Analog camera image split in two with a black line: fix the cable

You install an analog camera on a frame grabber, click Live, and the image arrives split into two identical halves with a black horizontal line between them — and the line jumps to a new position on every acquisition. Calibration then hands back a white, washed-out frame. Both symptoms look like trigger and software problems, and both usually are not. This guide runs the diagnosis in the order that wastes the least bench time: read the split image for what it says about sync, prove the cable, then fix the exposure mismatch hiding behind it. You finish with a tested video path, the correct switch positions, and a decision on whether to keep the analog link or replace it.

Why is the image split in two by a moving black line?

Because the frame grabber has lost sync lock, so it slices the video signal at the wrong points: the black bar is the vertical blanking interval landed mid-frame, and each half is the same field sampled twice. The picture content itself is healthy — the timing reference is not, and that reference travels in the cable.

An analog camera sends one composite signal: picture, blanking and sync pulses together, 1 V peak-to-peak into 75 Ω, with sync occupying the bottom 0.3 V of that range. The grabber phase-locks its sampling clock to the sync tips. RS-170 video runs 525 lines at 30 frames per second, a 15,750 Hz line rate; CCIR runs 625 lines at 25 frames per second, 15,625 Hz.

With sync intact, every line and every field starts at the same instant. With the sync conductor open, crossed to another pin, or floating without its ground reference, the lock drifts and each acquisition samples the signal at a new phase — so the blanking bar lands somewhere different every time. The doubling appears when the lock settles on a stable wrong offset, typically half a field, and the scene is painted twice around the stranded bar.

Is the fault in the trigger settings or in the cable?

It is the cable in most first installs that use a lead other than the grabber vendor's own: a split image that changes on every frame is a video-path fault, while a trigger fault gives a black, frozen or correctly framed image — never two clean copies of the scene. Read the screen before touching a single setting:

A ground loop drifts slowly and never duplicates the image; a sync fault doubles it sharply. When the cable is proven good and the image still tears, the grabber's own timing controls come next — analog video scanline distortion: check frame grabber settings walks those settings. Robot cells show the same failure class as a black band through the image, and the bench order is identical: see this camera black-line corruption case on a robot vision controller.

Why do the recommended switch positions give a black image instead?

Because those positions route the camera to external sync inputs or an external trigger arriving on pins your cable does not connect, so the camera waits indefinitely for a pulse that never reaches it and outputs no video at all. That black frame is a diagnosis, not a second fault.

Switch tables in analog camera manuals assume the reference harness: every sync, trigger, power and ground conductor present on defined pins. A generic or re-pinned cable breaks that assumption quietly. Video survives on the pins that happen to match, everything the recommended positions need is absent, and a camera waiting on external signals outputs a clean black — the cable has just told you which lines it lacks.

The combination that produced a stable live image put the camera on internal sync and free-running trigger; use it as the bench mode. Moving to hardware triggering later means proving the trigger conductor end to end first, and the timing chain behind that decision is covered in trigger call versus exposure time: understand capture timing.

Keeping the analog link? The software side of that grabber has its own checklist in analog frame grabber integration with vision software.

How do I test the camera cable pinout and shield?

Test it with a multimeter on continuity, both ends unplugged: every pin must read to its counterpart, the shield braid must read end to end below about 1 Ω, and any open, cross or missing ground on the sync line explains the split.

1. Unpower and unplug both ends before measuring — an ohmmeter across a live line pushes current into the camera output stage.
2. Write the two pinouts side by side from the camera manual and the grabber manual: video, video ground, sync, trigger input, power, chassis.
3. Map signal to signal. Any camera signal with no same-named grabber pin is the fault; stop and order the correct lead.
4. Buzz the assembled cable pin to pin, then braid to braid. Expect near 0 Ω per conductor and under 1 Ω on the shield.
5. Check adjacent pins for shorts, above all sync against power — a sync line crossed into a supply pin can damage the camera output stage.
6. Fit the matched lead made for that camera family if one exists. If the split clears, the case is closed.

With a scope, set the input to 75 Ω and look at the signal: healthy composite shows 1 V peak-to-peak with sync tips 0.3 V below blanking. Sync that is flattened or riding a noise floor convicts the shield and ground path.

