The rule, in one line

A surface collects water when it is colder than the dew point of the air touching it. Not colder than the air — colder than the dew point. That distinction is the entire subject, and it is why a glass of iced water sweats on a mild day while a window on a freezing one stays clear.

The two levers

Once you see the rule as a comparison between two temperatures, it follows that there are exactly two ways to stop condensation, and every remedy anyone has ever suggested is one of them:

  1. Raise the surface temperature above the dew point. Better glazing, insulation, moving air across the surface, or simply waiting for the sun.
  2. Lower the dew point of the air touching it, which means having less water in that air.

Nothing else does anything. Wiping the glass removes today's water and changes neither temperature, which is why it comes back. This is also the honest reason the "how to stop condensation" advice varies so wildly in quality: some of it moves one of these two numbers, and some of it does not.

What "margin" means here

The figure at the top is the gap between your surface and the dew point.

  • Positive — the surface is that many degrees above the dew point, and dry. The larger the number, the more the conditions would have to change before it is not.
  • Negative — the surface is below the dew point and already collecting water. The size tells you how aggressively: a surface one degree under gathers a faint mist, one ten degrees under runs.

The second readout, the air-to-dew-point spread, is about the air on its own. When that falls under about a degree the air is saturating everywhere at once, and you get fog rather than condensation.

Below freezing

When the dew point is under 0°C the vapour goes directly to ice and what you get is frost rather than dew. The threshold temperature is the same number from the same calculation — the calculator flags it so the wording on the page matches what you will actually see on the surface.

Strictly, saturation vapour pressure over ice differs slightly from over water, so a true frost point is a fraction of a degree different from the dew point computed here. The difference is small enough to be swamped by the accuracy of any humidity reading you are likely to have.