EN
Quick Links
Last Updated:08/31/2026 Author: Zeon
The early morning yield report reviewed 80.2%. Twenty percent of the day's manufacturing had actually failed at the exact same action-- the ICT (in-circuit examination) terminal-- and the failures were spread, inconsistent, and maddeningly periodic. The quality supervisor did what quality supervisors do: quarantined the falling short boards, pulled the failure logs, and arranged a style review. The engineering team believed a limited curcuit, a bad component whole lot, or a format concern. They invested a day and a half chasing a ghost.
Right here's the 2nd hook: when the service technician re-ran the precise very same 20% of boards with the specific same fixture, 98% of them passed on the second effort. The boards were fine. The design was fine. The elements were fine. The problem was a pogo pin-- one of the spring-loaded probes in the examination component, in a socket approximately 30 millimeters frome the one that was falling short, that had actually created a bad get in touch with and was lying about it.
That's things regarding examination components: when they fall short, they do not fall short noisally. They fall short silently, in the one place no one looks-- because everyone presumes the machine that measures whatever has to itself be fine.
A pogo pin is a deceptively straightforward device: a plunger, a barrel, a spring, and a suggestion that presses versus a test factor on the board. When the component shuts, each pin presses to its over-travel and the idea attacks into the test factor surface. The high quality of the electrical call depends upon the suggestion literally appearing whatever is in between it and clean steel-- oxide, flux residue, dust, or the natural film left by an OSP finish.
The electric specification is ruthless. A healthy and balanced pogo pin get in touch with steps in the tens of milliohms-- 10 to 50 milliohms is common. A limited call-- idea worn, spring tired, surface polluted-- can drift to numerous milliohms or oscillate unpredictably. A component that was flawlessly calibrated at 10 milliohms per call ends up being a different device at 300 milliohms, and the difference is invisible unless you measure it.
Currently multiply that by the variety of get in touches with in a regular test program. A board with 200 examination factors means 200 simultaneous spring-loaded get in touches with, each one combating its own battle against wear, contamination, and fatigue. The failure of a single pin can remove a whole test-- and if that pin is in a circuit that determines something refined, the failure can look exactly like a faulty board.
Pogo pin failing is hardly ever dramatic. It's a sluggish buildup of little destructions:
Idea wear. The suggestion is the factor of contact-- a crown, spear, or sculpt layout that concentrates pressure to pierce surface area movies. Every touchdown abrades it a little. The tip that begins at 0.4 mm size finishes its life rounded, brightened, and incapable to attack via anything. Suppliers price pins for numerous thousands or countless cycles, however those ratings think tidy boards, proper over-travel, and regular upkeep. Real-world life is usually a portion of the rated number-- a pin rated at 1 million insertions commonly requires replacing prior to 200,000 in manufacturing problems.
Contamination. Change deposit from soldering, dust frome the production floor, oils from handling, and solder bits from surrounding procedures all build up on the idea. Each deposit elevates get in touch with resistance. This is why fixtures on high-flux, no-clean production line break down faster than those on completely cleaned boards-- the pins are literally picking up residue every cycle.
Springtime exhaustion. The springtime is the pin's engine, and it's the component most conscious abuse. Exceeding the ranked over-travel-- the additional compression past very first get in touch with-- fatigues the springtime and minimizes the force it can use. A pin with inadequate force can't appear surface area films, so contact resistance climbs up. Fixtures that are slammed as opposed to shut, or boards that sit slightly high, eliminate springs quietly.
Misalignment. The pin needs to land within the examination point area-- typically a pad 0.8 to 1.2 mm in size. Drift of even 0.1 mm in fixture enrollment, tooling pin wear, or board growth can put the idea on the edge of the pad, where get in touch with area is tiny and resistance is unstable. The pointer might arrive on solder mask rather than copper-- which checks out as an open circuit every time.
Board warpage. The fixture thinks a flat board. A board distorted by also a fraction of a millimeter raises some test directs off their pins while over-compressing others. The affected nets review as opens or shorts. Aboard with uneven copper circulation-- the kind that warp a little after reflow-- the fixture can generate a consistent pattern of failures that looks like a layout flaw.
