Designing a pogo-pin test fixture that lasts

The first fifty boards get tested with a bench supply, a multimeter, and a programmer clipped onto whatever header happened to be convenient during bring-up. That works right up until the run size goes past what one person can hand-probe in a shift. Then someone goes looking for how to build a real fixture, and finds either a turnkey in-circuit test vendor with a quote sized for thousands of units, or a scattered handful of forum threads with mismatched pin part numbers and no explanation of why one choice mattered more than another.

A pogo-pin fixture for a few dozen to a few thousand boards a month is a mechanical design problem before it’s a test problem. Get the pins, the pitch, and the board layout wrong and the fixture will pass boards it shouldn’t and fail boards it shouldn’t, in ways that are maddening to diagnose because the fixture itself is the variable nobody suspects first.

Deciding what the fixture actually checks

“Test fixture” covers two different things, and mixing them up is the first mistake. In-circuit test (ICT) probes bare nodes — resistor values, shorts, opens, component presence — before firmware is even involved. Functional test (FCT) exercises the finished, programmed board against its actual interfaces: power it up, talk to it over its real communication bus, check the outputs it’s supposed to produce. Most small-batch hardware companies only need FCT. ICT earns its cost on high node-count boards with a real defect history to chase; on a first or second production run, the failure modes that matter are usually functional — a bad solder joint on a connector, a part placed backward, firmware that didn’t flash — and a bed-of-nails FCT fixture that powers the board and runs a test sequence over UART, SWD, or I2C catches those without the expense of probing every net.

That decision drives everything downstream: how many pins the fixture needs, how tightly they have to be placed, and how much the fixture itself costs to build and maintain.

Pins, tips, and pitch

A spring-loaded pin (a “pogo pin”) is a plunger riding a compressed spring inside a barrel, making contact by pressing straight down onto a pad. Vendors like Mill-Max, Ingun, and Everett Charles publish full mechanical drawings, and it’s worth pulling one before laying out test pads, because the pin picks the pad, not the other way around.

Two decisions matter more than the rest:

Pitch. The bed-of-nails standard is 100 mil (2.54 mm) center-to-center, and it’s the right default — off-the-shelf fixture plates, drill patterns, and pin sizes are built around it. Finer pitches (75 mil, 50 mil) exist for dense boards, but every step down means a thicker, more precisely drilled fixture plate and pins that sit closer to their neighbors, which raises both the fixture’s cost and the odds of intermittent contact between adjacent points. Route test pads to 100 mil whenever board space allows it; it’s the layout decision felt for the whole life of the fixture, not just the first build.

Tip style. The tip shape decides how the pin defeats surface contamination and how it treats the pad underneath it.

Tip style	Best against	Where it falls apart
Concave / cup	Round leads, vias, headers	Slides off a flat SMT pad
Crown / serrated	Flux residue, light oxide on flat pads	Can mark a gold-flashed test pad over many cycles
Spear / chisel	Oxidized or tin-lead finishes	Higher force needed, more scarring
Flat	Clean, gold-plated test pads	No bite through any contamination

A crown or serrated tip is the reasonable default for SMT test pads coming off a no-clean paste process, because it breaks through residue instead of riding on top of it. A flat tip only earns its place on pads that are consistently clean and gold-plated, which in practice means a dedicated test pad finish, not whatever the board’s main surface finish happens to be.

Holding the board still

The pins only work if the board is pressed onto them the same way every time. Three mechanisms cover nearly every small-run fixture:

A clamshell — a hinged top plate the operator swings down by hand — is the cheapest to build and the easiest to service, and it’s the right choice for prototype and low-to-medium volume runs. A vertical press riding on guide rods and linear bearings gives more force and tighter repeatability, at the cost of a more complex build; it earns its keep once volume is high enough that operator fatigue starts affecting contact quality. Vacuum actuation pulls the board down evenly across the whole pin field rather than at a few press points, which matters most on larger boards where a mechanical press can flex the board enough to lift pins on the far side from where the force is applied.

Whichever mechanism holds the board, alignment is a separate problem: at least two tooling pins, fixed to the fixture plate and mating with two tooling holes in the board, are what actually keep every pin registered to its pad. A fixture with good pins and a bad hold-down mechanism still fails randomly; a fixture with sloppy tooling-hole tolerance fails the same way for a completely different reason, and the two are easy to confuse from the outside.

Designing the board so the fixture can do its job

The fixture can only be as good as the pads it’s asked to hit, and that’s a layout decision made months before the fixture exists. Test pads in the 0.8–1.5 mm range give a pin room to land reliably; going below about 0.035 in (0.89 mm) starts trading fixture cost for a higher rate of false failures from marginal contact. Keep test points at least 0.1 in from the board edge, and 0.125 in if the fixture needs a gasket seal around the board for vacuum hold-down. Avoid routing a test point under a connector shroud, a tall electrolytic capacitor, or anything else that blocks straight-down pin travel — it’s an obvious rule until it’s the one net on the board that has no other place to go, which is usually the moment someone realizes test access should have been reviewed at layout — alongside the rest of the layout habits that only bite at assembly — not after the boards came back from the fab.

The other layout habit worth building early: two tooling holes, placed as far apart as the board allows, dedicated to fixture alignment and nothing else. Retrofitting them onto an existing design means finding board area that was never reserved for it.

Wear is the part that shows up after the invoice

Pogo pins are rated components, not indestructible ones. Vendor data sheets typically rate a spring-loaded pin between 100,000 and 1,000,000 cycles at its recommended mid-stroke compression, and that range is wide because it depends on tip style, plating, and how close to full travel the fixture actually compresses the pin on every cycle. A fixture running a few hundred boards a week won’t hit that ceiling for years; one running a few hundred a day will, and it’s worth knowing which situation applies before the fixture ships, not after intermittent failures start showing up on boards that test fine when re-run by hand — the classic symptom of a pin that’s lost spring tension or picked up enough flux residue to raise its contact resistance.

Budgeting for spare pins, and designing the fixture so a worn pin can be swapped without re-machining the plate, costs almost nothing at the design stage and saves a production stoppage later.

What to send us

If you’re past hand-probing and need a fixture for the next production run, send the board’s Gerbers or KiCad project, the test points you already have in mind, the interface the test sequence needs to talk over, and the volume you’re building against. From that we can tell you whether FCT alone covers it, what pitch and tip style fit your pad finish, and what the fixture itself should cost in build time.