Warning! On my way to work this morning I realize this is probably wrong. If you use this on a "dark" board and white stencil the colors will be inverted.
I'll have to investigate if it will still work with inverted colors, or if anybody knows please tell.
I think JTR deserves to be mentioned as well. He has found and corrected a lot of bugs while I have almost dropped the ball lately due to lack of time.
As i was preparing some boards for manufacturing (perhaps a 7400 entry) I thought it would be fun to partake in Itead's OpenPCB program. But that requires me to have some identifying mark on my board and I thought a QR code would be perfect.
The bmp import in eagle as some room for improvement so I skipped that and wrote my own. As I was writing I found out that the latest eagle (5.11.0) can do http get/post statements in ulp. So here it is, dynamic pcb qr code generation with the help of Web 2.0...
Hope somebody will find it useful. No rights reserved. It's just a little midnight hack, so expect bugs.
Isn't there some sort of zeroconf in the microchip stack? It does have some device discovery techniques.
Zeroconf/IP4LL is basically to choose a random IP in a specific /24 and then announce it via broadcast/multicast. If another host already has taken that address it will respond and the first host chooses a new random IP.
The downside is that with IP4LL addressing all (or at least some important ones) devices on the LAN pretty much has to support mDNS, multicast DNS
If you need a implementation of a pid controller, somewhat more elaborate but still not over engineered I read a very nice writeup by the guy that rewrote the pid controller for arduino. He made some excellent explaining without diving into mathematical symbol manipulation. Too bad I can't find it now, but I believe there where as post on hackaday.com
The few USB Device firmwares that I've developed with more than one Configuration managed to confuse Windows so much that the OS refused to work at all.
I hear you. Try making a hid with a joystick/helicopter combo as different configurations and you'll have fun bluescreening anything that came out of redmond for the last decade.
Try to take a voltage reading with your favorite voltage meter of the 5V USB supply voltage. Both when "idle" and when transmitting. If the voltage drops during transmit, it could be due to too much power draw, paired with a maybe not so low usb supply impedance.
If it drops really low, like .1V or something, it could be that your laptop really have some power management on the usb supply.
Also, try doing it with the charger for the laptop plugged in (I have no idea if or how this would affect the USB supply, but I can't hurt to try)
Unless he has some really cool piece of Operating System paired with an awesome motherboard a compile with a higher current consumption declaration in the descriptor, wouldn't probably change anything. At least haven't I seen anything other than some USB hubs complaining about too much power drawn (just a LED indicating that you should power the hub separately). Most mb and or OS'es just assume you are going to connect something like USB vacuum cleaner or charge your iPod-wireless-telephone through it and just ignores (within some limits) how many milliamps you declare or draw from the ports.
On the other hand it is in the standard and we are in a position to do it right, so maybe we should...
The reason for the rounded inside corners is that I imagined it milled when I designed it. By increasing the clearance parameter in the scad file it should accommodate sharp corners. I have to check when at home. No OpenSCAD in android market yet...
