On Mon, 2007-11-26 at 02:51 -0500, DJ Delorie wrote: > Spent the day building two r8c-based boards, thought I'd give a quick > status, since I did some new things with this. > > First off, I've got a board that has a USB chip (FTDI232R) connected > to an R8C/20, connected to a bunch of headers. In other words, it's a > programmable USB GPIO port. The only design flaw so far is that the > mini-B connector has an extra pin, and ground is on pin 5 instead of > pin 4 like all other USB connectors. Fortunately, pin 4 is not > connected on the device end, so I just shorted pins 4 and 5, and it > worked just fine. Haven't talked with the R8C yet, I need to adapt my > R8C programmer to the FTDI driver. This board also used the global > puller, as I was able to coerce it to give me useful results. > > Second, I've got an R8C/27 (TQFP-32) adapter (files are at > http://www.delorie.com/pcb/r8c-27-adapter/). Along with being a TQFP
The closing bracket breaks Evolution's auto-hyperlink thingy (tech-talk for software) -- Greg > to DIP converter, it includes some pullups and the oscillator, so > aside from programming you shouldn't need any support parts. This > board takes advantage of the ability to rotate parts 45 degrees. > > Now, for the new things... > > The DIP adapter is a two-layer board. However, I did it using two > separate SS boards, taped together. This lets me line up the two > etched boards after etching, instead of before etching. The clad is > 16 mil (1/64") with 1 oz copper, resulting in a 36 mil board. The > tape is 3M 7953 laminate adhesive, about 4 mil thick. > > The USB board is a home-etched four layer board. Yes, four. The core > is a 32 mil DS board, where I lined up the resist before etching. > What I did here is iron on one side, drill out the alignment holes, > and use those to line up the resist for the other side. This core > held the ground and power planes. Then I taped the outer layers to > it. Those were 8 mil 1/2oz SS, and are pretty much transparent, so > easy to line up and tape on. > > The tricky part is the vias. What I ended up doing is drilling big > holes in the outer layers before taping them on, then drilling small > holes through the full stack. That left some of the inner layer > copper exposed through the holes in the outer layers. For signal > vias, I only did the small holes, like a 2 layer board. If I do more > of these, I'm going to have to add some support to PCB for it, because > I mis-drilled a lot of the holes. For example, to connect a top > signal (layer 1) to power (layer 3), I had to drill a big (35-45 mil) > hole in layer 4, and no hole in layer 1, then a 14 mil hole through > the stack. To connect it to ground instead, I drill a 22 mil hole in > layer one and no other holes. To connect both sides to an inner layer > means drilling a 22, a 25, and a 14 mil hole. Plus, it would have > gone better if I could have had different size copper (or no copper) > on different layers based on connectivity. > > I had to print each layer separately, because PCB wouldn't let me > paste two boards into one layout and flip one, which would have let me > etch both sides on one board then cut them apart and tape them > together. > > I also tried using glossy magazine paper instead of my usual coated TT > paper. I wanted to use some of my 64 mil (1/16") clad that won't fit > through the laminator, too, so I ended up ironing it on. The paper > rubbed off like "they" said, but I was back to the original problem of > getting the toner to stick completely, without getting so hot it > smears. I think the paper might work in the laminator though, and I > have more clad than coated paper at the moment. > > > _______________________________________________ > geda-user mailing list > [email protected] > http://www.seul.org/cgi-bin/mailman/listinfo/geda-user _______________________________________________ geda-user mailing list [email protected] http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
