Page Title

The two coupling capacitors are a nuisance, although much less so than an input transformer. Still, it would be nice to do away with them. One possibility is to use two single cell batteries, AAs for example. One advantage to using batteries is that as they discharge, the idle current will decrease, not increase, as often happens in other battery bias schemes. Alternatively, we can just use resistors, as shown below.

Shifting the ground and output creates an amplifier out of the buffer topology. The amplifier below inverts the input signal at the output and uses two feedback loops to set the gain and lower the output impedance and distortion. Won’t the signal delivered to the load be compromised because of the electrolytic capacitors? Yes, it will. But guess what: the signal is compromised because of the electrolytic capacitors in a non-floating, conventional amplifier as well, as the output currents flow through power supply capacitors in both amplifiers equally. It’s just that we seldom think about the power supply’s parts as being in the circuit, although they are.

More ð

A note for the advanced practitioner, if the power supply’s capacitors didn’t find a connection to the power transformer’s center tap, then the output from this amplifier would be entirely capacitor coupled, which would eliminate the DC offset worries (a 1k resistor should then be connected from ground to the capacitors’ common connection, in case the capacitors differ in value). For me, though, the bigger problem is the mismatch between output devices. The next variation on the theme answers that problem.

As the graph above shows, the performance is excellent for so simple an amplifier. Translated into a THD percentage, the distortion comes in at about 0.1% at 1 watt output and 3% at full output, 35 watts. The output impedance is about 1 ohm. Bandwidth extends from DC to 200kHz. The 170pF capacitor is a cheat of sorts to extend high-frequency, as the 12k resistors working into the MOSFET’s input capacitance limit the frequency response. The gain is only 2.4 times the input, so the line amplifier will have to be able of putting out 10V peak. More gain would equal a higher output impedance and more distortion and a choice between a lower input impedance or a less extended bandwidth. Which, specification to the contrary, might not be a bad idea, as I have heard some excellent sounding amplifiers with high output impedances and 60kHz bandwidths.