It came up at work today and we all had this vague ‘I’ve heard of it before’ reaction but no one could remember what it was.
“Hysteresis”, perhaps? Usually used with reference to the magnetization of ferromagnetic material, it is the failure of some property induced by an external force to return to its original value when the force is removed (e.g., a ferromagnetic material can be permanently magnetized by an external field, as opposed to an electromagnet, where the magnetism disappears when the electric current inducing it is turned off).
(FTR, “hysteresis” has no conection with “hysteria”.)
Hysteresis IIRC is not only applied to magnetism but any time a force/energy is not conserved. Hence you could draw a “loop” consisting of an upper curve (the force originally applied) and a lower part (the force recovered eventually).
Historisis is the term used to describe the ability of a material to return to its original state after having a force applied to it. For instance, if you flex a well-constructed metal knife slightly, it will return to it’s original shape when you release the force on the knife. Similarly, rubber bands will return to their original shaped after being stretched.
There are practical limits to historisis, however. If you apply too great a force, the material will be permanently altered. This is hysteresis, as Akatsukami describes.
Etymology note : ‘hysteresis’ comes from Greek for ‘lagging behind’. You had a decent guess with a ‘histori-’ form (you appeared to know it’s time-related), though ‘history’ actually comes from a word for ‘to inquire’.
Anyways, in the most general terms, it refers to an effect that lags behind its cause, a temporary memory effect.
If you’re dealing with circuits, especially logic levels, it refers to the change in switching levels as you approach them from high or low. If you switch from low to high at some level, you will actually have to go lower than that level to switch back low. This provides better noise immunity than a circuit with a single threshold.
For example, suppose you have a circuit that switches between 0 and 5 V (low-high):
Start out at 0V, and increase the input. At some point you reach the trigger threshold (let’s say 2.3 V), and it switches high.
At high level, you decrease the input until it switches back down. If there were no hysteresis, it would switch low as soon as you got to 2.3 V. But let’s say instead it doesn’t switch low until you get down to 1.8 V.
Then you have (2.3-1.8 =) 0.5 V of hysteresis (the way it’s usually spec’ed).
I always use a thermostat analogy for hysteresis. Consider the thermostat in your home, say you have it set to 68 degrees F.
When the temperature in your house is high, say 70, the thermostat switch is off. As the house cools off, it approaches 68 degrees from the warm side, and at exactly 68 degrees the switch flips from off to on. However, the switch doesn’t immediately turn off, because that would lead to inefficient rapid cycling of the furnace. It waits until the temperature is higher, say 69 degrees, before switching off.
It’s essentially a system that is not described solely by its state, but has some historical background to it. In the case of the thermostat it is not sufficient to know that the temperature is 68.5 degrees to determine if the switch is on or off - you need to know the historical behavior of the temperature to know the on/off state of the switch. 68.5 degrees on the way down from 69, switch is off - 68.5 degrees on the way up from 68, switch is on.