You are checking internal gear teeth for presence and for broken or chipped teeth inside a bore of roughly 100 mm, and the lamp on the fixture lights only one side of the wall. The fix is geometry, not intensity: a 360° inspection lens sits on the bore axis and images the whole circumference in one shot, while an on-axis light reaches the wall evenly all the way round. This guide gives the decision in data-sheet terms: how far the wall sits from the probe, how many pixels the unfolded circumference needs, whether one focus setting survives the bore tolerance, and how to size the light. Each section ends with a value you can check against a lens data sheet before ordering.
Why does a side light only light part of a bore wall?
Light travels in straight lines, so a lamp set at one side of the bore mouth lights the arc of wall facing it — roughly a quarter to a third of the 360° circle — and the gear teeth behind that arc sit in shadow.
On a machined wall the lit arc is specular: the law of reflection sends each ray back where it came from, so the bright zone follows the angle, not the wattage. A bigger lamp widens the glare, not the coverage.
Grading teeth makes this worse. Thresholds that pass a lit tooth must also pass the same tooth one position around — a crescent of glare next to a dark arc fails that test on any polished bore. The same geometry governs bright field versus dark field lighting, and it is the recurring finding in pipe internal surface inspection: a single-direction source cannot be fair to 360° of surface. The cure is symmetry — light the wall from the axis, and image it from the axis.
What does a 360° inspection lens see inside a bore?
A panoramic probe sits on the bore axis and folds the view of the surrounding wall into one image, so one shot covers the full πD circumference — 314 mm on this bore — instead of one arc at a time.
Inside the probe, a mirror cone folded around the axis bends light arriving from every side into the lens. The image shows the wall as an annulus or a strip, with the bore's depth running radially across it. Nothing rotates, so one exposure freezes the whole surface — no part indexing, no stitching, no seam where two views meet.
The 360° inspection lens category splits into plain panoramic models and multi-angle models such as the LS-PCL-0057 and LS-PCL-0066, which view the wall at several angles at once. The split matters by defect: a missing tooth is diffuse contrast and shows in any view. A hairline crack or scratch is grazing-angle and may appear only in the steeper view, which is the reason to keep the multi-angle option in mind for surface-finish defects. The companion piece on the 360 degree inspection lens for inner wall, outer wall and bore imaging covers all three targets; this guide stays inside the bore.
How do you match the lens to the bore diameter?
The wall sits at about half the bore diameter minus the probe radius, so with a 30 mm probe in this bore the wall sits 35 mm from the probe, and the lens's printed working distance must cover that.
Wall distance w, bore diameter D, probe radius r. The catalogue sorts its 113 lenses by exactly this number: 79 print a working distance under 65 mm and 7 print 65-110 mm. This bore lands in the first group.
Whole outer wall or bore in one shot
113 models · from $899.00| Bore diameter | Probe diameter | Wall distance w |
|---|---|---|
| 60 mm | 30 mm | 15 mm |
| 100 mm | 30 mm | 35 mm |
| 140 mm | 30 mm | 55 mm |
| 200 mm | 30 mm | 85 mm |
For the tight rows the compact models take over: the LS-PCL-0042 works from 12.5 to 50 mm on a 1/3" sensor, and the LS-PCL-0057 is a 2/3" multi-angle lens with a printed working-distance range of 23-85 mm. That 23-85 mm band carries the whole 15-55 mm wall column of the table — one lens and one focus across four bore sizes.
For the far rows, the LS-TCL-0054 is a 2/3" C-mount lens whose data prints a working distance of 65-110 mm. With a 30 mm probe again, a wall in that 65-110 mm band maps to bores of roughly 160-250 mm. If pixel count drives the choice, the LS-PCL-0015 is the 1" option, with an 11-16 mm image circle and a working distance under 65 mm.
Probe diameter is the mechanical half of the match. The LS-TCL-0040 is a slim C-mount probe with printed dimensions of 30 x 138.18 mm. As a rule of thumb, leave about 5 mm between probe and wall so the light cone clears burrs and tool marks.
Decision for this bore: shortlist the panoramic lenses whose printed working distance covers a wall at 35 mm — the multi-angle 2/3" model or the 1" model above — and move to the 65-110 mm class only when the wall sits beyond 65 mm.
