BakeHQ

Cake sank in the middle

The cake looked good through the oven door and then dropped in the centre, either before it came out or shortly after.

Usually established at: baking

Narrow it down

Answer what you can. Anything you skip simply leaves its causes in play — there is no requirement to answer everything, and no answer is treated as more authoritative than what you actually saw.

How did the batter look before it went into the tin?
How old is the baking powder, and had it been tested?
When did it sink?
How high did it rise before it fell?
What does the cut surface look like?
Was the oven opened during the first two-thirds of the bake?
What was the centre like when you cut it?
How was the colour when the stated time was up?

All 7 possible causes are shown below. Answer a question above to start ruling them out.

Every possible cause

The complete model, whatever you answered above. Each cause states the mechanism — what physically happens — because a cause without a mechanism is folklore with a confident tone.

The structure set too late

Very common

A cake rises while its structure is still soft and sets when egg proteins coagulate and starch gelatinises. If the oven is too cool, or the tin too deep for the temperature, expansion peaks well before setting and the risen batter has nothing to hold it. The centre falls, usually after the cake looks nearly done.

What to change

  • Bake hotter, or lower the shelf

    Increase the oven temperature modestly, or move the tin to a lower shelf so the base receives more heat. Both bring protein coagulation and starch gelatinisation forward relative to the batter’s expansion.

    Cakes that rise well and then sink in the middle. Not appropriate where the cake is already browning too fast at the edges, which indicates the opposite problem.

    Trade-off: Too hot sets the outside before the centre has risen, producing a domed and cracked top.

Too much chemical leavener

Common

A chemical leavener expands existing gas cells. Too much expands them past the point the batter’s structure can support, producing large coalescing cells with thin walls. The cake rises dramatically, the cell walls rupture, and it sinks. Excess bicarbonate of soda additionally leaves an unreacted alkaline residue with a soapy taste.

What to change

  • Reduce the chemical leavener

    Reduce the baking powder or bicarbonate of soda modestly and check that the recipe’s acid balance still works. If the recipe uses bicarbonate alone, confirm there is an acidic ingredient present for it to react with.

    Cakes that rise dramatically and collapse, show large coalescing holes, or carry a soapy aftertaste.

    Trade-off: Reducing too far gives a dense cake, which is the fault in the opposite direction.

The oven was opened before the structure set

Occasional

Opening the door drops the oven temperature sharply. If this happens while the cake is still expanding and has not yet set, the delay in reaching setting temperature can be enough for the structure to fall. Once the cake has set, opening the door is essentially harmless — which is why this cause is far more specific than the folk version of it.

What to change

  • Leave the door shut until the structure has set

    Do not open the oven during roughly the first two-thirds of the bake, while the cake is still expanding. Once it has risen and the surface has set, opening the door to check or to turn the tin does very little.

    Cakes and other batters that set late. Deliberately narrower than the folk version of this advice: the general rule leads people to avoid checking a cake at all, which causes its own faults.

    Trade-off: You cannot judge colour during the period that matters most, so an oven that runs hot will over-brown before you see it. Verifying the oven temperature is the better fix for that.

The oven runs cooler than its dial

Common

Domestic ovens are frequently some way from their dial setting, and the discrepancy is usually invisible because it affects every bake consistently. A cool oven delays protein coagulation and starch gelatinisation, extending the window in which a cake can collapse and leaving crusts pale and bases damp.

The variable underneath

  • Oven temperature

    How to tell after the fact: A bake that consistently needs longer than stated and comes out pale is the signature of an oven running cool. One that browns before the centre sets is running hot. Several unrelated recipes failing the same way is the strongest single indicator.

What to change

  • Verify the oven temperature with a thermometer

    Put an oven thermometer in the middle of the oven, preheat fully, and compare the reading with the dial. Note the offset and set the dial to compensate. Check again after twenty minutes, since some ovens overshoot before settling.

    Any oven whose bakes are consistently under- or over-coloured. Fixes an entire class of faults that otherwise look like recipe problems.

    Trade-off: Needs an inexpensive thermometer, and the offset can vary across the cavity.

The batter was overmixed

Very common

Mixing past the point of combination develops gluten. In a batter this produces a tough, elastic structure that traps gas in large elongated channels rather than fine cells — visible as vertical tunnels in the cut surface — and a chewy rather than tender texture. It can also knock out air created during creaming.

The variable underneath

  • Mixing extent

    How to tell after the fact: Vertical tunnels running up from the base of a muffin or cake, with a peaked rather than domed top and a chewy texture, is the clearest single-variable signature in the catalogue.

What to change

  • Stop mixing as soon as the flour is moistened

    Fold wet into dry until no dry pockets remain and stop, even if the batter looks lumpy and unfinished. Remaining dry patches hydrate during the rest and the bake.

