Adjustable Floating Regulator

This is a rudimentary Adjustable Floating Regulator using an operational amplifier. The term "floating" means the regulator itself has no terminal that connects directly to circuit ground. Floating regulators can be used with very high input and output voltages as long as the difference between these voltages does not exceed the regulator's maximum specification. This circuit demonstrates the basic concepts of three-terminal adjustable floating regulator, a configuration which is employed in the popular LM317. However, the LM317's bandgap reference and the way the output driver works will not be emulated here. Paramount to the operation of three-terminal adjustable floating regulator is the reference voltage and its excitation current. The reference voltage must be low since usually this voltage is also the minimum value that the output can be adjusted to. The current that flows out of the reference terminal should be low, predictable, and stable since a change in this current will cause an error in the output voltage. For this reason the regulator's quiescent current is returned to the output, not to the adjustment, terminal (a 5-terminal op-amp is needed to accomplish this, it should be noticeable in the circuit). As a consequence, the quiescent current should be kept low so that the minimum load required at the output is also low. My objective is for this circuit to be opened and simulated in Free Subscription version of Multisim Live. Due to some limitations such as limited device models, Components per circuit constraint, etc., the circuit is crude in its functionality. Here are the shortcomings of this circuit: Voltage Reference - the most appropriate voltage reference currently available in the Free Subscription version of Multisim Live is Zener diode. I set the Zener voltage to 3 V, 1.25 V is common in popular voltage regulators. Reference Excitation Current - I used the 20 mA default value in the model, this is too high compared to what is used by popular voltage regulators (the typical value for the LM317 is 50 uA). Although I could have reduced this value to refine the circuit I decided not to since if the circuit will be built this value will depend on the actual Zener diode anyway. To generate the 20 mA excitation current I simply used a voltage-controlled current source to avoid exceeding the number of components limitation. To minimize the error in the calculated output voltage, the programming current is kept much greater than the current that flows out of the adjustment terminal. In this way, fluctuations in the current from the reference circuitry will have a negligible effect on the output voltage. In the circuit shown the current through Rset is 3 V / 15 Ohms = 200 mA. The current that flows to Radj = 200 mA + 20 mA = 220 mA. The voltage drop in Radj is 220 mA x 0.6(50 Ohms) = 6.6 V. The output voltage is 6.6 V + 3 V = 9.6 V. The output voltage will be close to 9 V if Radj's position percentage is adjusted to 54.545 %. Note: A voltage-controlled current source, instead of just a current source, is used to excite the Zener diode to have a reminder that the quiescent current in the reference circuitry should be returned to the output (not to the adjustment) terminal. The wire that connects the bottom of reference excitation current control circuit to the output node can be deleted and the result of the simulation should have a negligible difference.