Two balls in a can
Two steel balls roll to whichever end of the can is lower. At the lead end they rest across the two contacts and close the switch; at the other end they touch nothing. One maker's sheet puts the change about 10 degrees either side of level, and between those two angles the balls stay where they last stopped. Level is not a reading.
A weight that closes a gap
Inside the brass can are two steel balls and nothing to hold them. At the lead end, the end with the two wires, the switch's contacts sit where the balls come to rest. When that end is the lower one, gravity rolls both balls down onto the contacts and they join the two leads: the switch is closed. When the other end is lower, the balls roll away to it, where they touch nothing, and the switch is open.
Drag the angle. Past about 10 degrees either way the answer is decided, and tipping further changes nothing: this switch has two states and no sense of how far. Between the two, watch what the balls do: nothing. On a nearly level can they have no reason to roll, so they stay at whichever end they were, and the switch keeps whatever state it had. Come to 0° from the left and it reads closed; come from the right and the same angle reads open.
The inside is drawn to explain it, not measured. Both makers draw the part from outside only; that there are two balls, and that the lead end is the closed end, is theirs.
What the two sheets say
The part on this board is Shouhan's SW-520D. Its sheet calls the lead end the ON end: the switch closes when shaken, or when the lead end sits below the horizontal "at a suitable angle", and it opens again only when it is still, with the other end below the horizontal. It gives 12 V and 10 mA as its limits, and no angle.
Bailin makes the same part and does give one: laid horizontal, it closes once the lead end is more than 10 degrees below level, and opens once the other end is. It also puts the closed resistance under 10 Ω and the life at up to 100,000 cycles. The book uses its 10 degrees, and says so each time, because it is one maker's figure for a part this board did not buy from it.
Level is the dangerous place
Shouhan's sheet adds the useful warning: laid horizontal, the switch is the easiest position of all for a shake to trigger; with the closed end pointing up, the hardest. Near level the balls are balanced, and a tap or a vibration moves them. A rattle is not a tilt covers what that does to a reading, and how a sketch ignores it.
When it does not work
Bailin's SW-520D sheet says about 10 degrees past level, either way; Shouhan's, the maker of the part on this board, gives no angle. Nobody has measured this block. Tip it slowly while watching the LED: that is the honest answer for yours.
Near level the balls are free to roll, so a tap can move them onto the contacts or off them. Shouhan's sheet says as much: laid horizontal is the position a shake triggers most easily, and with the closed end up it is hardest to trigger.
It does the same job without the mercury: a conducting weight that rolls to the low end and bridges two contacts. Metal balls bounce where mercury flowed, which is why this one rattles and needs a filter.
Where the can sits on this board, and which ways of tipping it do anything.
Which end is lower →Edit this page — content/books/tilt-sensor/two-balls-in-a-can.mdx
Questions about this product
See what other owners have asked, and read their solutions.
Tilt Sensor
Loading discussions…
Discuss this article
Ask about this page. The answer stays here, on the page it belongs to, for whoever hits the same wall next.