|
The right side uses an ideal constant current source while the left is the actual circuit. This kind of driver is pretty commonly available and useful for controlling current through a load with a not-so smooth power supply. In my experience these drivers work very well for low current applications (<50mA) such as controlling a string of LEDs being powered from a boost converter with some switching noise on the output. The 10k resistor and two BJTs are the internals of the driver; the low value resistor is (usually) an external component you’ll need to select based on your application. The switch on the left shows how digital logic like a comparator output, microcontroller pin, etc. can be connected to the center node as an open drain and pull it to ground to act as an output disable function.
The working principle of this circuit is to use negative feedback to adjust the voltage drop across the collector-emitter path in the BJT in series with the load to create a stable current through it. The current passes through the load (LED here), through a transistor, then through a low value resistor into ground. Using ohm’s law and knowing the current through the low resistor you want and roughly the Vbe of the bottom BJT, about 0.7V normally, lets you calculate the resistance needed to set the current at a desired value. An increase in the power supply voltage leads to an increase in current through the load that in turn increases the voltage drop across the bottom resistor. That increased voltage makes the transistor on the left conduct more, reducing the voltage at the top BJT’s base causing it to conduct less and hence decreasing the current in the load branch. Likewise a drop in the power supply’s voltage decreases the voltage on the Iset resistor, increasing voltage at the top BJT’s base and decreasing its collector-emitter’s impedance to increase the load current.
A drawback of this kind of circuit is that it’s pretty low precision because of its reliance on the voltage drop across a resistor in parallel with the base of a BJT. The voltage across Vbe will not always be 0.7V but varies with current through it and hence as current changes due to power supply fluctuations the Vbe will bounce around as well and offset the Iset. Secondly, this circuit is only suitable for low current applications. As Iset increases the bottom resistance gets lower and lower until eventually it’s just a short to ground and Vbe can no longer be realistically guaranteed. At that point the current in the load will become very unpredictable and prone to swaying wildly. That said I’ve found this kind of circuit to be a cheap and effective way to control a load in a project with good enough accuracy given a noisy power supply and low precision needs.
|
