/diy/ - Do It Yourself

>>1406372
figuring out how you fucked up is part of the fun

>>1406430
it kicks ass because you can use the same stamping dies, mold inserts, EDM electrodes, etc. for every position
>one huge pad
which requires special consideration in paste application and reflow. in real volume manufacturing, consistency is king

>>1406492
thank you for your tracking info

>>1406556
3/10
first, traces are way too thin for a high-current load. there are trace current calculators out there that will tell you the max current given a width, copper thickness, and allowable temperature rise (keep it to about 10°C). in fact, one is included with KiCAD
also remember to use through-hole component pads as vias when needed, and feel free to route thru them
U3 (U? it's a connector, call it J or CN) pins 3 and 5 could be routed outside the pins (Pic related, blue). now you can route other traces topside, without crossovers (light blue)
move D1 west of the M1 connector, now you can route it all topside, without crossovers (sea green)
wrap the reversing switch around M2 pin 2 and wrap the common pin of "U2" through the big space (light yellow)
boom, now it's a single-sided layout, which you can run single-sided to save cost, or duplicate onto both sides of the board for added current capacity
another note: if you can't meet some criterion any other way, and you can do so, it's often worth rearranging your input connector pinout to accommodate your layout

>>1406567
>he believes the sharp corners meme
wew

>>1406620
better, but that acute angle near "U3" pin 3 and "U1" pin 4 is bad practice. it can trap etchant on particularly low-tech processes. for the first, just route thru the pad, or route +12V between the coil terminals as in Pic related. for the second, use a T junction