Version 2.0: Featuring improved appearance and bug fixes.
The process of solving the ground-loop problem involved trial and error, which produced useful modifications to the power supply:
Figure 6. An overhead view of the revised layout.
The first change was the replacement of the EI/frame-type transformer with a toroidal unit, guaranteeing no appreciable transformer hum in spite of the continuous loading by the Class A amplifiers. The new unit is rated at the same 50VA and 24V output as the original unit, and still rectifies and filters to about 32V. The toroidal design, however, offers a compact structure and a smaller EMI radiation field.
The second change was a revision of the filterbank layout. Unlike the original design, all four capacitors are now centered around a proper star ground, with the smoothing resistors moved to the outside for improved dissipation.
With the noise problems solved, the only thing left was to install the top cover and put the unit to good use:
Figure 7. The front side of the completed amplifier.
The power switch is found at far left, with a white LED indicator lamp mounted next to it in a chrome housing. The two output jacks and volume control occupy the right-hand side.
Figure 4. The rear side of the completed amplifier.
Viewed from left to right are the input jacks and channel-reversing switch, the heatsink, and the AC inlet jack and master fuse holder. Simple and effective.
The Version 2.0 headphone amplifier was more expensive that the first, and well worth the investment. The heatsink only becomes warm to the touch with both amplifiers biased to about 250mA. There is no mechanical hum or noise, and the output is electrically silent at idle even if the input jacks are left floating.
In short, with the benefit of prior experience and a growing knowledge base, I was now able to build a superior Class-A headphone amplifier, and it was used regularly for several years afterward. §
aaronv dot net -at- gmail dot com