BJT current mirror PUBLIC

Two NPN bipolar transistors form a current mirror.

A current mirror essentially "copies" a current injected into one transistor (Q1 in this case) and creates a current source (into the collector of Q2 in this case). While this circuit doesn't do much on its own, it's a core part of lots of analog amplifier circuits, as it can ensure that two parts of the signal chain have the same bias current running through them.

This is also an example of a translinear circuit, as Q1 acts as a diode, converting a current to a (logarithmically-related) voltage. That voltage is applied to the base of Q2, which acts as an (exponentially-related) voltage-controlled current source.

For the technically curious, also see these recitation notes from MIT Prof. Joel Dawson about current mirrors.

A few things to try in CircuitLab

DC

Run the DC simulation, and you'll see the two collector currents. Both are within a few percent of the 1mA control current. (The few percent of error is due to the base currents required to drive both BJTs.)

DC Sweep

The current source I1 can be swept over decades of current, from microamps to milliamps, but the currents match within a few percent over the entire range.

Bode

How fast is our current mirror, and how quickly can it respond to small variations in the control current? Run the Bode analysis to see. It's quite fast -- the frequency response is flat to tens of megahertz! Note that this is relative to the 1mA bias current, so if we reduce this to instead be 1uA, we'll find that the corner frequency moves down into the hundreds-of-kHz range.