Symmetrical Output Stages
       In spite of a truly complementary set of MOSFETs, the symmetrical output stage is the standard. This is easy enough to understand: two symmetrical source followers working into the same load. It can be run in Class-A, AB, B and it can be driven from a single phase. This last point constitutes a major feature for many, as phase splitters confuse most beginners. Furthermore, a supremely simple amplifier has its charms. For example, all that is needed is one triode, two MOSFETs, and one interstage transformer as is shown below.

     Many will balk at the use of an input transformer. Wasn't the great advantage to solid-state amplifier that they eliminate the need for signal carrying transformers? Yet, a good transformer solves as many problems as it creates. The operative word in the last sentence was "good." Good transformers are both expensive and hard to find. The obvious solution is to replace the transformer with a coupling capacitor. (DC coupling is impossible, as the bottom triode's grid is -120 volts below ground level.)
     The amplifier below is a simplification of a   design sent from David S., a reader  wishing to build the simplest hybrid possible. (A similar circuit appeared in a mid-80s MJ Stereo Technic article by Sidewinder.) It use a coupling capacitor to block the huge voltage offset between the input and the bottom triode's grid.

      While many bells and whistles are required, such as grid and gate stopper resistors, fuses, LED bypass capacitor, gate protection zeners, and source resistors, the amplifier is actually functional as is and could be described as the world's simplest hybrid. The amplifier does not even invert the output signal phase, as the input transformer inverts the phase for us. To make a truly useable amplifier, a bit more complexity is demanded.
     One obvious danger with this existing topology is that at turn-on the tube is not yet conducting and the gates will see +120 volts. Replacing the plate resistor with another triode configured as a current source will help a bit, as both tubes will heat up in unison.
    Another problem area is the lack of a DC offset nulling control. The simplest solution is to add a potentiometer to half of the circuit so that the output can be brought to zero volts. The slicker and safer approach is to use a DC servo to steer the output stage to zero volts.

    With the coupling capacitor comes a problem. I am sure that many of you have caught on to how passionately I hate noise. So naturally, that is the first aspect I look at in an amplifier's design. The bottom triode has two inputs the first obviously is the grid, the second is cathode. Grounding the input makes this clearer. The grid sees the ground's zero noise voltage while the cathode sees the full negative power supply rail noise voltage. This noise is then amplified in-phase at the plate. Now, if this amplifier used a great deal of negative feedback, the true source of the noise at the output might not be found. But with a zero feedback amplifier, less is hidden.

< PREVIOUS

NEXT >