Speed is a scalar (how fast); velocity is a vector (speed plus direction). Drag the sliders below to set an object circling at a constant speed and radius, and watch the velocity vs speed distinction play out live: the velocity keeps changing direction even as the speed holds perfectly still.
Circular Motion
An object circles at a constant speed. Its speed never changes — yet its velocity changes every instant, because velocity has direction.
Watch the gold arrow: its length (the speed) stays fixed while its direction (the velocity) keeps turning.

The velocity vs speed simulator is a free interactive physics lab that runs in your browser — nothing to install and no sign-up. An object circles at constant speed while its velocity keeps turning — speed is a scalar, velocity a vector.
| Control | Range | Step |
|---|---|---|
| Speed | 1 – 10 m/s | 0.1 |
| Radius | 1 – 3 m | 0.1 |
Your speedometer would look boring in this sim. Set the Speed slider to any value and watch the number sit perfectly still, minute after minute, while the object loops around the circle. Yet the velocity arrow attached to that object refuses to hold a pose: it swings, rotates, and points somewhere new at every instant. That single picture captures the whole distinction. Speed is a scalar, the plain magnitude of how fast you are going. Velocity is a vector, bundling that magnitude together with a direction.
Because the arrow always lies along the tangent to the path, its direction changes continuously as the object circles, even though its length never grows or shrinks. So a constant speed does not buy you a constant velocity. Drag the Radius slider wider and the loop stretches; one lap now covers a distance of 2·π·radius, yet the reading stays fixed while the direction keeps turning. Any change in velocity, direction included, means the object is accelerating, so it accelerates the entire time, pulled by a centripetal acceleration aimed straight at the centre.
Confuse the two words and half of rotational motion stops making sense; you would expect no acceleration whenever the speedometer holds steady, and you would be wrong. Keep the vector and the scalar separate and orbits, turns, and spinning wheels all fall into place. Pin down a single reading with the velocity calculator, then step sideways into the wider collection of motion sandboxes.
Start with the sim's own limitation: its speed is an input, not a measurement. The slider pins it, so nothing can slow the object down. Every real circling body bleeds speed to friction and drag and needs a tangential force to top it up, and the instant a force has a tangential component the acceleration stops pointing at the centre. It tilts off the radius, and a = v²/r then gives only one component of the total. A car braking through a bend is the everyday case: turning and slowing at once, its acceleration points somewhere between inward and backward, and neither the speedometer nor this sim captures that.
The bigger idealisation is instantaneous velocity itself. It is a limit, not something any instrument reads. A speedometer, a GPS fix and a radar gun all average over a finite interval, and on a circle the average velocity follows the chord while the average speed follows the arc. The sim draws both: the wine line is the chord, the distance readout is the arc. Run one full lap at the default settings and the gap becomes total: the distance readout reaches 12.57 m while the displacement returns to 0.00 m. Over that lap the average speed was the full 4 m/s; the average velocity was exactly zero. Halfway round, at 6.28 m of distance, the displacement reads 4.00 m — the diameter, not the arc.
Then there is the frame. The sim silently picks the centre of the circle as a stationary origin, so "constant speed" is a claim about that choice, not about the object. Put the same circle on a moving train and every velocity vector gains a constant offset: the speed now peaks where the object moves with the train and dips where it moves against it. The one quantity the sim calls constant is not constant at all. Acceleration survives a change of inertial frame unchanged, which is why Newton's second law is written with acceleration and not velocity.
Finally, taking a magnitude destroys information you cannot recover. Two objects running this circle in opposite directions give identical speed readouts of 4.0 m/s and opposite velocity vectors; if they meet, they close at 8 m/s, not 0. That is why speeds cannot be added or averaged the way velocities can, and why averaging speeds along a journey with turns overstates how far you actually got.
Keeping a satellite in orbit. The International Space Station orbits at about 7.66 km/s, roughly 6,770 km from the Earth's centre. Its speed barely varies over months; its velocity swings through a full 360° every lap. Feed those into the same two relations the sim uses and you get a centripetal acceleration of 8.7 m/s² and a period of about 93 minutes — and that acceleration is almost exactly local gravity at that height. Gravity is not switched off up there; it is entirely used up turning the velocity vector. That is what weightlessness really means.
Judging whether a bend can be taken. A car at 20 m/s (72 km/h) on a 100 m radius slip road needs 4.0 m/s² inward, about 0.41 g — comfortable for a dry tyre. Because the acceleration goes as the square of speed and inversely with radius, the margin vanishes fast. Push the sim's sliders to Speed 10 m/s, Radius 1 m and the readout jumps to 100.00 m/s², twenty-five times the road case, while the speed readout still shows a placid 10.0. The speedometer gives no warning at all; only the vector does.
Throwing a hammer. A men's competition hammer weighs 7.26 kg and swings on a radius of roughly 2 m, leaving the thrower's hands near 28 m/s. That is 392 m/s² of centripetal acceleration, about 40 g, so the athlete hauls on the wire with roughly 2,800 N — the weight of three adults — purely to bend the hammer's path, none of it making the hammer faster. The reward is a throw of about 80 m. Set the sim's Radius slider to 2.0 m for the geometry; the real release speed sits nearly three times past the slider's ceiling.
Speed is a scalar — how fast, with magnitude only. Velocity is a vector — the speed together with a direction. Speed is the magnitude of the velocity.
Yes. An object moving in a circle at a steady speed has a constantly changing velocity, because the direction of its motion keeps turning even though its speed does not.
Yes — even at constant speed. Its velocity changes direction continuously, and any change in velocity is acceleration; here it is a centripetal acceleration pointing toward the centre.
Along the tangent to the circle — the instantaneous direction of motion. The speed is simply the length of that velocity vector.
Yes, and sharply: the centripetal acceleration is v²/r, so halving the radius doubles it. Hold the Speed slider at 4 m/s and drag Radius from 3.0 m down to 1.0 m — the acceleration readout climbs from 5.33 to 16.00 m/s² while the speed readout never leaves 4.0.