On 30/10/2010 15:13, Olin Lathrop wrote: > Pay attention to details. OK, all of these points are very valid. However, I do this as a hobby=20 and relying on Eagle for some parts saves a lot of time! I will take=20 more care to smash the parts and fix labels in future, I tend to spend a=20 lot less time than I should on the schematic. > Yikes, you've got silkscreen overlapping pads all over the place! Yes, the silkscreen has issues. Turns out I exported the tDocu layer=20 along with tPlace. I will make sure I am more careful when exporting=20 the gerbers for the board house. (Note that the Eagle USB connector still has silkscreen all over the=20 pads on tPlace, I'll fix that up) > You apparently didn't use net classes at all, since everything is in the > default class and its width and other values were never set. I had not heard of these, so some research is needed for future. Thanks=20 for the pointer. > You also clearly didn't run a DRC check, which would have found a number > of clearance violations. Actually the board I uploaded passes the Sparkfun/BatchPCB DRC just fine. > You do need to get used to making your own parts and definitely > not rely on the Cadsoft libraries, but you have to learn how to do it rig= ht > first. Several of my libraries are available in my Eagle Tools release a= t > http://www.embedinc.com/pic/dload.htm. I have made a few parts so far but have a lot to learn, clearly. I have=20 grabbed your tools and will look through the libraries for inspiration. I am hoping that the new release of Eagle with XML libraries will make a=20 big difference to how easy these things are to share/modify (even just=20 importing a good footprint from one library to another is a pain). > The bypass cap for IC1 is poorly placed. Noted. > As a challenge and to demonstrate some of the things I am talking about, = I > edited your board and schematic. The route was a mess, so I ripped it al= l > up. Somewhere in this process I did something wrong or Eagle got confuse= d > about the GND net connections. Even though I deleted the GND polygon fro= m > the top layer, it still thought everything on the GND net was connected a= nd > showed no airwires. That's because GND airwires were hidden (Eagle does show this in the=20 status bar when you run ratsnest). Didn't you read the manual at all? ;) > I created a ground polygon on the *bottom* layer, and then rerouted the > board. I also replaced your 0805 resistor with my own package since your= s > had that bizarre copper ring around them. The "bizarre copper ring" is not copper, it's on the keepout layer. It=20 doesn't get used except for a visual guide of where not to place components= .. The poly on the bottom layer is much better. > Note the more readable > orientation of the text. I didn't have to do anything special as I have > separate parts for vertical and horizontal resistors. I also fixed the t= ext > of T1 as a example. Yes, that is much more readable. > The oblong holes for the connector pads were just silly. Very sensible. I generally use a pin header for insertion into a=20 breadboard, hence the 0.1" spacing. I'll make some PINHDR variations=20 with round pads. > I then manually routed all the GND1 SMD pads each with their own via to t= he > bottom (ground) polygon. That leaves as much routing flexibility for the > remaining signals as possible. I routed a few obvious connection and pla= ces > where I thought the auto router might paint itself into a corner, then au= to > routed the rest. That's a really neat example, thanks. I have learned a lot creating=20 this and reading through the feedback. Unfortunately BatchPCB don't allow buried vias and the clearance between=20 some of them is also too small. A few of the vias could be moved away=20 from pads, I will attempt to do that later. I might just send off my=20 board and learn from what I get back. The few I have made before I am=20 slowly documenting on my website. Thanks so much for taking the time to look at it. David --=20 http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .