Placing the series resistor at the bottom also yields advantages. Some safety enhancement can be had from having one of the resistor's leads at ground potential and the other at much lower voltage than the top arranged series resistor. (Remember, voltage is analogous to velocity with moving objects, double the voltage or the velocity and wallop increase fourfold: kinetic Energy = ½MV² and Joules = ½CV². But the big advantage is that by referencing the ground at one of the resistor makes the other end a negative power supply. And negative power supplies are handy to have in tube equipment, but often difficult to create. For example, a power amplifier's output stage might need an negative power supply to bias the output tubes, but the high voltage transformer may not have a low or medium voltage tap.

    From this viewpoint, the ground is the signal and any deviation from that signal at the output is countered by the feedback until it is eliminated.
    So which is better? Better for what? Better for needlessly heating the room? Better for using a battery power supply? Better for isolating delicate signal currents in an amplifier? Each finds its own place in tube audio designer's palette. My preference is to use kind for kind: a shunt regulator for steady average current draw circuits, for example Class A amplifier stages, and a series regulator for variable current draw circuits, for example Class AB, B output stages.

An Inverted Shunt Regulator
   The usual topology is to place the series resistor in between the load and the B+ connection, just as a choke would normally be placed in a non-regulated power supply. However, just as the choke can be placed at the bottom, between load and negative power supply, the shunt regulator's series resistor can be placed at the bottom as well. What advantage can be wrestled from this inversion? In the case of the choke, the increased safety of moving the choke closer to ground stands out as a large advantage. With the choke at the top, its winding sees the full B+ voltage and should the wire's insulation breakdown, the voltage difference between the B+ voltage and the chokes ground level metal structure will make for some fireworks. But when the choke is placed at the bottom, the choke's winding and case only see the voltage developed across the DCR of the choke plus the ripple voltage, a total of 30 volts let's say. Now, the insulation should not breakdown and even if it did, 30 volts makes for much smaller sparks.

    And when a transformer does have this tap, it is seldom matched with another tap on the other side of the center tap, which means that only a half-wave rectifier circuit can be used. The problem with half-wave rectifier circuits is that they are useful only if a minuscule amount of current is drawn from them, as they yield much more ripple and much less current than the full-wave rectifier circuit. (Excessive current draw through the half-wave rectifier also tends to magnetize, i.e. saturate the power transformer's core, as the current is being only drawn in one direction through the power transformer secondary winding.) And the alternative, using a full-wave bridge rectifier circuit, yields too much negative voltage and usually precludes using tube rectifiers.