When I received the Big Box of Parts USA a few months ago I found an unpolulated Bus Pirate v4.0c board. I've been wanting a BP for some time and enjoy building whatever I can, so I grabbed this board and set off ordering parts. I was able to source all parts through Mouser (almost, explained later). I've been wanting to try my hand at reflow using a toaster oven as well, so why not pull two projects into one. The parts arrived, but to my dismay I realized the CD4066 chip I ordered was SO-14 (as noted on the wiki parts list) instead of TSSOP14, as this board required. Also, the shrouded header I received was not quite what I had expected so I tossed it in the parts bin and decided to use a normal unshrouded header instead. Everything else was as expected. I taped each of the part strips onto a printout of the parts list for easy identification while populating the board.
The next hurdle was going to be solder paste. I didn't want to spend money on a "custom" stencil for this one-off, and I didn't want to make a project out of making a stencil, either laser cut or etched. I decided to hand dispense solder paste on the cheap, so I ordered a tub of Mechanic paste from ebay. I also acquired a 5ml syringe and blunt tip. I knew with some of the smaller pads this would be tough. I carefully began placing the beads of paste on the pads. I had several spots where the paste bridged onto adjacent pads, but I hoped it would sort itself out during reflow. For the PIC and USB connections, I used a thin strip of paste instead of trying to keep to the individual pads. Due to the fact that I used a plunger-style syringe, the dabs of solder paste were inconsistent in size from one pad to another, some getting what appeared to be too much, others too little. Once complete, parts were hand placed using tweezers, and tapped or pressed gently into the paste to hold them down and hopefully avoid tombstoning.
The next hurdle was the reflow itself. I've been wanting to build a toaster reflow controller for some time now but wasn't sure how well it would do the job. For this project I decided to use a K-Type thermocuple digital thermometer to monitor temp in the oven, a stopwatch for timing, and my hand for on/off control of the oven to maintain the reflow profile. A few times the temp began coasting past my desired temp, so I cracked open the oven door a bit to bleed off a few degrees. This was not that hard to control, even manually. Once the temps got into the 210-220*C range parts started popping into place. I let it climb to 230*C at the peak and held it for about 20 seconds while watching to make sure the larger pads like the USB grounds reflowed, then I turned the oven off and let it cool to 180*C, then opened the door until about 150*C, and finally pulled the board out for final cooldown and inspection.
Everything reflowed very nicely considering the crude process I had followed. I had no tombstones, deformed connectors, or part-to-part solder bridging, but there were a few issues. Several of my pads had too much paste. This was hard to judge before reflow as the paste loses some mass when the flux cooks off but I wasn't sure how much. I wicked a bit of the excess solder from them using copper braid and a 25-watt pencil iron. A few other pads were connected, but had "too little" solder on them. The PIC had only a few bridges which were easy to wick away. The biggest issue though was with the PIC orientation. One side was perfect, the two adjacent sides were very good, but the last side was shifted about a pin's width out of line. The connections were held in place between the pads instead of being on the pads. As I don't have a hot air tool (now justified and added to my list), this would be an issue. I decided to pop the board back into the oven and redo the process. I also added a bit more paste to a few pads that seemed lacking. When the board hit a sweet spot around 230*C I popped the door open and tapped the PIC in hopes of shifting it into place. This was harder than expected and took several tries. Once satisfied, I ramped back to 230 and followed the proper cooldown to ensure no cold joints. The chip still wasn't aligned quite right, but had moved enough to eliminate the shorts.
At this point, the reflow work was done, so I hand soldered the headers in place, then connected the USB cable to see if it'd light up. It did! This was a good sign after all the board had gone through up to this point. Now, I still need to await the CD4066 IC, which I'll hand solder into place. I also need to get a boot loader onto the PIC. Unfortunately I only have a JDM2, so I'll be building a PicKit2 clone to flash the bootloader.
All in all, this project was a success! The process worked well enough to repeat and the reflow toaster puts an end to my fears of QFN and BGA chips. It looks like I'll be moving forward on the reflow controller project now that I've seen the results of an off the shelf kitchen toaster. I'll also have a BPv4 to add to my toolbox along side my OLS. Pretty sweet for a free PCB!