Two Stage Buffers
      The next step is to address the problem of the MOSFET's high input capacitance. High power MOSFETs carry the penalty of a large amount of capacitance from the gate to the drain, which requires a fairly heavy current to charge and discharge quickly enough to provide an adequate slew rate and a wide bandwidth. In fact, capacitance against slew rate equals current :
      I = SR x C.
The gate to source capacitance is not as much a liability, as the source will follow the gate closely, resulting in just a few volts of gate swing relative to the source. What is, in truth, an added hassle is the interconnect's capacitance, which can be substantial for long runs.   So then, how do we drive all of this capacitance? Two possible solutions stand out: use a more robust line-stage amplifier; for example, one that used a 6BX7 or 2A3 as the output tube; or add a second stage to our high-power buffer.
     The first approach is obvious enough, but the second approach entails a few subtleties.     Adding a pre-buffer to our present circuit can take two forms, if were restrict ourselves to solid-state devices: adding a N or a P-device. (If we decide to use a vacuum tube our only choice is N-device, as that is the only flavor tubes come in.)

      Buffer with an active constant current source

      Here is a case where topological blindness can occur. If the input to the buffer had been an IC that fed and controlled the top MOSFET, many would complain that they did not want some cheap high feedback Op-Amp controlling the output. Yet if we move the same Op-Amp to the bottom of the circuit, there are no complaints. (A similar situation occurs when dealing with high-voltage regulation: few complain about an LM741 controlling a high-voltage transistor that is 100% in the signal path, although they would scream if the Op-Amp was moved to the middle or the bottom of the circuit. Very strange indeed.)
      Overcoming the Op-Amp's sonic contribution is easy enough. Adding one resistor and one capacitor is all that is needed. In the following circuit, two feedback loops are made.      The Op-Amp sees two feedback loops: one AC and one DC. The AC feedback loop extends from its negative input to its output. The DC feedback loop extends from its negative input to the top of the 1 ohm sensing resistor. Thus any AC signal coming from the voltage reference will be passed to the Op-Amp's output, but any AC signal from the sense resistor will be shunted away though the .5 µF capacitor, while any DC signal will be preserved.

< PREVIOUS

NEXT >