All three candidates fit C-mount cameras; their printed rows:
| Specification | LS-PCL-0057 | LS-PCL-0015 | LS-TCL-0054 |
|---|---|---|---|
| Working distance | 23~85 mm | 44.3 mm | – |
| Sensor format | 2/3" | – | 2/3" |
| Mount | – | C-mount | C-mount |
| Price | $1,715.00 | $909.00 | $3,639.00 · quoted |
Values as printed in each product's data sheet; open a part for its drawing and full specifications.
How many pixels does the bore wall need?
Unfold the circumference first: this bore wraps 314 mm of wall into one image, and catching a 0.5 mm chip with 3 pixels across it needs a pixel footprint of 0.17 mm, about 1,880 pixels across.
C is the unfolded circumference in millimetres, px the pixel footprint on the wall, N_px the sensor's pixel count across the image, n the pixels you demand across a defect of width w. For this bore and a 0.5 mm chip at 3 pixels across it: N_px = 3 × 314 / 0.5 = 1,884.
| Camera class | Pixels across, typical | Footprint on 314 mm wall | 0.5 mm chip | 0.1 mm crack |
|---|---|---|---|---|
| 2 MP | 1,920 | 0.164 mm | 3.0 px | 0.6 px |
| 5 MP | 2,448 | 0.128 mm | 3.9 px | 0.8 px |
| 12 MP | 4,096 | 0.077 mm | 6.5 px | 1.3 px |
A missing tooth is several millimetres of gap and every row sees it; the chip or crack edge is what sets the camera. If the pass/fail feature is a 0.1 mm crack, the requirement becomes 3 × 314 / 0.1 = 9,420 pixels across — beyond any single panoramic image at these classes. Split the task: one panoramic pass for tooth presence and a close-up second station for the crack call.
The camera resolution calculator runs this same division on your own bore and defect:
Will the whole wall stay in focus across bore tolerance?
With a pixel footprint of 0.128 mm on the wall and 3.45 µm pixels, effective magnification is about 0.027x, and the thin-lens estimate gives roughly 39 mm of depth of field at F/4 — enough to hold focus across a bore tolerance of tens of millimetres.
p is the sensor's pixel pitch, c the blur you accept, N the f-number. In the 5 MP row, px = 0.128 mm and p = 3.45 µm give m = 0.00345 / 0.128 = 0.027. At F/4 the estimate is 2 × 4 × 0.00345 mm × 1.027 / 0.027² = 39 mm of total depth, about ±20 mm around the focused wall. Stopping down to F/8 doubles it to about 78 mm, but the Airy disk grows to 2.44 × 0.55 µm × 8 = 10.7 µm — 3.1 sensor pixels — and diffraction starts erasing the defect edges you stopped down to keep. F/5.6 is the practical middle: about 55 mm of depth with a 7.5 µm Airy disk.
Set focus once near the wall distance of your median bore and the estimate holds across tolerance swings of tens of millimetres. Treat the depth figures as a thin-lens estimate with the mirror fold ignored, then confirm on the shallowest and deepest wall you will see. The trade is set out in aperture, f-number and depth of field in machine vision cameras, and the diffraction limit versus pixel size shows where stopping down stops paying.
Which light reaches the whole circumference?
A source that is symmetric about the bore axis lights every degree of the wall equally, so mount a coaxial light of about the bore diameter over the mouth and the wall gets even light around the full 360°.
The coaxial lights category works here because of how the light is built: a beam splitter folds the LED ring into the optical path, so light leaves along the same axis the lens looks down. Mounted on the front of the 360° lens, the light travels down around the mirror cone and strikes the wall at near-normal incidence at every angle. Each tooth reads with the same brightness, and one threshold set survives the whole annulus. The same axis-light logic is what makes coaxial lighting work on engraved characters in recessed metal.
Size the light with one geometric rule: the clear output must be at least the bore diameter, or the wall just inside the rim sits in shadow, because light cannot bend around the mouth edge.
The LS-CXL-0067 is a 102 x 100 mm coaxial light source that runs from 24 V — its clear output matches this bore mouth. The LS-CXL-0092 is a 70 x 70 mm unit for the 50-100 mm size class.
- Size
- 100-200 mm
- Voltage
- 24 V
Colour follows the defect: white when the failure is a colour — rust, stain, coolant film — and red with a bandpass filter against ambient light when only geometry matters and plant lighting must be rejected. The wall sits tens of millimetres from the output, so intensity is not the constraint; run continuous on a static fixture and strobe sub-millisecond when parts index.
To pair the light with the bore and drive it from the trigger, start the configurator at the lighting step:
The configurator matches a camera, lens, light, light controller and cables to each other, step by step, and lists only parts that fit what you already picked.
