The example presented here is a four-element LED bar-graph display. The LEDs light up sequentially from left to right as the current (I) is increased. Because the circuit is entirely analog, the LEDs turn on gradually until the current limit for each stage is reached. Once the limit is reached, the next LED begins to turn on. A current source isn’t strictly needed to drive the circuit; a voltage source with a series-limiting resistor will work just as well.
The npn transistor of the first stage (Q104) is connected as a diode. This is shown for reasons of symmetry and can be omitted in a practical circuit. As the current (I) increases, it flows through R100 and DS100, causing the LED to begin to glow. As the current increases, the LED intensifies. This continues until the voltage drop across R100 reaches about 0.65 V at room temperature. At this point, the base-emitter junction of Q100 becomes forward-biased, which turns on the second stage of the array.
As the current is further increased, the intensity of the first LED remains constant due to the current-limiting action of R100 and Q100. Additional current then flows through the second stage of the array and the intensity of the second LED increases until its current limit is reached.
As the current through the network continues to rise, this process is repeated for each successive stage in sequential order. There is no theoretical limit to the number of stages that can be connected.
The main design criterion for this circuit is the current limit of each stage. This is simply determined by the resistor across the base-emitter junction of the pnp transistor. In the example shown, the current limit is 0.65 V/100 Ω= 6.5 mA (see the figure). Typically, all stages would be identical, but this isn’t required. By using different current limits for each stage, some interesting functions are possible, such as a logarithmic bargraph display. A logarithmic display also would include resistors to shunt excess current around the LEDs so that the full intensity of each LED would be the same.
One drawback of this simple design is that the current limits are a weak function of temperature due to the nature of the pn junctions in the transistors. Series resistors aren’t needed at the bases of the MMUN2211 devices because they contain internal resistive networks to serve the same purpose.
The usefulness of this circuit as a bar-graph LED display is already evident from the example. A side effect of the recursive stages is that they clamp the voltage appearing across all of the stages. In this regard, the circuit behaves as a shunt voltage regulator. The clamp voltage depends on the loads driven by the stages. If the loads are LEDs with a voltage of 1.6 V at 6.5-mA bias, the clamp voltage for the network will be approximately 2.9 V. In the example, this is the sum of the voltages across Q104 (connected as a diode), the LED, and the resistor.
Of course, shunt voltage regulators are extremely inefficient for most applications. However, they do possess a unique property that can be quite useful in certain circumstances. Shunt regulators are superb at dissipating excess energy from a source, such as when float-charging a lead-acid battery with an array of photovoltaic solar cells.
As the battery becomes fully charged, the excess energy from the solar cells would typically be consumed as electrolysis in the cells of the battery. This depletes the battery of electrolytes, eventually destroying it. A shunt regulator connected across the battery prevents excessive current from entering the battery, saving it from destruction.
The main advantage of this circuit compared to other shunt-regulator topologies is the ability to discern how much extra current is available from the source. In the aforementioned battery example, this circuit acts as a kind of fuel gauge that indicates the current battery-charge level. If more stages are on, less current is entering the battery and the battery is in a higher state of charge.