Product
Recordings
New
Watch every Build Better session on demand.
Pin inspection FAQ
How true position measurement works, how accurate it is, how it handles false calls, and what it takes to deploy on your line.
True position measurement checks how far each connector pin sits from where the drawing says it should be. Instrumental images the connector, locates every pin's X and Y coordinates against datums from the connector supplier's drawing, and calculates true position from them. It also keeps the raw X/Y vector, so engineers can see both the distance and the direction each pin moved.
Instrumental's pin inspection detects bent, splayed, missing and misaligned pins in high-density connectors. It also catches shifts between pin groups, debris and foreign objects in pin windows, and damage to the connector housing or ground plane. It can flag pins that are non-parallel or off-center even when they still pass the true position spec, which helps teams catch weakened pins before they fail.
High-density connector pins are thin and fragile. They can move or bend at many points between the time they come from a supplier and are added to a server rack. Common causes include press-fit forces, board thickness, handling, packaging and shipping, debris, and contact from test probes. Because one connector can carry hundreds of pins, a single bent pin can cause a test failure or a latent signal integrity problem downstream.
Instrumental inspects high-density connectors used in AI servers and racks, including midplane, backplane, mezzanine, edge, and blind-mate connectors. It covers both mating pins and press-fit pins, on loose connectors or on assembled boards. Inspection can run at incoming quality control, after press-fit, around electrical test, at final assembly, and in failure analysis.
Camera-based inspection handles reflective, gold-plated pins that laser measurement often struggles with. Laser systems can also miss gross damage such as a crushed housing or a damaged ground plane. Instrumental pairs telecentric optics for undistorted measurement with AI that flags any visual anomaly, so one image delivers both a true position value and a damage check.
Automated pin inspection measures every pin on every unit, while manual inspection relies on an operator spotting a few bent pins among hundreds. Instrumental records an image and measurement for each pin, tied to the unit's serial number. That record lets engineers prove whether a pin was straight when it left a station.
True position measurement is a numeric check of each pin's location against the drawing spec. Anomaly detection is AI that flags anything that looks different from normal, such as debris, housing damage or an unusual pin shape. Instrumental reports them separately: measurement usually drives the go/no-go decision, while anomaly detection catches defects no spec describes.
Traditional AOI needs hand-written rules and often a golden unit for every defect it checks. Instrumental's AI learns what normal looks like from production images and is pretrained on an extensive real-world dataset, so it can catch new, unexpected defects during new product introduction. Its measurements and images also connect to test and process data, so teams can trace a pin problem to its root cause.
In a cross-site Gage R&R study, Instrumental's pin positions at two factories differed by at most 35 µm, with 13.2 µm RMS, against a 400 µm true position spec. Repeatability was measured by removing and reloading the unit for every capture. Identical images processed by the same software return the same result every time.
Instrumental validates pin measurement in three steps. First, verify the datum structure against the connector supplier's official drawing. Second, run a Gage R&R that reloads each unit between captures. Third, correlate pin-by-pin X/Y data against the supplier's CMM or factory data. Calibration uses a ceramic checkerboard target with about 10 µm accuracy to correct lens distortion and set scale.
Yes. Instrumental shares the same true position measurement logic across sites and compares raw X/Y data from each station to confirm they agree. When results drift, the usual causes are a software version mismatch or differences in fixtures that tilt the board. Cross-site Gage R&R on the same units finds and quantifies the gap.
Usually because pins moved between sites. Instrumental's serialized images compare the same pin at each station, which shows whether it was straight when it left and bent on arrival. Handling, shipping and undersized dust caps are common causes. If the images match but the readings don't, the gap points to a difference between the two inspection setups.
Not when pin inspection is set up the way Instrumental deploys it. On a connector with 800+ pins, even a 0.5% per-pin false call rate would flag nearly every unit, so measurement and anomaly detection are separated. True position measurement drives go/no-go against your spec, and an operator reviews flagged images to confirm real bent pins before a unit is rejected.
No. Instrumental sets the go/no-go limit to your true position spec, so a reading just inside tolerance passes. Engineers still see every measurement in the web app for trend analysis and failure analysis. Lines can also add a buffer station so operator review happens without slowing the line rate.
Yes. Limits follow the spec you choose, and different stages can use different limits, such as a tighter supplier outgoing spec and a looser incoming spec. Instrumental's images and test data help teams set limits based on evidence, such as how far a pin can bend before it causes a mating failure.
No. Instrumental's pin inspection doesn't need a golden reference unit. Its unsupervised AI begins flagging bent pins, debris and other connector anomalies after imaging about five units. Targeted defect models can then be trained on roughly 30 units, so inspection can start during NPI, when good reference units are often unavailable. Instrumental's True Position Measurement (TPM) measures against the product specification drawing, so it needs no golden unit either.
A typical Instrumental deployment takes about four weeks from a signed statement of work, followed by up to two weeks of tuning on 30 to 50 production units. Timelines depend on the station design and on access to connector drawings and CAD. Custom fixtures for tight spaces, such as midplanes inside a chassis, can add lead time for optics.
Yes. One Instrumental station can inspect several connector types by swapping the nest that holds the part, while the optics stay fixed. New nests can be designed from the connector's CAD. If connectors arrive at the same time on a high-volume line, separate stations may be faster.
Yes. Instrumental inspects both sides by rotating or flipping the part between captures, or by using top-down and side-view cameras together. The right setup depends on available space, lens size and cycle time targets.
Yes. Instrumental's platform is hardware-agnostic. It can analyze images from its own stations or from existing equipment and fixtures, received over SFTP or an API. Precise true position measurement needs undistorted images, so telecentric optics are recommended for the measurement itself.
Yes. When connectors and units are serialized, Instrumental links pin images and measurements from every station into one record per serial number. That chain of custody can span supplier outgoing inspection, incoming quality control, assembly, test and returns. It shows exactly where along the way a pin moved.
Yes. Instrumental returns pass/fail results to the line through an API or PLC signals and can integrate with MES. Pin images and measurements can also be correlated with electrical test data such as TDR and ICT, so engineers can link a visual defect to a functional failure and find root cause faster.
Yes. Instrumental exports measurements by serial number as CSV, including raw X/Y coordinates and true position for every pin and every inspection. Teams use the exports for their own analysis, Gage R&R studies and correlation with supplier data.
Inspect pins wherever they can move and before they get covered up. Instrumental customers commonly inspect at incoming quality control, after press-fit, before and after electrical test, before a cover or module is installed, and at final assembly. Inspecting at several points pinpoints the station where damage happens.
Test probes can bend or damage connector pins. Instrumental compares images taken before and after test to detect relative pin movement, even when the pin is still within spec. That evidence shows the test is safe to run on every unit, and it catches damage caused by the tester itself.
Yes. Some teams run Instrumental's pin inspection periodically, such as once per shift, or on demand after a test failure. This catches bent pins in racks and backplanes before they cause repeat failures, without adding a step to every production cycle.
Our engineers can walk through our methodology and line setup.
Talk to an expert