Wiggle the harness — especially where it crosses pivots, exhaust heat, or the bellhousing.
Step 5 — Unattended capture in a customer car
- Confirm the trigger fires on a deliberate test glitch before handing the car back.
- Hard-wire scope power to a fused, always-hot 12 V source (cigarette lighter outlets often power down with ignition).
- Disable the scope's screensaver / sleep timer.
- Set single-shot, deep memory, widest practical capture window centered on the trigger point.
- Strap the scope down so leads can't shift. Tape a "do not touch" note over them.
- Ask the customer to call the moment the fault occurs and not restart the car — the captured trace is on screen until you collect it.
Reading the capture
- Pre-trigger samples (left of the trigger marker) show the conditions leading up to the fault — usually more useful than post-trigger.
- Compare the bad capture against a known-good reference of the same signal on the same vehicle.
- If the suspect signal looks clean but the fault occurred, the cause is upstream — capture power and ground at the same time.
- Save the full screen, including timebase, V/div and trigger settings. A trace without scale is unreadable a week later.
Frequently asked questions
What is glitch triggering?
A trigger mode that fires when a pulse is shorter (or longer) than a window you set. It's the only practical way to catch a one-in-a-thousand event — the scope ignores every normal pulse and only saves the bad one.
When should I use roll mode?
For very slow events (parasitic draw, fuel-trim drift, intermittent module wake-ups) where you'd rather watch a scrolling chart than wait for a trigger. Roll mode plots samples continuously as they arrive — perfect for timebases of 100 ms/div or slower.
How do I leave a scope recording unattended in a customer's car?
Use single-shot with a strict trigger (glitch or pulse-width), deep memory, and a wide capture window. Disable the screensaver, plug the scope into a hard-wired 12 V source (not the cigarette lighter — it powers down), and tape a note to the wheel so the driver knows not to touch the leads.
My intermittent only shows up after a heat soak. How do I capture it?
Set a glitch trigger on the affected signal, put the scope in single-shot, and run the engine to operating temp. Use a heat gun (or freeze spray, depending on whether heat or cold triggers it) on suspect connectors and modules while watching for the trigger.
Related guides
Caught the glitch — now what?
Upload the screen — WaveWrench Pro compares it to a known-good reference, highlights the anomaly, and suggests the most likely root cause.
Open the analyzer