I don't either, but I said "that will be easy, just edit the pcb svg in core and look". Wrongo! there doesn't appear to be an svg in the core svg directory (or I don't know what it is called). I suspect they create the svg on the fly in the code. After some head scratching however I prevailed. I exported you pcb (with jumper) as an svg and then edited that with Inkscape which tells me the pad is .075 inch diameter which likely makes the hole .0.35 (the stroke-width isn't the standard 20 but 1.44 which is why the probably on the hole size). As to the bypass capacitors indeed all the bypass capacitors are just caution, more needed on the LCD board than the main one, because as you say the CPU will have internal bypass capacitors. The primary issue here is wire length from the source of power (which is going to be longest in the LCD case), as the wires add a small amount of inductance (which resists changes in current by changing the voltage, where a capacitor resists changes in voltage by changing the current). What happens is that if multiple data leads change at the same time and the inputs (which is likely true of the LCD controller) have a high capacitance, when they switch (and depending on how fast they are switching, which with modern logic can be very fast) they draw a lot of current. If the power lead is long and with enough inductance, that will cause a downward power spike on the power line to the LCD (and thus the LCD controller). Without the bypass capacitors the inductance of the wires may temporarily drop the voltage see n at the LCD display. If it drops below about 4 to 4.5V for a 5V supply the LCD controller (or most any other logic chip) may malfunction, and since the LCD controller is essentially a CPU inside, malfunction may be stop working until it is reset. The bypass capacitors near the LCD fix that by resisting the voltage change caused by the increased current draw by supplying the energy required from the energy they have stored in them until the inductance finally allows the voltage to change (which it will quite quickly, just not always quickly enough). Within broad limits (the notable exception being rf which is much more complex, although this isn't all that easy Image may be NSFW.
Clik here to view.
) more bypass capacitors are better. The usual rule of thumb is a .1uf ceramic capacitor close to the power leads of each chip and larger tantalums every 4 or 5 chips. If you don't use them it will likely work just fine, but its possible that it will work fine %99 of the time but every once in a while the system will hang until you power cycle and reset everything. That is a very hard and frustrating problem to find and fix (and is expensive if you have cut and assembled a bunch of boards and it turn out to need more bypass caps!). As a result I tend to stick extras in.
Peter
