Cyclic damp heat and temperature/humidity cycling: what they test, and what to measure

A steady-state damp heat test asks one question: given time, does moisture soak into this product and harm it? It is deliberately run so that water never condenses on the specimen.

Many products never see that in service. They see days and nights: warm and humid, then cool, then warm again. Each cycle draws moist air into the product and, on the way back up, condenses water on its surfaces. A seal that is fine at a constant condition can pump water in a little at a time. A coating that resists humidity can fail under a film of liquid water. Cyclic damp heat and temperature/humidity cycling tests find those failures.

This article covers what the cycling tests do, how they differ from each other, what a modern chamber does on every transfer, and what has to be measured for a cycling record to mean anything.

Why cycling finds what steady state misses

Two things happen in a cycle that never happen at a steady condition.

The specimen breathes. As the temperature falls, the air inside an enclosure, a connector or a gap contracts and draws in air from outside, and at high humidity that air is full of water. As the temperature rises, the air expands and some is pushed out again, but water that has condensed or been absorbed stays behind. Over many cycles, an enclosure that would take months to wet through by diffusion can collect water quickly.

Water condenses on the specimen. A specimen always lags the air. When the temperature rises at high humidity, the air reaches the new condition first, and every surface still colder than the air’s dew point collects water. The numbers are unforgiving. At 65 °C / 93 %RH the dew point is about 63.4 °C, so any surface more than about 1.6 °C below the air is wet. At 25 °C / 93 %RH the dew point is about 23.8 °C, and the margin is about 1.2 °C. In a steady-state test this condensation is a fault to avoid; in a cyclic test it is part of the point.

Two cycling tests

The IEC 60068-2 series defines two cyclic tests that are often confused with each other and with the steady-state ones:

• Cyclic damp heat (IEC 60068-2-30, Test Db) cycles between a lower and an upper temperature over a 24-hour cycle, keeping the humidity high throughout. It is the test for products that will see humidity combined with temperature changes, and the condensation on the rise is expected.
• Composite temperature/humidity cyclic (IEC 60068-2-38, Test Z/AD) uses the same humid cycling and adds a cold subcycle below freezing, with the humidity system switched off. Water that got into cracks and gaps during the humid part of the cycle freezes there and expands. It is aimed mainly at component-type specimens, where that mechanism can open a seal or a crack that humidity alone would not.

The product standard or customer specification decides which test, which severity and how many cycles. The lab’s job, as with any environmental test, is to show that the cycle was applied as specified.

Cascade control: the part and the air are different things

Fast cycling chambers are often controlled by cascade on the part. A sensor sits in or on the specimen, and the temperature loop controls that sensor, using the chamber air as its inner loop. To bring a lagging part to temperature quickly, the controller drives the air well past the setpoint and backs off as the part arrives.

That changes what has to be measured. On a cascade chamber the air is not supposed to sit at the profile’s setpoint during a transfer. In the ProcessView+ demonstration, the controller may drive the air as far as 75 °C on the rise to a 65 °C dwell, and as far as −20 °C on the way down to −10 °C. Judging the air against the part’s setpoint would flag a chamber doing exactly what it should. So ProcessView+ judges the two as separate channels: the part against its loop’s setpoint, and the air against the air setpoint the controller computes for it.

Humidity during a cycle

Two pieces of chamber behaviour shape how a cycling run is judged.

Humidity cannot track a fast transfer. A chamber that moves 40 °C in six minutes cannot hold 93 %RH all the way. As the air heats, its relative humidity falls until the humidifier catches up; as it cools, the chamber has to remove water to keep from saturating. So in the demonstration, humidity is judged at the dwells and recorded, not judged, through the transfers. The record says so.

In the cold subcycle, the humidity system is off. Below freezing a chamber does not control humidity, so on those steps there is no humidity band, and the cold dwell is left out of the humidity figures. The QA report prints why: humidity was not controlled for the whole step. Below freezing, the QA report also does not calculate a dew point or a condensation margin, because the formulation it uses is not valid there, and the record says so. The live Humidity Diagnostics window can show a frost point instead when the dew point falls below 0 °C, as long as the air itself is above freezing.

Programmed time, or wait for the chamber?

Controllers such as the Watlow F4T can hold a profile at the end of a ramp until the process arrives, often called guaranteed soak. Many cycling profiles run with it off, so that the cycle keeps to the clock the standard specifies. Then the profile moves on whether or not the part has arrived, and the record has to show whether it did.

The demonstration runs this way: on programmed times, like an F4T with guaranteed soak off. Each transfer is programmed to be at least as long as the chamber air needs, and each dwell has a settle window, during which the run is watched but not judged, so the part’s approach is kept out of the judged time. In the demonstration the window is six minutes on every dwell, three of the part’s time constants. One window governs every channel on the step.

What to measure

Was each humid condition held? The joint window measures the share of each humid dwell with temperature and humidity both in band at the same time, with the longest unbroken hold. Each held condition gets its own row in the condition table on the QA report cover. Our article on reading temperature and humidity together covers the joint window in detail.

Where would water condense? On a cascade chamber the dew point is taken from the chamber air, which is where the water is, and the condensation margin counts the part as one of the surfaces water can form on. Neither is calculated below freezing. Like every psychrometric figure in ProcessView+, they are measurements for the reviewer, not part of the verdict.

Did the part, the air and the humidity each stay in tolerance? Time in tolerance is reported per step and per channel: the part, the air and humidity, each against its own setpoint, with the time in warn and in alarm, the number of excursions and the longest one. The transfers are monitored but not judged.

Was the chamber being watched the whole time? A real cycling test runs for many cycles over days. Each logged reading is marked as observed or repeated during an outage, and the sample coverage is stated beside the verdict. An outage longer than a short threshold raises the verdict to REVIEW.

What was the verdict, and why? PASS, REVIEW or FAIL is decided by each step’s criticality and the worst excursion on it, and the reason is printed under the verdict. Mean kinetic temperature is a pharmaceutical storage figure, and the cycling demonstration’s report leaves it out.

Try it without a chamber

The ProcessView+ free trial includes the T/H Cycle Demo on a built-in controller simulator: a small potted module on a simulated 25-litre benchtop climatic chamber under cascade control. It dwells at 25 °C / 93 %RH, rises to a hot dwell at 65 °C / 93 %RH, returns, transfers to a cold dwell at −10 °C with the humidity system off, and returns to a final humid dwell. The chart draws the part, the chamber air and the air setpoint the controller computes, so you can watch the cascade drive the air past the setpoint and the part follow it. A run takes 109 minutes, or about 27 minutes with the demo clock at 4×.

It is a demonstration, not a conforming test. Its structure is based on IEC 60068-2-38 Test Z/AD, but it runs one compressed cycle, and the record states that it is not a conforming run. The chamber’s heating and cooling rates are representative of a class of benchtop chamber, not measurements of a particular one. Demonstration runs are simulated and are not test results, and none of them qualifies a product.