Recently I wrote about my first experience with surface mount technology.  As part of that ongoing project, I needed a capable chip mounted on a development board.  That chip turned out to be an Atmega169 TQFP, which I will clock at 16Mhz.  To date, I'm still working only on single sided boards.  This would be my most ambitious project yet.

Like the last board that I made, I printed this one with a lithographic process.  The mask was printed on velum and then transferred to the positive sensitized board using a fluorescent lamp.

  Velum actually works surprisingly well.  It doesn't warp when printed, and holds toner quite well.  Even though this board is very basic, you can see that the routing is quite dense.

(Click on the images to enlarge) The actual mask that I used was a bit different from the one pictured, but not too much.  The traces are about .010" thick.  Unfortunately that makes then prone to over-etching.  The etching process was taking too long in the basement (about 60f) so I moved the acid container to the oven.  Then it went too quickly and killed a couple of my traces. 

The soldering is pretty ugly, for two reasons.  First...I bought a new iron a couple months ago, and it's way too hot.  Second, I failed to remove all the photo resist and it ended up burning and also preventing the solder from sticking well.  Usually I just sand it off, however I was reluctant to sand too much on this board because of the small trace size.

Despite my misgivings, I was actually able to connect to the board via the programming header on the first try!