Semiconductor curve tracers were frequently used in analog circuit designs many decades ago, at a time when discrete semiconductor devices were in dominance and ICs were scarce. While curve tracers are no longer widely used nowadays due to the ubiquity of digital circuitry and computer aided designs, they are still quite popular in the educational world and among hobbyists. In this and the next post, I will discuss some of the design considerations of a curve tracer that I built and show some real-world measurement results.
A typical use of a curve tracer is to generate the V-I characteristics of the device under test (DUT). The following diagram illustrate how a typical configuration works when the DUT is a BJT.
In the figure above, a staircase waveform is fed into the base of a NPN transistor via a base resistor R2. For each step in the waveform, a different base current (Ib) is generated. When the input waveform voltage is sufficiently large, the voltage drop between the base and emitter junction (Vbe) can be ignored and thus:
where Vbi is the voltage of the ith staircase. This arrangement is usually good enough for most cases and this is what I used in my design as well. If we need more accurate characterizations of different Ib’s, we can either use a variable current source (which increases the circuit complexity) in place of the staircase waveform generator or simply use the same design but add a current sensing Op Amp to read back the actual base current.
A sweep voltage is applied between the collector and the emitter. The sweep waveform is usually a symmetrical triangular wave so that characteristic curves for both sweep directions can be captured. When we only care about the characteristics in one direction (say when the voltage is ramping up) we could use a saw tooth waveform instead. The linearly of this swept waveform is not significantly important, and in many curve tracer implementations (like this one) the sweep waveform is simply bridge-rectified from the transformer directly. Since the rectified AC signal is sinusoidal, the end of the curve tends to be brighter than the rest of the curve compared to that using a triangular swept signal.
In the diagram above, the collector current is measured at the emitter side via the voltage drop across the emitter resistor. While technically speaking, the measured voltage corresponds to Ie not Ic, but usually Ib is small enough and can be safely ignored without any significant impact to the overall measurement accuracy. For some power devices however, the common emitter gain (hfe) can be quite small (e.g. in the teens) and the collector current measurement can be off by quite a bit when the measurement is done on the emitter side.