HOW-TO GUIDE
How to Trigger an Oscilloscope on a Runt Pulse
By the OhmSavvy Research Team ·
Safety: know where the ground clip goes
- On a mains-powered bench scope, the probe ground clip is connected to mains earth. Clip it only to the circuit's ground; touching it to any other voltage shorts that point to earth.
- Runts often show up on gate-drive and control signals inside switching supplies and motor drives, where the local ground may sit at mains potential. Use a rated differential probe or an isolation transformer on the device under test, and never float the scope.
What a runt pulse is
A digital input sees three regions: below its low threshold it reads 0, above its high threshold it reads 1, and in between the result is not guaranteed. A healthy pulse crosses both thresholds quickly. A runt crosses the first threshold and turns back before reaching the second, so it spends its whole life in the undefined band. One gate may count it and another may ignore it.
That inconsistency is what makes runts so troublesome. A clock line with an occasional runt can advance one counter and not another, so two parts of a system disagree about what time it is. The symptom is often a rare crash, a corrupted byte or a board that fails only when warm, with nothing obviously wrong when you look at the signal with a normal edge trigger.
The rule a runt trigger applies
1Check which trigger types your scope has
Open the trigger menu and look through the type list. Many digital scopes include a runt trigger, sometimes alongside a window trigger or a pulse-width (glitch) trigger, but availability varies by model, firmware and installed options. Our product data for the scopes we review does not confirm which of them offer a runt trigger, so check your own manual rather than assuming. If you have one, carry on with steps 2 to 4. If not, skip to the workarounds. Before any of it, compensate the 10x probe and use a short ground; see how to connect an oscilloscope probe correctly.
2Set the upper and lower thresholds at the logic levels
The two thresholds define what counts as a runt, so they should match what the receiving chip treats as a valid low and high. Its datasheet lists them as VIL (maximum input low) and VIH (minimum input high). Set the lower threshold at or just above VIL and the upper one at or just below VIH. Measure the normal high and low levels first so you know both thresholds sit well inside the real swing.
| Logic family (typical) | Supply | Lower threshold (VIL) | Upper threshold (VIH) |
|---|---|---|---|
| 5 V CMOS, e.g. 74HC | 5 V | 1.5 V | 3.5 V |
| 5 V TTL, e.g. 74LS / 74HCT inputs | 5 V | 0.8 V | 2.0 V |
| 3.3 V LVTTL-compatible | 3.3 V | 0.8 V | 2.0 V |
These are common textbook values; individual parts differ, so take the real numbers from the receiving chip's datasheet. If you are unsure, a starting point of roughly 30% and 70% of the supply catches most runts on a CMOS signal. Make sure the probe attenuation set on the channel matches the probe, or every threshold will be off by a factor of ten.
3Choose polarity and, if offered, a width limit
A positive runt rises from low and fails to reach the upper threshold; a negative runt falls from high and fails to reach the lower one. Choose the polarity that matches the fault, or "either" if your scope offers it and you do not know yet. A normally-low clock or chip select line usually shows positive runts; a normally-high reset or active-low line usually shows negative ones.
Some runt triggers also let you qualify the runt by width, such as narrower or wider than a set time. Leave that off at first so you catch everything, then use it to separate two different kinds of runt once you have seen them.
4Arm Normal or Single mode and wait
In Auto mode the scope sweeps on its own when no trigger arrives, which would overwrite the rare runt within moments. Use Normal mode to see every runt as it happens, or Single to freeze the first one. Place the trigger point near the center of the screen so you can see what led up to the runt as well as what followed it, then run the circuit through whatever conditions bring the fault on.
When it fires, measure the runt's peak, width and timing relative to neighboring edges with cursors. If you have a spare channel, put the clock, the supply rail or the driving signal on it, because the relationship between the runt and those signals is usually what reveals the cause.
Measure the runt once you have it
A single frozen capture is evidence, but a few numbers make it useful to whoever has to fix the circuit. Record these for each runt you catch:
- Peak level. How far it got towards the upper threshold. A runt that just misses VIH behaves very differently from one that barely leaves ground.
- Width at the lower threshold. Use time cursors where the runt crosses the lower threshold going up and coming down. Compare it with the narrowest legitimate pulse.
- Position relative to the clock. A runt that always sits a fixed time after a clock edge has a timing cause; one that floats at random usually comes from noise, crosstalk or an asynchronous input.
- Rate. Count how many appear per second or per operation with the scope in Normal mode. A rate that changes with temperature, supply voltage or load is a strong clue.
Keep the thresholds identical while you change one thing at a time in the circuit. If the runt rate drops to zero after a change, you have probably found the cause. For width measurements in general, see how to measure pulse width.
