Trigger an external instrument from a frame grabber trigger output

You have a frame grabber pulling video in one window and an instrument — a digitizer, an oscilloscope, a strobe controller, a PLC input — that needs to know exactly when each frame begins. The grabber already knows: it locks to the video's sync and starts each frame at the same point every time. Its trigger output line can hand that instant to any instrument with a trigger input, so both devices stay phase-locked over a single wire. This guide walks the job end to end: which hardware can emit the pulse, how to wire and configure the line, the timing arithmetic that decides whether one pulse per frame freezes your capture, and why a composite (CVBS) source locks while a component YUV source rolls. Everything uses parts already on the bench.

What you need

Three things, and the first one decides the whole scheme.

• A grabber with a configurable trigger output. Frame grabbers built for Camera Link and CoaXPress cameras carry trigger and generic I/O lines as part of their standard connector set. Adaptors that only move network packets do not. Check the model's data sheet for a trigger or DIO connector before you plan anything else.
• The grabber's I/O breakout cable. The I/O lands on a multi-pin connector, and the matching cable splits the pins into open wires you can terminal onto the instrument.
• The instrument's trigger specification. Take four values from its manual: minimum trigger pulse width, threshold voltage, input impedance, and whether the input is single-ended or isolated.

A Camera Link grabber is the class built for exactly this job, trigger lines included:

If your grabber turns out to be a packet mover, take the pulse from the camera instead — most industrial cameras expose digital lines your code can drive as outputs. Using a camera trigger line as a digital output covers that route step by step; the instrument side of this guide applies unchanged, because the instrument only cares that a clean pulse arrives.

Wire the grabber output to the instrument trigger input

Open the grabber manual and find the I/O connector pinout. It names one pin as the trigger or generic output and its neighbour as ground. Then land three conductors:

1. Output pin to trigger input. Run the output wire to the instrument's external trigger terminal.
2. Ground to ground. A logic pulse is a voltage measured against ground, so the instrument must share the grabber's ground reference or it sees noise instead of an edge. Land the ground wire on the instrument's ground terminal, beside the trigger wire.
3. A supply, if the line is opto-isolated. An isolated output switches an external supply rather than driving its own pin voltage; the pinout page states the pin, the polarity and the current the loop needs.

Keep the trigger and ground wires together as a twisted pair, keep the run short — a metre or two when both devices share a rack — and ground any shield at one end only.

The breakout cable made for the grabber's I/O connector does the pin-splitting for you; this one ends in a DB9 female at the grabber and open leads at the far end, with 7 m of cable in between:

Configure the grabber to emit a pulse on frame start

The electrical path is done; now tell the grabber when to fire. In your SDK or the GenICam feature browser, open the digital line controls and set three properties:

1. Line direction: output. The same physical line usually works as input or output; the direction setting picks which.
2. Driving event: frame start. Choose the acquisition event that drives the line. Frame start fires once per frame as readout begins; an SDK may also expose exposure start or per-line events, but frame start is the one that matches frame-based capture.
3. Polarity and pulse width. Set the idle level low and the pulse active-high unless the instrument's manual says otherwise, and program a width that clears the instrument's minimum trigger width — the worked example below picks 1 µs.

If the line refuses to output, the usual culprit is the line feature group not being enabled in the right order. GenICam line settings that block a hardware trigger walks that group from the input side, and the ordering logic transfers directly. For bench bring-up before a video source exists, most SDKs can also emit pulses on an internal timer instead of an acquisition event: generating a trigger pulse train without a function generator shows that variant.

Set the instrument and check the timing budget

Set the instrument to external trigger, rising edge, threshold about a third of the way up the pulse — roughly 1.5 V on a 5 V pulse — so neither rail's noise can false-trigger it. Then run the arithmetic that decides whether one pulse per frame is enough.

NTSC-rate composite video is the classic case:

• Frame rate 29.97 frames per second, 525 lines, interlaced.
• Line rate 29.97 × 525 = 15734 lines per second.
• Frame period 1 / 29.97 Hz = 33.37 ms, so the frame-start pulse arrives every 33.37 ms — about 30 pulses per second.
• Line period 1 / 15734 Hz = 63.6 µs. Check: 525 × 63.6 µs = 33.4 ms ✓.
• Pixel period at the standard-definition 13.5 MHz luma sampling rate (ITU-R BT.601): 1 / 13.5 MHz = 74 ns.

Worked example — freeze one line of CVBS on a digitizer. Assume the digitizer's manual lists a 10 ns minimum trigger pulse width. Program 1 µs: a 100× margin, costing 1 µs / 33.37 ms = 0.003 percent of each frame interval. Trigger on the grabber's frame-start pulse, then use the instrument's trigger delay to pick the line: line 200 sits about 200 × 63.6 µs ≈ 12.7 ms after frame start. Sweep the delay around 12.7 ms until the trace shows the line you want. A trace that then holds still proves the whole scheme: constant pulse-to-video delay only offsets the picture, never smears it.

