Connector inspection looks like a two-dimensional problem and mostly is one — until coplanarity enters the drawing. The useful split is between features defined in the image plane, which a one-shot system handles well, and features defined in Z, which need a different sensor no matter how good the camera is.
In-plane features: pin pitch, pin width, pin length, terminal width, the position of each pin relative to a datum, window openings in the housing, and the presence or absence of a latch. These are projection measurements and a backlit one-shot system is the natural instrument.
Out-of-plane features: coplanarity of the contact row, seating-plane height, terminal bend out of the plane, and the height of a press-fit shoulder. These are defined along Z. A pure 2D shadow only shows them as a slight change in silhouette width, which is not a measurement of the feature.
Surface features: plating colour, contamination, scratches, solder wetting. These are appearance judgements. A measuring system can image them, but it is not a plating-thickness instrument and should not be presented as one.
Pitch is an edge-to-edge distance between adjacent pins. Backlight turns each pin into a silhouette with a high-contrast boundary, which is the most stable condition for sub-pixel edge finding. Coaxial or ring light produces specular highlights on plated pins, and the edge position then depends on the plating finish — the same part can read differently after a plating batch change.
The accuracy question is decided by the field of view, not by the camera count. A connector row that spans 40 mm cannot be measured in an 18 × 18 mm field without stitching, and stitching brings the length-dependent term into the result: ±(2+0.02L) on PRO and PRO-H, ±(1.5+0.02L) on MAX and MAX2. At L = 40 mm that term is 0.8 µm and 0.8 µm respectively on top of the base figure, which is usually acceptable — but it has to be stated, not assumed.
Model range, as stated in the IMAGE 3 series specification sheet. Repeatability here is the figure printed in the sheet; it is not a guarantee for your own part, your fixture or your operator. Treat it as the floor you start from when you run a trial, not as the number you can quote to a customer on day one.
| Model | Precision-mode range (X×Y×Z) | Wide-field range (X×Y×Z) | Lens field of view (precision / wide-field) | Repeatability (precision / wide-field) | Camera | Height measurement |
|---|---|---|---|---|---|---|
| IMAGE 3 | 18 × 18 mm (fixed platform) | Ø100 mm | 18 × 18 mm / Ø100 mm | ±0.5 µm / ±1 µm | Two 1" 20 MP B/W | Not available |
| IMAGE 3 PLUS | 118 × 118 × 75 mm | 200 × 200 × 75 mm | 18 × 18 mm / Ø100 mm | ±0.5 µm / ±1 µm | Two 1" 20 MP B/W | Not available |
| IMAGE 3 PRO | 218 × 118 × 75 mm | 300 × 200 × 75 mm | 18 × 18 mm / Ø100 mm | ±0.5 µm / ±1 µm | Two 1" 20 MP B/W | Not available |
| IMAGE 3 PRO-H | 218 × 118 × 75 mm | 300 × 200 × 75 mm | 18 × 18 mm / Ø100 mm | ±0.5 µm / ±1 µm | Two 1" 20 MP B/W | Not available |
| IMAGE 3 MAX | 205 × 104 × 75 mm | 223 × 118 × 75 mm | 6.4 × 4.8 mm / 18.8 × 14.1 mm | ±0.25 µm / ±0.5 µm | 1" 20 MP B/W | 8055 spectral confocal, WD 55 mm, range ±3 mm |
| IMAGE 3 MAX2 | 206 × 106 × 75 mm | 225 × 125 × 75 mm | 6 × 6 mm / 25 × 25 mm | ±0.25 µm / ±0.5 µm | 1" 20 MP B/W | 8055 spectral confocal (optional) |
The full specification sheet, including the camera and lighting options and the workpiece-to-model mapping, sits on the IMAGE 3 series product page. If you only want a first-pass sanity check on whether your part fits any model, the instant measurement selection tool takes the part envelope and returns the models that can cover it.
Coplanarity is the spread of the contact points along the seating axis. On the IMAGE 3 series, height measurement is only available on MAX and MAX2, through the 8055 spectral confocal sensor: working distance 55 mm, measuring range ±3 mm, repeatability ±2 µm, spot size 17 µm, linearity 0.02 %F.S. On MAX it is fitted as standard; on MAX2 it is optional.