Workarounds if your scope has no runt trigger
A plain edge trigger cannot isolate a runt on its own. Set the level below the runt's peak and both the runt and every normal pulse cross it; set it above the peak and only normal pulses fire. You can still find and measure runts with a combination of tools:
- Edge trigger plus infinite persistence. Trigger on the normal rising edge in Normal mode, turn on infinite persistence and leave it running. A runt appears as a faint pulse that stops short of the bright band of normal ones. Our persistence guide covers this method in detail.
- Pulse-width trigger. Many runts are also narrower than normal pulses. A trigger set for pulses shorter than the shortest legitimate width often catches them.
- Trigger on what causes it. If you suspect a particular event, such as a relay closing or a bus changing direction, trigger on that signal and watch the line with the runt on another channel.
- Long capture and zoom. Capture a long record in Single mode and scroll or zoom through it looking for short pulses. This works best when runts are frequent.
01
Probe set up
10x, compensated, short ground
02
Measure levels
Normal high and low
03
Thresholds
At VIL and VIH
04
Arm
Normal or Single mode
05
Correlate
Clock, rail or source on CH2
What causes runt pulses
A runt says that something tried to switch the line and either stopped or was overpowered. The timing of the runt against other signals points to which:
| Likely cause | Typical clue on the scope | What to check |
|---|---|---|
| Logic race or hazard | Runt lines up with another input changing | Gate logic, add a register or synchronise |
| Metastable flip-flop | Runt or late edge after an asynchronous input | Synchronizer stages on the input |
| Bus contention | Runt when two drivers are enabled at once | Enable timing, direction control |
| Weak driver or heavy load | Slow edges that sometimes do not finish | Load capacitance, pull-up value |
| Supply droop or ground bounce | Runts coincide with dips on the rail (CH2) | Decoupling, ground path |
| Crosstalk | Runt aligns with edges on a neighboring line | Routing, spacing, edge rates |
Before chasing the circuit, rule out the measurement: a long ground lead or a 1x probe can turn a clean narrow pulse into something that looks short and low. If the runt changes when you change the probe setup, it was never real. Signals that look grainy rather than short may need the noisy waveform checks instead, and a single tall spike rather than a short pulse is covered in how to capture a brief voltage spike.
Which scope you need for runt hunting
A narrow runt only looks like a runt if the scope has the bandwidth and sample rate to show its true height; too little and a short full-height glitch is rounded down into a fake runt. A second channel for the clock or supply rail makes finding the cause much faster. For this kind of digital debugging a bench scope is the right tool, and our best beginner oscilloscopes roundup compares them. Check each model's manual for its trigger list before buying.
Bench scopes in our reviews
100MHz, 4 channels, 1 GSa/s; more bandwidth than the Rigol, for projects that will get more demanding.
50MHz, 4 channels, 1 GSa/s; the default beginner recommendation for years.
100MHz, 2 channels, 1 GSa/s; an oscilloscope and waveform generator in one unit.
As an Amazon Associate I earn from qualifying purchases. Prices and availability are shown on Amazon.
Frequently asked questions
What is the difference between a runt pulse and a glitch?
A glitch is usually defined by time: a pulse narrower than the circuit should ever produce, often at full amplitude. A runt is defined by amplitude: it crosses one logic threshold but not the other, whatever its width. One fault can be both, a narrow pulse that never reaches full height, so either trigger may catch it.
Can a runt pulse cause a microcontroller to malfunction?
Yes. Because a runt sits between the logic thresholds, one input may read it as a high and another as a low, or a flip-flop clocked by it can be left in an undefined state for a while. The result is an extra or missing count, a corrupted bit, or a fault that only appears on some boards.
Why does my runt trigger fire on every pulse?
Usually the upper threshold is set above the normal high level, so every pulse fails to reach it and counts as a runt. Measure the normal pulse top first and set the upper threshold comfortably below it. Also check the probe attenuation setting, because a wrong 1x or 10x setting scales every threshold you enter.
Does probe bandwidth matter when hunting runts?
It can. A narrow runt passed through a 1x probe or a scope near its bandwidth limit is rounded off and reads lower than it really is, so a short full-height glitch can look like a runt. Use a compensated 10x probe with a short ground and keep the bandwidth limit off while you hunt.
Can a logic analyzer catch runt pulses?
Not reliably. A logic analyzer compares each input against a single threshold and records only high or low, so a runt either looks like a normal pulse or vanishes, depending on which side of that threshold it peaks. It is still useful for finding when the fault happens and what the bus was doing at the time; then use the scope's runt trigger to see the real amplitude.