The same pin-by-pin discipline settles digital cases too: USB camera keeps disconnecting: isolate cable and hardware faults applies it to a different connector. When you order the replacement, buy the lead made for the camera family rather than a generic one — the pinout is the product:

Right-angle connectors fit tight camera backshells where a straight lead will not:

Browse the full camera cables list or the accessories finder for the connector your camera uses.

Why does Live look right but the calibration frame come out white?

Because live view and single-shot capture run different exposures: live shortens exposure to keep the frame rate up, while Acquire applies the stored value — and 20 ms of stored exposure against the 0.5 ms you watched live is 40 times the light, which saturates analog white completely. The blur is the same overexposure plus residual sync jitter, which is why the cable is fixed first.

Worked example, with the numbers you would read off a working set-up:

At the same f/4, 5.3 stops over means the signal clips at white and the calibration grid washes out. Two ways back:

1. Lock the still exposure to the live value, 0.5 ms — manual exposure, every auto mode off, before calibrating. This is the decision.
2. Or close the iris by the square root of 40, about 6.3 times: f/4 becomes roughly f/25. Many C-mount irises close no further than f/16, so the iris alone cannot absorb 5.3 stops.

Mind the frame-period ceiling as well: one frame lasts 33.3 ms at 30 fps and 40 ms at 25 fps, and an exposure demand above it cannot fit in a single frame. The camera accumulates charge across frames and the still arrives smeared on top of being white.

The live-versus-still brightness mechanism is the one behind video exposure and frame rate: why still images look brighter.

Which settings should I run while the correct cable is on order?

Run internal sync and internal free-run trigger — the positions that produced the stable live image — and let the grabber lock to the camera's video; that needs nothing from the cable except video and ground, which isolates the fault to the wiring.

• Sync source: internal. The camera generates its own timing, and nothing has to arrive from outside.
• Trigger: internal free-run for set-up. Move to hardware trigger only after the trigger conductor reads end to end.
• Shutter: fixed, never auto — auto shutter chases the lighting and moves the calibration with it.
• Gain: fixed at the lowest value that gives usable contrast. Auto gain changes image statistics between live and stills.
• Exposure: manual, at the value live view settled on, 0.5 ms in the example above.

Recalibrate after every change. A plate captured on auto exposure will not match production frames, and calibration fails again for a reason that has nothing to do with wiring.

When should I stop repairing analog and move to a digital camera?

Replace the analog link when the matched cable is unobtainable, the camera is end of life, you need frame-accurate trigger-to-image timing, or ground-loop hum survives single-point grounding — a digital link moves sync into the protocol, where a wrong cable produces no image instead of a half-image. Resolution is the other lever: an analog frame digitizes to roughly 0.3-0.45 MP, while a current digital model starts near 5 MP.

Camera Link keeps the deterministic timing analog never had, over a standardized cable pinout, and the compatibility questions for that route are covered in Camera Link frame grabber compatibility for line scan. On the Camera Link side, grabber and cable are bought as a matched pair:

The LS-CBL-0191 is a 5 m Camera Link camera cable, SDR to SDR.

For longer runs and simpler cabling, GigE carries video, power and trigger on one lead — start from the area scan cameras list or the camera finder.

What should I confirm before ordering parts?

Confirm the camera connector's part number and pinout, the grabber connector's part number and pinout, the conductor count your signals need, and the length and flex the route sees — in that order, because the pinout is the fault you are chasing and a missed conductor recreates it.

• Pinout mapped signal by signal, camera manual against grabber manual.
• Continuity under 1 Ω on every conductor and the shield, measured on the unplugged cable.
• Grabber input termination at 75 Ω, with no passive tee on the line.
• DIP switches photographed; camera on internal sync for the bench test.
• Exposure locked to a measured value, auto modes off, before calibration.
• Cable rated for the route: drag-chain grade where it moves, shielded where power runs alongside.
• Analog replacement parts ordered as one matched set — camera, lead, grabber.

The next step is mechanical: run the system configurator to pair a camera with its lead and grabber, or send the two pinouts to us through the contact page and we will map them for you.