Cold Test Chamber Temperature Range And Tolerance Per IEC 60068 2 1

The preferred temperatures

The standard lists a ladder of preferred low temperatures so labs around the world test to the same numbers instead of inventing their own. The common steps run +5, -5, -10, -25, -40, -55, and -65 degrees, with a few finer points spaced between them. A product specification names one of these, and the chamber has to settle there and stay.

Lower is harder.

Every step down the ladder asks more of the refrigeration. A box that reaches -40 with a single-stage system needs a cascade pair to touch -65, and the jump from -55 to -65 often doubles the cost and the complexity of the cooling side. The temperature a specification calls for, then, shapes the whole machine before a single sample goes in.

Down is expensive

Every step colder asks more of the refrigeration. The jump from minus forty to minus sixty-five can double the cost of the cooling side.

Cold finds the brittle

Cold rarely melts anything. It stiffens, shrinks, and embrittles, and the failures it brings out are cracks and leaks rather than burns.

What tolerance means here

Reaching a temperature is half the job. Holding it is the other half. The standard sets a tolerance band around the target, commonly plus or minus 3 kelvin for a cold test, and the chamber has to keep the specimen inside that band for the length of the soak.

Two numbers sit behind that band. Fluctuation is how much the reading wanders over time at one point once the box has settled. Gradient, or uniformity, is how far the temperature varies from one corner of the working space to another at the same moment. A chamber can hold a tight fluctuation at its sensor and still miss the gradient if a far corner runs warm, so a careful test measures both.

Inside the working space

The tolerance applies to the working space, the central volume where specimens sit, rather than to every cubic centimetre behind the door. The standard pulls the measuring points clear of the walls, the floor, the fan, and the air inlets, since those spots always read hotter or colder than the zone a product occupies. A chamber rated for a tight gradient earns that rating only across the defined working space, and a lab that overloads the box or crowds the walls gives that rating away.

Still air and the heat-dissipating case

IEC 60068-2-1 splits its methods by whether the specimen makes its own heat. A part that sits dead, drawing no power, takes the still-air methods, Test Ab for a gradual change and Test Aa for a sudden one. A part that runs and warms itself, a powered board or a small motor, takes the heat-dissipating methods, where the standard watches the temperature of the air around the specimen rather than a single chamber number.

The distinction changes the wiring.

A dead specimen needs only a tray and a thermocouple. A powered one needs a feed-through for its cables, a way to switch it on inside the cold, and extra sensor channels so the lab logs both the chamber air and the specimen's own surface as the run goes on.

Rate of change and the soak clock

The standard cares about how fast the box cools as well as where it lands.

A gradual cold test ramps the specimen down slowly so thermal shock does not cloud the result, while a sudden one drops it in fast to probe exactly that shock. Once the specimen reaches temperature, the soak clock starts, and the dwell, often two, sixteen, or seventy-two hours, gets counted from the moment the specimen itself settles, not from the moment the air hits target.

That gap matters for a thick part. A heavy casting lags the air by a wide margin, so the lab waits for the part's core to reach temperature before it starts the count, and a thin foil reaches it almost at once.

Frost, condensation, and the dry side

A cold chamber has a moisture problem even though the cold test itself does not control humidity. Any water vapour in the air freezes onto the coldest surfaces, so the cooling coil and the specimen both gather frost. Frost on the coil chokes the heat exchange. The chamber answers with a defrost cycle or a dry-air purge to keep working through a long soak.

Dry air is part of the spec.

Many cold chambers feed a trickle of dried air into the space during a low-temperature run, holding the dew point below the set temperature so nothing frosts where it should not. That dry purge keeps ice off the specimen as well, so the part under test meets cold rather than a coat of frost that would change what it feels.

Where the sensors sit

A cold chamber carries two kinds of sensor, and the difference decides whether it meets the band. The control sensor sits in the air stream and tells the refrigeration when to cut in. The measuring sensors sit in the working space, near or on the specimen, and report what the part went through.

A reading taken only at the control point flatters the box. A careful cold test maps the working space with a grid of probes, often nine for a small chamber and more for a large one, and checks that every probe stays inside the band. The lab calibrates those probes against a reference before a qualification run, since a half-kelvin sensor error swallows half the margin the standard allows at a tight tolerance.

Approaching the target

Hitting -55 is easy. Stopping there without first dipping to -60 takes a careful control loop. A box that slams its compressor on and off undershoots the target, then drifts back up, and the reading hunts around the band instead of resting in it. A cold chamber tuned for tolerance eases toward the set temperature, trims the cooling in small steps near the target, and holds a flat line once it settles.

Sizing the refrigeration to the band

The tolerance band and the lowest temperature together set the cooling design. A wide band at a mild temperature suits a simple single-stage system. A tight band at -65 needs a cascade, heavy insulation, and a loop that can trim the cooling in fine steps so the box neither overshoots nor hunts.