When is a 360° inspection lens the wrong choice?
Reach for another tool when you need defect depth, when the bore is barely wider than the probe, or when the bore runs deeper than the focus band — a 360° lens judges presence and surface condition, not 3D shape.
Crack depth needs range data and a 2D image has none. If the disposition of a defect depends on how deep it runs, add a 3D station rather than push the panorama past what it measures. Clearance is the second limit: when the bore is barely wider than the probe, there is no room for the light cone, and a rigid side-view borescope or a rotated two-pass layout replaces the panorama.
Reach is the third limit. The probe images a band of wall about one depth-of-field thick around itself, so a bore much deeper than that band needs several stations or an axial scan between shots. Dimensional checks tighter than roughly 0.05 mm belong elsewhere too: rotate the part under a telecentric machine vision lens instead — the entocentric versus telecentric lens comparison gives the selection logic.
Within its limits — presence, chips, cracks, corrosion and coating damage, all around the wall in one shot — the panoramic lens is the right tool, and no stitched mosaic matches its fairness across 360°.
What do you check before ordering?
Pin five numbers before you order: bore diameter and its tolerance, wall distance against the printed working distance, pixels across the unfolded circumference, probe clearance, and the light's clear output against the bore mouth.
- Bore diameter and tolerance. Measure D and its tolerance, compute w = D/2 − r, and require the printed working distance to cover w across the whole tolerance, not just nominal.
- Probe clearance. Check the printed probe diameter; leave about 5 mm to the wall so light and burrs clear.
- Sensor and image circle. The sensor must fit the printed image circle; formats run from 1/4" to 1", and 107 of the 113 lenses are C-mount.
- Pixels. Carry the N_px ≥ n × πD/w result with margin — the 1,884-pixel figure buys a 5 MP sensor, not a 2 MP one, when cracks matter.
- Focus budget. Compare the depth-of-field estimate with the bore-to-bore spread; if it fails, plan per-SKU refocus with locked recipes.
- Light. Clear output at least the bore diameter; colour chosen by defect; a bandpass filter if plant lighting leaks in; strobe capability if parts index.
- Worst-case exposure test. Shoot the oiliest, darkest part and check uniformity around the whole annulus, not one arc.
Then size the parts: the lens finder filters the 360° series by working distance and sensor format, the light finder by size and colour, and the lens configurator walks the chain from bore to sensor. For a bore outside the standard ranges, contact us with the drawing and the defect list. The station-level order of decisions — lighting, lens, then exposure — is laid out in stable 2D vision inspection systems, and further lens material sits on the machine vision lens topic page.
Questions
Can one 360° lens cover several bore sizes on the same machine?
Often yes. With a 30 mm probe, a working-distance band of 23-85 mm spans bores from roughly 75 mm to 200 mm across: the probe radius sets the near limit and the optics set the far one. Keep focus mid-range so the ±20-30 mm depth-of-field estimate absorbs the step between sizes, or refocus per SKU and store the recipe with each part program.
Does the panoramic image need unwrapping software?
No. Template, blob and defect tools work on the raw annulus as it comes off the sensor; the polar-to-rectangular unwrap is cosmetic and, when you want it, a few lines in any vision library. Calibrate before you measure on the unwrapped map: the scale varies with distance from the image centre, so measurements belong on the annulus where the scale is known.
Is a 2 MP camera enough for a 100 mm bore?
For presence checks, yes: 1,920 pixels across a 314 mm unfolded wall gives 0.164 mm per pixel, so a 0.5 mm defect still spans about 3 pixels. For crack edges it is thin: 5 MP raises that to 3.9 pixels and 12 MP to 6.5. A 0.1 mm crack needs about 9,400 pixels across, which means two views rather than a bigger sensor.
Can the same idea inspect the outside of a shaft?
Yes. The LS-PCL-0081 is a 360° imaging system for outer surfaces on a 2/3" C-mount camera: it looks at a round part from the side and captures the full circumference in one shot. It suits O-ring grooves, threads and surface defects on shafts and housings that cannot be rotated under a standard lens.
How do I keep the probe window clean in production?
Treat it as a wear point: every ray to the sensor crosses the probe's front optics, so one oily fingerprint smears the whole 360° image at once. Plan an air purge or a wiper at the bore mouth, check uniformity on a plain target at shift start, and hold a spare probe for stations running oiled parts.