Project history & acknowledgementsHaving been an active participant in several DIY audio and headphone related forums for a few years, and launching the successful M³ headphone amp project (with Morsel, the other member of Team M³) in 2004, I felt that the time is ripe for a "bigger brother" to the M³. While the M³ is an excellent amplifier, it was designed with a keen eye towards high price/performance ratio. The new amplifier would instead push much farther into the high-end, where cost and complexity concerns are less of a constraint. Nevertheless, I am a pragmatic designer and I don't like high cost and complexity as an end to itself. Thus was born the β22 amplfier project and its companion power supply, the σ22.I had long been fond of using power MOSFETs as output devices. The β22 and σ22 follow the footsteps of the M³ in applying these devices to good use, this time pushing the bar even higher. All-discrete and complementary topologies and a number of other design highlights contribute to lower distortion, higher output, wider bandwidth, and most importantly, better music. The β22 amplifier topology had actually been in gestation for several years prior to my announcement on headwize.com. I had shown various iterations of the schematic diagram to visitors at my "mini-meets", but other projects had taken precedence until September 2006, when I decided to let the cat out of the bag, and turn this project to reality. Elements of the β22 design were direct beneficiaries of work done in the past by Erno Borbely, W. Marshall Leach, Nelson Pass, John Curl, Richard Marsh, Walt Jung, James Bongiorno, Douglas Self, Rod Elliot, and others. These gentlemen have generously published their insights and the state of audio art had advanced because of them. Unlike past projects, I took full advantage of PSPICE simulation in the design phase of the β22 and σ22. This is partly because of the all-discrete nature of these projects, and also due to my acquisition of the OrCAD software suite. It turned out that simulations paid big dividends in allowing me to explore the effects of various design elements, and made it possible for me to tune the circuit without having built a real prototype. Changing something and seeing the result is a snap, which would have been cumbersome and expensive with real circuits. In fact, instead of the usual breadboard prototype, I went from simulation directly to a prototype PCB. Actual circuit testing is still of paramount importance, of course, because simulations do not account for all real-world issues. I was pleasantly surprised, though, to find that the simulations were quite accurate. Participants in the headwize.com DIY forums provided valuable comments and suggestions, which helped the circuit and the board layout evolve into its current form. Gratitude is extended to all who took the time to write comments and feedback. The first β22 and σ22 printed circuit board prototypes were fabricated in October of 2006, on which I was able to build a working system. The production board, with only very minor modifications from the prototype, went to fab late November 2006. General availability of boards began in December. Look for more exciting DIY projects from AMB in the future! Main: β22 Main | Prev: β22 gallery | Next: Buy |