The same per-line relations govern line-scan timing on production lines — set your own numbers here:

When the pulse rate comes from the camera side rather than the video source — a second camera triggered off the first frame, for instance — frame rate with external hardware triggers gives the matching pulse math.

Why composite locks and component YUV rolls

The reason sits in the signal. A composite video signal carries the picture and its line and frame sync pulses on one wire, as the composite video reference summarises, so a grabber locked to CVBS knows exactly where every frame starts — and its frame-start output inherits that certainty.

Component YUV or YPbPr splits the picture across three wires. The luma wire may carry sync, or the source may provide no composite sync at all; what the grabber receives depends on the source's convention and on whether the grabber's input format is set to match it. If the grabber finds no sync edge it recognises, it free-runs: the image rolls, and a frame-start pulse fired from a free-running readout lands at arbitrary phase — worthless as an instrument trigger. That is why the YUV capture rolls where the CVBS capture holds still, on identical wiring.

Fix the sync first, then trust the trigger:

• Feed the grabber a composite copy. Sources with component output usually also have a CVBS output. Lock the grabber to that, feed the component signal to the instrument, and the pulse is valid.
• Regenerate the sync. A sync separator rebuilds a clean composite sync from a video signal; between source and grabber it hands the grabber the edge it needs.
• Use a dedicated sync line. Some sources expose a separate sync output — wire it to the grabber's sync input and set the input format to match.

Confirm the picture is locked and still before you believe the output pin. Signal checks for a grabber that cannot detect its video source covers the detection side, and scanline distortion on analog video covers the capture settings that follow. For the grabber class itself, a walkthrough of analog frame grabber selection compares the choices for capture-software use.

Drive several instruments or a second card

One frame-start pulse can serve more than one listener. Two or three high-impedance TTL trigger inputs can share the grabber's output wire in parallel. Beyond that, or over longer runs, buffer the pulse through a line driver or trigger distributor so the grabber's output is not loaded down. The electrical budget is the same one behind triggering two cameras from one grabber input: one grabber input, two cameras works through it in detail.

Between two acquisition cards in one PC, an internal trigger bus carries the pulse with no external wire at all — card-to-card frame-trigger sync has its own frame-trigger sync troubleshooting guide. Between instruments from different worlds, the external trigger wire is the common language — everything above is that wire.

Verify the link end to end

Four checks, in order:

• Scope the output pin first. Expect a pulse at the expected rate — about 30 per second for a 29.97 Hz source — with clean edges and the programmed width. No pulse means the direction or the driving event never applied.
• Trigger the instrument and watch the trace. A trace that sits still proves constant pulse-to-video delay. Occasional jumps point at sync or threshold, not at the wire; step changes in brightness that track the trigger are a known symptom family — the gray-level jump diagnostic for pulse-triggered captures sorts them.
• Pull the video source and watch the pulse. It should stop or degrade with the video. A pulse that keeps firing metronome-regular with no source means the grabber is in a free-run or timer mode, and the lock was illusory.
• Check the width at the instrument end. Long wires and heavy loads round the edges; if the pulse arrives under the instrument's minimum width, widen it at the grabber.

Mistakes that cost an afternoon

• Wrong polarity. The instrument waits for a rising edge; the pulse idles high and falls. Select the other edge, or invert the line.
• Threshold at the rail. A 4.9 V threshold on a 5 V pulse misses on droop and noise. Set it a third of the way up instead.
• Ground loop. Two mains-referenced instruments sharing a monitor make hum ride on the trigger. One ground wire, shield grounded at one end.
• Assuming the pulse exists. An adaptor without a DIO connector cannot be configured into one; plan the camera-line route from the start.

When a frame-start pulse is not enough

Frame start marks frames, not lines or pixels. If the capture needs exact alignment at line or pixel precision across the whole frame, one pulse per 33.37 ms is the wrong clock: take line timing from a sync separator's line output, or genlock the chain instead. And when the second device is another camera rather than an instrument, paired-camera hardware sync is its own latency problem — triggering two USB cameras together covers that case with the same trigger discipline.

Parts and next step

If the pulse comes from the camera rather than the grabber, its I/O and power connector breaks out over a dedicated cable; this one carries the 12-pin connector over 5 m:

Browse the frame grabbers and cables categories for the grabber and breakout pair, filter by interface and trigger needs in the accessories finder, or start at the extras step of the configurator and add the trigger wiring to your system. When a capture still will not lock after these checks, the contact page reaches the applications team.