Two consequences follow. First, coplanarity inspection is a point-scan operation, not a single exposure, so the cycle time scales with the number of contacts probed rather than with the number of parts in the field. Second, the 17 µm spot sets the smallest contact that can be probed cleanly — narrower contacts need a different sensor, and this is a common place where a connector project quietly changes scope.
If the drawing says coplanarity but the budget is a 2D system, the honest position is to define what can be guaranteed: in-plane position and pitch, with coplanarity declared out of scope. Writing that down prevents an acceptance argument later.
| Model | Precision mode, single field (µm) | Precision mode, stitched (µm) | Wide-field, single field (µm) | Wide-field, stitched (µm) |
|---|---|---|---|---|
| IMAGE 3 | — | — | — | — |
| IMAGE 3 PLUS | — | ±(2+0.02L) | — | ±(4+0.02L) |
| IMAGE 3 PRO | ±2 | ±(2+0.02L) | ±4 | ±(4+0.02L) |
| IMAGE 3 PRO-H | ±2 | ±(2+0.02L) | ±4 | ±(4+0.02L) |
| IMAGE 3 MAX | ±1 | ±(1.5+0.02L) | ±1.5 | ±(2+0.02L) |
| IMAGE 3 MAX2 | ±1 | ±(1.5+0.02L) | ±1.5 | ±(2+0.02L) |
L is the measured length in millimetres. The stitched column applies when the feature is longer than one field and the system has to join several images; the single-field column applies when the whole feature sits inside one shot. Read the sheet in this order: which mode, stitched or not, then the length — the three together decide the number.
A position tolerance without a stated datum is not a specification. In connector drawings the datum is usually the housing edge or two locating posts, and the optical system has to see that datum in the same image as the pins — otherwise the measurement is stitched across a datum that has its own uncertainty.
The practical check is whether the datum feature and the worst-case pin both fit in one field. If they do, position can be evaluated per shot with no stitching term. If they do not, either the fixture has to repeat the datum reference mechanically across shots, or the tolerance has to be evaluated on a stitched image with the length term declared.
Three cases. When the gating feature is plating thickness — that needs a coating gauge. When the part is a mated assembly whose critical dimension only exists after insertion — that needs a functional gauge. And when the contact pitch is below what the chosen field can resolve with a stable edge: at some pitch the silhouette merges and no amount of magnification recovers independent edges, because the limit is the optical resolution rather than the pixel count.
Pitch itself is a well-conditioned projection measurement with backlight, and the stated repeatability floor is ±0.25 µm in precision mode on IMAGE 3 MAX and MAX2. The limit is not the repeatability but the field of view: a row wider than 18 mm needs stitching, which adds a length-dependent term of ±(2+0.02L) or ±(1.5+0.02L) depending on the model.
Only with a height sensor. On the IMAGE 3 series that means IMAGE 3 MAX with the 8055 spectral confocal sensor fitted as standard, or MAX2 with it as an option: working distance 55 mm, range ±3 mm, repeatability ±2 µm, spot size 17 µm. A 2D backlit system shows silhouette change, not true seating-plane spread.
Backlight, because it produces a high-contrast boundary independent of the plating finish. Ring or coaxial light on plated contacts produces specular highlights whose position depends on surface finish, which turns a plating change into an apparent dimensional change.
No. An image measuring system evaluates geometry in the image plane and can flag appearance, but plating thickness requires a coating thickness gauge. Treating an optical channel as a plating measurement creates an acceptance claim the instrument cannot support.
A field at least as wide as the row plus margin, otherwise stitching applies. IMAGE 3 covers Ø100 mm in wide-field mode, PLUS covers 200 × 200 mm, and PRO and PRO-H cover 300 × 200 mm. The precision-mode field stays 18 × 18 mm on those models, so the mode used changes the accuracy term.
It is the most reliable option. Pockets that are noticeably larger than the part let the pose vary between loadings, which shows up as position scatter that looks like process variation. A loose tray is only acceptable when a locate step re-establishes the datum per part.
The fastest way to close this question is a real part, not a discussion. Send one sample with the drawing and the features you actually gate, and we will run it and report what the measurement returns — including the features that come back as a no. A short written feasibility note is more useful than a brochure.
Simply enter your email to receive the latest news and insights from Pomeas. Stay connected with Pomeas and be the first to discover new innovations in optical excellence.