Positive exponentials (against time)in electronics anywhere?Thought "experiment"

[Infraviolet]

Just some idle ponderings but...
Can anyone think of an example of a positive exponential voltage (between any two points) or current (with respect to time in whichever case) occuring anywhere in electronics?

There are plenty of negative expontentials "e^(-const*t)", and of "const - e^(-const*t)", such as in the discharging and charging of capacitors, but what about positive expontentials which run away with themselves, until some practical limit is reached and the behaviour is stopped at a steady state (including by the circuit being ruined in the process).

The closest thing I can think of is where the resistance inside a MOSFET rises with the MOSFET's temperature, in which case the temperature can go up expontentially until heat losses reach equilibrium or the mosfet is destroyed, but that's a temperature rising exponentially, rather than a strictly electrical parameter, like a voltage or current.

[IanB]

It's going to happen anywhere you have positive feedback. Practical examples would be with a comparator, or a Schmitt trigger.

If you plot the change in state over a short enough time scale on an oscilloscope you should be able to see the exponential rise of the signal.

[Benta]

Current through an NTC resistor at constant voltage?
Voltage over a PTC resistor at constant current?
Both have to be self-heating, of course.

[twospoons]

Thyristor turn-on?

[tom66]

Avalanche breakdown (in diodes, zeners > 5.6V) has a positive exponential term in it, if I remember correctly. 

[SiliconWizard]

An exponential is a reasonable approximation of a diode I-V characteristic (Shockley model). Generate a ramp across a silicon diode in series with a resistor, look at the voltage across the resistor and you'll have roughly an exponential as long as you stay away from breakdown of course. The breakdown part is a different story which I'm not covering here.

[ebastler]

An exponential is a reasonable approximation of a diode I-V characteristic (Shockley model). Generate a ramp across a silicon diode in series with a resistor, look at the voltage across the resistor and you'll have roughly an exponential as long as you stay away from breakdown of course. The breakdown part is a different story which I'm not covering here.

That approximation is good enough to be widely used in analog computers to calculate log and exponential functions. 

Including the use in multipliers which calculate a*b = exp (log(a)+log(b)). Typically using a transistor in diode connection, rather than a plain diode, for improved dynamic range to my knowledge.

Edit: Oops, only now have I registered the "against time" in the thread title. So log or exponential amplifiers don't count... As others have pointed out, positive exponential growth can be generated via electronic components or circuits, but I can't think of a common technical application.

[SiliconWizard]

Yes. The relatively "common" application is the log amplifier: https://en.wikipedia.org/wiki/Log_amplifier

[Circlotron]

Take the output of an opamp integrator, put it through a X1 inverting amplifier then feed the output back to the integrator input.