Feel the floor

Method

How this is modelled

These floors vibrate by hundredths to tenths of a millimetre, from under a hair's width to several sheets of paper. We draw it hundreds to thousands of times larger so you can see it.

The floor and the cup are drawn on one log scale for every floor, so small motion is magnified more: on this screen hundreds to thousands of times. The note under the stage gives the factor at the cup’s largest movement in the walk. Motion is shown 4× slower; every number stays in real seconds. True size draws the floor at its real movement: you’d feel it, not see it.

Draw a path on the floor, then press Walk, or drag the cup somewhere else. Turn the view by dragging around the floor with the mouse, or with two fingers on a touch screen; double-click to bring the default view back.

Vibration, not strength, often sets how far a CLT floor can span [1]. Bare CLT is lightly damped (about 1% in lab tests), so each footstep rings on [2].

These simplifications move the numbers a little, not which floor moves most, how long each rings, or that long spans tune in.

• Stiffness. Each floor is treated as one layered timber plate. Its stiffness comes from the way the layers are stacked, with a correction for shear [15][18]. Each floor is one solid panel; seams and board gaps are not modelled.
• Ways to move. We find the shapes the floor likes to move in, and how fast each one shakes [14]. Two walls carry the floor; the other two edges are free.
• Motion. Each shape is moved along through the walk and then left to ring down. The cup feels their sum [13]. Every shape gets the same damping.
• Walker. A typical walker: 700 N at 1.85 steps a second, using published footstep and walking-force models [1][6][10]. Our walker dances; the floor feels a normal walk (0.7 kN at 1.85 steps a second). A typical walker at a fixed pace, not a design load. CCIP-016 uses a firmer footstep (54 rather than 42) [11].
• Damping. Typical values from the literature: bare CLT about 1% in lab tests, screed 2% from a design table, joists 3% [2][5]. “What if 1%?” uses the EN 1995-1-1:2004 default [8]. Damping is the least certain input, and it sets how long a floor rings.
• Felt meter. It compares the cup’s motion with the ISO 10137 base curve of human perception [9], as given by WoodWorks [1]. At 1, most people start to notice. It reads the strongest moment of the whole walk, with every footstep and every mode counted, so it sits a little above a design guide’s single-footstep check: on WoodWorks’ worked CLT floor it reads about 12 to 15 here, against 11 in the guide [1]. It shows when most people notice movement. It is not a design limit.
• Span band. The CLT Handbook’s vibration-controlled span for the bare panel [2].
• Screed. The screed counts as weight only: 100 kg/m², its dry weight in service [16].
• Drawing. Motion is shown 4× slower. Ripples on the floor and in the cup are drawn on one log scale for every floor. The cup ripples at min(1, log₁₀(1 + x) / 3) of its full height, where x is the cup’s motion over √2 times the base curve: x = 1 is where “Still moving” stops, and 999 is full height. The floor is drawn at sign(w) · min(1, log₁₀(1 + |w| / 10 µm) / log₁₀ 101) of its full height. In the default view, which Reset brings back, at full scale the floor moves 34 px on a laptop screen and 24 px on a phone, the same for every floor and span, so 10 µm to 1 mm fills it; the note under the stage gives the exaggeration on your screen, which stays the same as you turn the view or zoom it (pinch, or scroll with ctrl or ⌘ held). The boards, layers and joists are drawn for show, as in the Awatif CLT app; the model treats each floor as one plate. The walker is drawn for show, and the cup is not to scale.
• Real size. The factor under the stage is the drawn height over the real height at the cup’s largest movement in the walk. The scale is logarithmic: small movements are magnified more, so the factor changes from floor to floor, and a floor that moves 40 times more is drawn only a few times higher. The number is the vibration at the cup; the floor’s sag under the walker’s weight is left out. True size draws the floor and the walker at their real movement, too small to see; the cup’s rings take the cup’s real vertical movement, drawn larger with the mug, which is not to scale. A hair is about 70 µm across [19] and a sheet of 80 g/m² office paper about 0.1 mm thick, at about 800 kg/m³ [20].

Illustrative, not a design check. Motion shown 4× slower and drawn on one log scale for every floor. The cup is not to scale. Our walker dances; the floor feels a normal walk (0.7 kN at 1.85 steps a second).

Sources