I was looking into assembling this headphone amplifier for a long time, since I've learned about it's existence from Bob Katz' article. It's really an interesting design, featuring Class A operation and an active ground channel. I'm not sure that the latter actually makes a lot of sense (there are both pro and contra articles about it). But the combination of Class A operation and a discrete second stage based on MOSFETs definitely sounds interesting and is different from any other headphone amp I have.
From all the different options this design offers, I have chosen to build it with adjustable gain (instead of bass boost, since I'm not really interested in the latter). I've calculated the required resistors ratio to achieve 2x to 6x gain range—this is the same range that is used by default in Objective 2 headphone amp, and in my M3 it's possible to increase the gain gradually.
I've ordered all the parts I could from AMB's shop (that cost me around $100), and the remaining parts from Mouser (also about $100). After parts from Mouser has arrived, I discovered that the opamps (AD8610AR) have SOIC-8 package instead of DIP-8 that I was hoping for, so I had to solder them onto adapters (my first SMD soldering experience BTW!).
I was also thinking how to make all the external connections (inputs, outputs) detachable, and came up with a decision to use Molex 2 wire assemblies I found on Sparkfun. This has turned out to be very handy while I was experimenting with placing the board into the case and had to re-connect them multiple times.
In general, soldering the components wasn't hard at all—the board is quite spacey. The range of components' heights varies greatly from low-profile resistors and capacitors to big capacitors and heatsinks of MOSFETs, so make sure you arrange them by height prior to soldering.
I didn't even think about building my own, after reading about all the necessary precautions one must to follow when working with high voltage circuits. I decided to just buy a ready made linear power supply. I've found one based on Zero Zone board. It can go up to 24 V at 0.6 A, which is enough for M3.
It's actually turned out to be a neatly built power supply, ground-referenced case, with fuses, and as a bonus—an LCD displaying current voltage. The V- ground is floating (not connected to mains ground), just as M3 design requires.
That has turned out to be the most interesting experience for me. Assembled board is quite big (7" x 5") and tall, and a ventilated case is recommended. I found these requirements to be quite close to those of Mini-ITX PC form factor, and started looking for a nice looking computer case. I ended up buying Morex 557 Universal Mini-ITX case. Here is where the challenges has started.
First, the case doesn't have a back panel—it is supposed to come on the PC board. Second, ITX motherboard mounting holes obviously didn't match against the holes on the M3 board. Third, the front panel had to be modified to accommodate the volume pots and the giant Neutrik 1/4" jack (NJ3FP6C-B) I've bought. The good part, though, was that the case had a toggle power button with LED. The button itself wasn't of any use, but the case's LED connector fit nicely the Molex connector I've put on the M3 board for LED.
I've bought a sheet of medium thickness ABX plastic. I used it both for making the back panel, and for carrying the board's rear part. The front part is resting securely on volume pots (and this also makes them grounded, as M3 design recommends) and a little shelf used for latching the front plastic panel onto the inner metal panel. This way, I also have solved board mounting problem.
For the volume pots, a regular 8 mm drill bit worked fine. The Neutrik jack has the same diameter as XLR jacks (that is, enormous), which is called "size D" and is about 15/16" in diameter. I couldn't find a hole saw of this exact diameter in our local Orchard store, so I ended up using a 1" (25 mm) saw, and it has turned out that the jack covers it nicely. But I would still recommend trying to get a saw of 15/16" diameter if you would like a more exact match.
The fact that the rear panel is plastic helped in ensuring that the power jack is isolated from the ground. But this also mean that the RCA input jacks are not grounded either. So I soldered a wire to one of IG holes on the board, and secured the wire to one of the case's screws.
The final stroke was to cover the holes for USB ports on the front panel. See what I came up with :)