Temperature Cycling of the Electronic PCB Assemblies

Google Cyclic Humidity test.

What does the functional product specification write about humidity??

Usually, it says something like "Maximum 90% relative humidity, non-condensing". The non-condensing bit is important and means moisture does not condense inside the equipment (and then freeze, since you mention -20 C). In the usual case, equipment is in a cabinet with enough self-heating or anti-condensation heaters to prevent condensation (research dew point) and enough ventilation to keep the air similar to outside air. In the extreme case, if you totally sealed the equipment at high temperature and humidity and then cooled it below freezing, you would certainly get ice.

Thanks--so we must have the chamber which is noncondensing type over the full range.

Unless the customer requires that you test function on every unit over the full temperature range, I would not expect to cycle to such a low temperature.

If your object is to reduce "field failures" by cycling every equipment made, it will not do to "test to destruction".

Since it is obviously expensive to provide coolers and de-humidifiers, it is usual to cycle only from factory ambient to maximum specified functional temperature for this purpose. If you just raise the temperature, relative humidity is reduced and condensation is not possible.

Note also, since it is usually considered that failure rate doubles for every 10 Celsius rise in operating temperature, it is the hot part of the cycle (and heating/cooling) which will find defective parts.

If you actually meant a "Type Test" on the equipment, then, of course, you would need cooling & dehumidifying equipment - but just for a small test chamber, not for a whole room full of normal production units.