How I Judge Baked Custard Doneness

A baked custard has a narrow useful finish: enough egg protein has set to hold a smooth spoonful, but not so much that the proteins tighten, expel water and become grainy. I read several cues together rather than trusting minutes alone.

I look for a set perimeter and a small controlled tremble in the centre. The middle should move as one soft mass, not ripple like liquid. If the whole ramekin is rigid in the oven, it will usually be firmer after cooling than I want.

Shallow ramekins show a smaller gradient than deep ones, so geometry changes both time and the meaning of the visible movement.

The ceramic, water bath and outer custard remain hot after baking. Heat continues moving inward while the ramekin is removed and begins cooling. I stop before the centre reaches its final chilled firmness.

I remove ramekins from the hot bath promptly. Leaving them in the water extends the bake even after the oven is switched off.

For crème brûlée I may use a centre reading around 78–82°C / 172–180°F as a working reference, but probe position and thermometer accuracy matter. A reading does not replace the movement and surface cues.

Different formulas, depths and desired textures require their own validation. I do not turn one number into a universal safety or quality guarantee.

Excessive heat makes egg proteins aggregate too tightly. The smooth gel becomes grainy and may release watery liquid. I check actual oven temperature, ramekin depth, bath level and baking time before altering yolk quantity.

Direct hot spots can overcook the edge even when the centre is correct, which is why the bain-marie and suitable ramekins matter.

If it remains fluid after complete chilling, it was underbaked or the formula lacks sufficient egg structure. I first confirm that the centre was genuinely set and that the ingredient weights were correct.

A warm custard is naturally softer than a cold one, so I do not diagnose the final texture before the defined chill.