As Samuel mentioned earlier, I've been having some problems building his design on boards supplied by Fabio. At work, I commonly hand-solder parts with pads spaced three to a millimeter, so I'm fairly confident my soldering skills weren't to blame.
So far, I've built three SGA-SOAs working very meticulously - measuring or testing every component and double checking the location before soldering it in place. Samuel's documentation is very useful and his assembly sequence was extremely helpful. Still, the first one didn't work.
The first module was drawing excessive current. The problem was traced to a short to ground at the base of Q3. No fault was visible even under intense magnification, but measurement showed the pad was definitely shorted to ground. My theory is that the solder is able to make a bridge just under the silk screen layer out of sight just under the white ink. The affected pad had white ink between the pad and the ground plane which formed a neat radius around part of the pad. Here's an image of the pad after I had fixed the short (click for enlargement):
I removed the transistor and used solder wick on the top pad and the short was gone. To ensure that problems didn't recur, I carefully scratched off the ink around the pad.
Number two gave me similar problems, though on a different transistor. When I pulled the transistor and used the same trick with the solder wick, I still had a short. Finally I discovered that while removing the transistor with a heat pen one of the ceramic capacitor pads had gone short as well when the solder reflowed. The same sort of bridge was found with white ink surrounding part of the radius of both pads.
On my third board, I tried deliberately to make a short, but found I was unable to do so. Closer inspection revealed that the ink did not make a complete bridge between the pad and the ground plane. Even so, I removed any ink from between the pads and ground plane before stuffing the whole board. This one worked without any problem first shot.
Hope this helps anyone who is experiencing problems with the batch of boards from Fabio. I believe Samuel is reworking his files to prevent this problem from occurring in the future.
Can anyone else share their experience?