Vibration and Shock

Generic engineering material, applicable to any module from any supplier.

A separate team-process reference can be found read more.

Vibration is continuous and small. Shock is occasional and large. They damage different things and the remedies differ, which is why treating them as one requirement produces a design that survives neither.

What vibration does

Blurs, where the movement during the exposure is comparable to a pixel.

Which is arithmetic: displacement during exposure against millimetres per pixel — and a short exposure fixes it, provided there is light for one.

An independent reference for adjacent engineering topics is Thorlabs.

Distorts, on a rolling shutter, since different rows are read at different points in the vibration cycle — producing a wobble that changes frame to frame and that no exposure setting removes.

Loosens things. Adjustment rings, connectors without latches, fasteners without thread locking.

And fatigues cables, where a flex is subject to continuous small movement it was not specified for.

What shock does

Moves alignment. A single impact can shift a module in its mount or a lens in its barrel, and the system then images correctly but points somewhere slightly different.

Which is the insidious failure, because nothing looks broken and the processing simply starts failing.

Cracks solder joints and connectors, particularly on a heavy component with a long lever arm.

And breaks glass, where a window or lens element is unsupported.

The remedies differ

For vibration: shorten the exposure, stiffen the mount so its resonance is above the driving frequency, and lock everything that can turn or unplug.

Isolation mounts help and introduce a trade — a compliant mount reduces transmitted vibration and reduces alignment stability, which for a measurement system may be the wrong exchange.

For shock: support the module close to its mass, avoid long unsupported cantilevers, and design the mount to relocate positively so that a shifted module returns to position rather than needing realignment.

Resonance

The case that turns modest vibration into a large one.

A bracket has a natural frequency, and driving it near that frequency amplifies the movement substantially.

Which means a mount that is adequate under general vibration can fail badly at one specific machine speed — and the failure appears when the line runs at that rate and disappears when it does not.

Stiffen rather than damp, where possible: moving the resonance above the excitation is more reliable than absorbing at it.

Establishing what the environment actually is

Nobody knows without measuring, and estimates are consistently low.

A recording accelerometer on the intended mounting point, for a shift, produces the frequency content and the amplitude — which is the input every decision on this page needs.

Where that is impractical, the camera itself is an instrument: a long-exposure image of a fixed target smears in proportion to the movement, and the smear is measurable.

And ask the machine builder, who frequently has figures and is rarely asked.

The exposure arithmetic

Blur = amplitude × 2π × frequency × exposure time, approximately, for sinusoidal motion at the peak.

Which converts a vibration specification into a maximum exposure — and that into a lighting requirement, since a shorter exposure needs more light.

The chain runs vibration to exposure to illumination, and it is the reason a vibration problem frequently has a lighting solution.

Where the light cannot be increased, the alternatives are isolation, stiffening, or triggering capture at a point in the machine cycle when movement is least.

Triggering around the motion

The cheapest remedy where the machinery is cyclic.

Most machines have a quiet moment — between index steps, at the top of a stroke, while a part is clamped.

Capturing then removes the problem entirely rather than mitigating it, and it costs a trigger signal the machine controller usually already has.

Which is worth asking about early, since retrofitting a trigger is harder than specifying one.

Transport and installation

A product that survives service can be destroyed in a van.

Shock during shipping frequently exceeds anything in operation, and alignment set at the factory arrives disturbed.

Which argues for a positively located mount and for a documented check on installation — a reference image compared against the commissioning one, which takes a minute and catches it.

In one line

Measure the environment before designing the mount — estimates of vibration are consistently low.

The connector, specifically

The component vibration finds first.

A fine-pitch connector without a latch works loose over months, producing intermittent faults that arrive long after installation and are blamed on everything else.

Specify a latching connector where vibration exists, and strain-relieve the cable near it so the joint is not carrying the load.

And check it at service rather than assuming, since reseating a connector is thirty seconds and diagnosing one is a day.

Where the camera is on the moving part

A different problem from a camera watching one.

A module on a robot arm, a gantry or a rotating head experiences acceleration rather than ambient vibration — large, predictable, and cyclic with the motion.

Which makes triggering easy, since the controller knows exactly where the arm is, and makes the cable the hard part: a dynamic flex rated in cycles rather than a standard one.

And the settling time after a move becomes the limiting factor on cycle rate, which is a mechanical specification rather than an imaging one.

What to write in the requirement

Frequency range and amplitude, or a reference to a standard test profile appropriate to the application.

Shock magnitude and duration, separately.

And whether the equipment must operate during vibration or merely survive it — which are very different requirements and are frequently conflated in a single line.

A camera that must produce measurements while a machine runs has an operating requirement; one that must not break during transport has a survival one.

The short version

• Vibration is continuous and small, shock is occasional and large, and they damage different things
• Vibration blurs, distorts on a rolling shutter, loosens fasteners and connectors, and fatigues cables
• Shock moves alignment without breaking anything visible, which is why the failure looks like a processing problem
• For vibration: shorten the exposure, stiffen the mount, lock everything that can turn
• Isolation mounts reduce transmitted vibration and reduce alignment stability, which may be the wrong trade for measurement
• Resonance turns modest vibration into a large one at one specific machine speed, and stiffening beats damping