Heat leak fights the band. Every cable feed-through, every door seal, and every viewing window leaks a little warmth into a box sixty or more degrees below the room. A cold chamber built for a tight tolerance watches those leaks closely, since each one shows up as a warm spot that widens the gradient and eats into the band the standard allows.

Conditioning before the cold

A cold test does not start at the cold. The standard sets a conditioning period at room temperature first, so every specimen begins from the same known state. Skip that step and a part that arrived warm from a previous test reads differently from one that sat overnight, and the comparison between samples falls apart. Stable starting conditions are part of why two labs running the same standard land on the same answer.

Reaching past the ladder

Some products call for colder than the preferred ladder reaches. Space hardware, high-altitude avionics, and certain research samples ask for -70, -80, or lower, past the comfortable limit of a two-stage cascade. A third refrigeration stage or a liquid-nitrogen boost takes the box deeper, trading running cost and complexity for the extra reach.

The control problem grows with the depth. A tiny heat leak that a -40 box shrugs off becomes a visible warm spot at -80, the insulation has to grow thicker, and the tolerance band gets harder to hold the further down the run goes.

What the cold brings out

The point of the test is the damage cold draws out, and the failures cluster around a few mechanisms. Plastics turn brittle and crack under a load they carry fine at room temperature. Elastomer seals shrink and stiffen, so a gasket that held pressure warm starts to leak in the cold. Greases and oils thicken until a mechanism that moved freely begins to bind. Different materials in one assembly contract by different amounts, and that mismatch pulls on solder joints, adhesive bonds, and press fits until something lets go.

Power-up adds its own stress.

A part switched on at the bottom of a cold soak meets a second shock as its own heat fights the surrounding chill, and the steep gradient across a board can split a joint that survived the steady soak untouched. A cold test that powers the specimen catches that failure, while a passive soak walks straight past it.

The record behind the result

A cold qualification lives or dies on its record. The standard expects the report to name the temperature, the tolerance achieved, the soak time, the rate of change, and the state of the specimen during and after the run. A bare pass or fail means little without those numbers, since a part that scraped through at a loose tolerance has not earned what a part that held a tight one did. The chamber logs feed that record, and a gap or a wandering excursion buried in the trace can put the whole soak in question long after the part has left the box. That is the quiet reason the measuring chain matters as much as the refrigeration. A probe reads only as accurately as its last calibration, and a sensor that has drifted half a degree turns a passing run into a fiction that cannot be defended. Labs that take cold work seriously calibrate against a traceable standard on a fixed schedule, place enough probes to map the whole space rather than trusting one, and keep the certificates with the data, so the numbers in the report can be stood behind years later when a field failure sends someone back to the file.

The door and the viewport

The weakest points of a cold box are the moving and transparent ones. A door seal that flexes on every cycle hardens in the cold, loses the springiness that let it follow the frame, and eventually lets warm room air seep in, frosting the gap and dragging the nearby corner out of band. The leak rarely shows on the control sensor in the centre of the space, so it hides until a mapping run finds the cold corner that will not come up to setpoint. A single-pane viewport behaves no better: it fogs, then ices over from the inside, so a watcher sees nothing, and the bare glass bleeds heat straight to the room and plants a warm spot on the wall beside it. Cold chambers built for a tight tolerance answer both with heated multi-pane windows and seals rated for the lowest temperature in the range, often a silicone or a fluoroelastomer that stays flexible far below freezing. The aim is a box that keeps its band through thousands of door openings rather than only on the day it left the factory, since the seal a lab trusts in year five matters more than the one it bought.

What it takes to reach the floor

Reaching the bottom of the preferred ladder is a refrigeration problem that grows harder with every degree. A single-stage system can hold minus forty without much strain, but minus sixty-five sits far below what one refrigerant pulls down to in a single go, so the box stacks two circuits in a cascade, the first chilling the condenser of the second, to step the temperature down in two leaps. The deeper the target, the heavier the insulation, the more careful the refrigerant choice, and the more a small heat leak matters, because a watt of warmth that a minus-twenty box would never notice becomes a visible warm spot on a shelf at minus sixty-five.

Holding the band at the floor is its own fight.

Every cable feed-through, every door seal, and every viewing window leaks a little heat into a box sixty or more degrees below the room, and each leak shows up as a corner that runs warm and widens the gradient the standard allows. A cold chamber built for the deep end watches those leaks closely, eases toward the setpoint in fine steps so it does not overshoot and then hunt, and proves through mapping that the whole working space rather than the control sensor alone sits inside the tolerance before a single specimen goes in.

Pulling it together

A cold test reads simple from the outside: make it cold, keep it cold, see what breaks. IEC 60068-2-1 turns that into numbers a chamber has to meet, a temperature off the preferred ladder, a band of a few kelvin, a uniform working space, and a soak counted from the specimen rather than the air. A box that reaches the low end of the range and holds the band across the whole working space, frost and heat leak and all, is one a lab can trust to give the same answer on a Monday as it gives on a Friday.