    Muffins, quick breads and any batter mixed by the muffin method. Does not apply to creamed batters, where the creaming stage requires sustained beating.

The chemical leavener had lost its activity

Common

Baking powder contains an acid and a base kept apart by a dry starch buffer. Moisture from the air lets them react slowly in the tub, so an old or badly closed container has already spent part of its capacity before it reaches the mixture. The result is an evenly dense, flat bake with no tunnels and no large holes — the gas simply never arrived. Bicarbonate of soda alone does not degrade this way, but it depends on an acid in the formula, so the equivalent failure there is a missing or weakened acid rather than a spent base.

What to change

  • Test the baking powder in hot water

    Stir half a teaspoon of baking powder into a little HOT water. Vigorous immediate fizzing means it is active; a weak reaction means it has lost capacity. Use hot water rather than cold — a double-acting powder’s second acid is heat-activated, and cold water under-reports its activity. For bicarbonate of soda, test with a spoonful of vinegar instead, since it needs an acid rather than heat.

    Any chemical leavener older than a few months, opened a while ago, or stored somewhere humid.

    Trade-off: Two minutes, and a negligible amount of leavener.

The batter split before it went in

Common

A creamed batter is fat holding air with water dispersed through it. Adding cold liquid quickly drops the temperature of the fat and the mixture curdles, releasing the air that creaming introduced. A split batter bakes into a denser, greasier, coarser cake — and the split is visible in the bowl, which is why it is one of the few cake faults that can be caught before baking.

What to change

  • Bring the eggs and liquid to room temperature first

    Let eggs and any milk sit out before starting, and add the eggs a little at a time rather than all at once. If the mixture does begin to curdle, a spoonful of the measured flour beaten in usually brings it back. A split batter can be baked and will give a coarser, denser cake.

    Any creamed cake. Irrelevant to melted-fat and all-in-one methods, where there is no emulsion to break.

    Trade-off: Requires planning an hour ahead, which is the commonest reason it is skipped.

Preventing it next time

  • Check the oven’s real temperature with a thermometer, since a cool oven causes this fault and is invisible without one.
  • Measure chemical leaveners accurately, and reduce rather than increase them if the cake rises dramatically before falling.
  • Leave the door shut until the cake is set — not for the whole bake, but through the period where it is still rising.

About opening the oven door

The received version of this — that opening the door makes a cake sink — is too strong, and BakeHQ records it as a commonly repeated claim in that form.

What is defensible is narrower. Opening the door drops the cavity temperature sharply, and if that happens while the cake is still expanding and has not yet set, the delay can be enough for the structure to fall. Once the cake has set, opening the door does very little.

The distinction matters because the general version leads people to avoid checking a cake at all, which causes its own faults. The specific version tells you when it is safe to look.

Related

Where this happens

  • Chocolate cake

    Cocoa’s absorbency and the acid-leavening interaction give this product more ways to reach the fault than a plain sponge has.

  • Muffin

    Muffins reach it by a distinctive route — overmixing produces tunnels and a peaked top alongside the collapse.

  • Sponge cake

    The reference creamed cake, and the family the model was written from.

  • Banana bread

    Deep, wet and heavily leavened, which is the combination that sinks.

  • Carrot cake

    A wet batter that sets late is the classic carrot cake collapse.

  • Pound cake

    Ordinary structure-timing causes.

  • Fruit cake

    A heavy fruit load in a batter that set late.

  • Cupcake

    Over-leavened small volumes fall dramatically.

  • Red velvet cake

    Ordinary structure-timing causes.

  • Angel food cake

    A fat-free foam cooled the wrong way up compresses under its own weight.

Possible causes

  • The structure set too late

    The general mechanism behind most instances: expansion peaked before coagulation and gelatinisation could support it.

  • Too much chemical leavener

    Identifiable by a dramatic rise before the fall and large coalescing holes in the cut surface.

  • The oven was opened before the structure set

    Recorded in a narrower form than the folk version. The temperature drop can cause a collapse during the setting window specifically; after the cake has set it does very little.

  • The oven runs cooler than its dial

    Invisible without a thermometer because it affects every bake consistently, and it causes faults that look like recipe problems.

  • The batter was overmixed

    Distinguished by vertical tunnels and a chewy texture rather than by the large thin-walled holes excess leavening gives.

  • The chemical leavener had lost its activity

    Distinguishable from the other causes of a poor rise: a spent leavener gives an evenly dense cake with no tunnels and no large holes rather than one that rose and fell.

  • The batter split before it went in

    A split batter has lost the air creaming introduced, so it rises less and can collapse around a coarse, greasy crumb.

Covered in

If none of this matches what you saw, the model has a gap. Tell us — that is how this section gets better.

BakeHQ would like to record anonymous usage — which pages are read, whether searches return anything, whether a diagnosis reaches an answer. No personal data, no advertising profile, and nothing you type is ever sent. The site works identically either way.