Setting Triggers & Timebase for Automotive Signals — Capture Settings, Sample Rate, Probing

Why capture settings matter more than the scope

A $400 USB scope with the right settings will beat a $4,000 bench scope set up wrong. Three knobs do 90% of the work: timebase (how much time fits across the screen), voltage range (volts per division), and trigger (when the scope arms and starts drawing). Get those right and the shape practically diagnoses itself.

Timebase quick-reference

Ignition secondary (single event): 1–2 ms/div
Ignition parade (all cylinders): 20–50 ms/div
Fuel injector (saturated): 2 ms/div
Peak-and-hold injector: 500 µs/div to see the peak phase
CKP / CMP at idle: 5–10 ms/div
Narrowband O₂: 200 ms/div – 1 s/div
CAN bus frame: 50–100 µs/div
Relative compression (cranking): 100–200 ms/div, 5 s total capture
Fuel pump current: 2 ms/div for commutator ripple, 1 s/div for run-current trend

Voltage range — leave headroom

12 V chassis signals: 5 V/div (60 V full screen of headroom for inductive kicks).
Injector inductive kick can reach 80 V — use 10 V/div or a ÷10 attenuator if your scope clips.
Secondary ignition: use the dedicated capacitive pickup with the scale it specifies (often 10 kV/div).
Low-voltage sensors (O₂, knock, MAP analog): 200 mV/div to 1 V/div with the channel DC-coupled.
If the trace clips at the top or bottom of the grid, the reading is wrong — back off the V/div before you trust the shape.

Trigger modes, demystified

Auto vs. Normal vs. Single

Auto: scope draws a trace even with no valid trigger. Good for hunting an unknown signal, bad for stable analysis.
Normal: scope only updates when the trigger condition fires. This is what you want once you know the signal.
Single: arms once, captures one event, then stops. Use this with a glitch or pulse-width trigger to catch intermittents.

Edge trigger

Pick the rising or falling edge of the channel with the cleanest, most repeatable transition. For an injector, trigger on the falling edge of the ground-side pull-down. For ignition, trigger on the primary turn-off (or secondary firing line). Set the trigger level above the noise band, not inside it.

Pulse-width / glitch trigger

The single most underused feature on modern scopes. Tell the scope "trigger when the low pulse is > 5 ms" and you'll catch a sticking injector that only drags every 200 firings. Combine with single-shot mode for hands-free capture.

Sample rate and memory depth

Sample rate sets the highest frequency you can resolve (Nyquist: at least 2× the signal frequency, in practice 5–10×). Memory depth sets how long you can capture at that sample rate. A scope advertised as "1 GS/s" with only 10 k of memory drops to 10 kS/s the moment you stretch the timebase out for a cranking capture — and your starter-current humps turn into staircases.

Always check the actual sample rate displayed on screen after you set the timebase.
For relative-compression and intermittent captures, prioritize deep memory (≥1 MS per channel).
If your scope has "high-res" mode, use it for low-voltage sensor work — it averages samples into 12-bit values and kills noise.

Probing — where most readings go wrong

Ground placement matters. Ground at the battery negative or engine block. Grounding to a sensor return adds 50–200 mV of garbage.
Back-probe at the connector with a proper back-pin, not by piercing wires. Pierced insulation corrodes and creates the next intermittent fault.
Keep probe leads short. A 2 m clip-lead loop is an antenna — it'll pick up coil noise and show false oscillations.
Current clamps: always zero the clamp around the wire before measuring. A 50 mA offset wrecks fuel-pump and relative-compression readings.
Differential probes are required for CAN-H/CAN-L if you want to see the differential signal without ground-loop artifacts.

A repeatable capture checklist

Pick the channel and probe type for the signal.
Set V/div so the expected peak fills ~60% of the screen.
Set timebase so 2–5 cycles of the event fit on screen.
Choose trigger source, edge, and level — confirm the trace locks in Normal mode.
Verify sample rate is at least 5× the highest frequency in the signal.
Capture, then save the full screen including grid, scale, and trigger markers.

Frequently asked questions

What timebase should I use for an ignition waveform?

1–2 ms/div for a single event (firing line, spark line, ringdown), 10–20 ms/div if you want to see several cylinders in a row at idle. Anything slower than 5 ms/div will smear the ringdowns into a blur.

How much sample rate do I really need?

For most chassis work 1 MS/s per channel is enough. CAN bus at 500 kbit/s wants at least 10 MS/s to resolve clean edges; FlexRay and high-speed Ethernet need a faster scope. Slower signals (O2, MAP, TPS) are happy at 100 kS/s.

Edge trigger or pulse-width trigger?

Use edge trigger for periodic signals (CKP, ignition, injectors) — it locks the trace at a repeatable point. Use pulse-width trigger to catch a too-short or too-long pulse such as a dragging injector or a missing CKP tooth.

Why is my trace jittery and unstable?

Almost always a trigger problem: wrong source channel, level set inside the noise band, or auto-trigger fighting a slow signal. Switch to normal mode, raise the trigger level above the noise floor, and pick the channel with the cleanest edge.

Related guides

Not sure if your capture settings are right?

Upload the screenshot — WaveWrench Pro flags clipped traces, undersampled signals, and bad trigger placement before it diagnoses the fault.

Open the analyzer