The frustrating part of pogo pin failings is that retesting "repairs" them. The exact same boards pass the second time, and that truth persuades everyone the problem was transient-- a glitch, a taking care of issue, perhaps the operator. The system is physical and absolutely mundane:
- The very first goal of the pin idea can mechanically pierce or displace the oxide/contamination movie it fell short to appear the very first time. The 2nd touchdown lands on the freshly revealed metal. Call improves.
- The act of putting and clamping the board a 2nd time can seat it in different ways, changing which pins are in contact and with just how much pressure.
- A pin that was minimal-- polluted suggestion, weary springtime-- can behave differently cycle to cycle, passing intermittently and stopping working periodically.
The retest pass price is really a diagnostic hint: when a high percentage of failing boards hand down retest, the first thing to suspect is the component, not the boards. High retest pass prices are a red flag for call concerns, not a confidence that every little thing is great.
The price of a false failing isn't just the retest time. It's the equipment of uncertainty that surrounds it:
Quarantined stock. Every failing board sits in a hold location while the "issue" is checked out. On a high-volume line, that's thousands of boards per hour waiting on a ghost.
Engineering time melted. Design reviews, schematic checks, and element evaluation all get invested in boards that were never defective. The rela problem-- a worn pin-- is invisible to every one of those examinations.
Wasted rework. Boards that fall short, obtain flagged, and get "reworked" (typically a basic retest that passes) carry the price and the stigma of rework without the benefit. Some percent of false-failure boards obtain junked or repaired unnecessarily, and every touch-up is a real flaw threat on an excellent board.
Distorted return data. If the component is degrading slowly, return drifts down week over week, and no one can clarify why. The production team starts "repairing" the procedure-- tweaking profiles, changing paste, changing positioning-- to compensate for an issue that exists completely inside the examination component.
Fortunately: pogo pin troubles are mechanical, and mechanical issues are avoidable. The self-control is monotonous, and it works:
Track call resistance, not simply pass/fail. An examination program that logs per-net resistance-- or a minimum of flags nets whose resistance has actually slipped over a threshold-- turns a concealed deterioration into a noticeable fad. When the same web shows rising contact resistance throughout a week of production, the fixture is informing you it requires solution before it begins failing boards.
Make use of a golden board. Maintain a known-good board, evaluated and doccumented, and run it with the fixture at the start of every change. If the golden board begins stopping working-- or passing with abject margins-- the component is suspect neccessarily. This single practice would certainly have captured every pogo pin failure I've ever before seen, days prior to it struck the yield record.
Maintain the pins on a routine, not on a failure. Tidy ideas at a defined period-- the interval relies on your flux and board sanitation, however every 10,000 to 20,000 cycles is an usual starting factor for high-throughput lines. Change pins before they're worn, not after. A pin ranked at a million cycles that obtains changed at 150,000 due to the fact that it's on the schedule costs a few bucks. A pin that stops working at 180,000 costs a day of manufacturing.
Design test points to be probed. A test factor that's a bare via, a pad under a component, or a pad with an OSP finish is an examination factor that battles the probe. Dedicated probe pads with a surface appropriate for repeated contact-- and sized for the pin idea-- expand pin life and support call resistance. This is a design-for-test (DFT) choice, made at layout time, and it settles in fixture dependability for the life of the item.
See the retest price. If your retest pass rate is high-- over a few percent of failing boards-- treat it as a fixture caution, not an enigma. The ratio of "stopped working on initial test, handed down retest" to "fallen short on both" is just one of the most effective very early indications of contact degradation in the whole examination process.
The awkward truth is that an examination fixture is an using mechanical assembly, not a tool of absolute fact. It has springs, tips, and get in touch with surfaces that weaken on a timetable nobody jotted down. When produce decreases and the boards are fine, the component is the starting point to look-- not the last.
The 20% of boards that fell short that early morning? They returned through the line, passed, and delivered. The actual failing was never in the boards. It remained in a solitary used probe idea, 30 millimeters from where anybody was looking, quietly lying to a maker that was expected to tell the truth.
Article written by Zeon
Hi, I'm Zeon — 20 years in PCB and electronics manufacturing. Front-end design and R&D, components sourcing, precision SMT, DIP through-hole assembly, and complete unit assembly. That's the full path from concept to finished product, and it's the path I've walked for two decades.
Hot News2026-09-17
2026-09-09
2026-09-07
2026-09-04
2026-09-03
2026-09-01
2026-08-31
2026-08-27
