/sci/ - Science & Math

bump with a little picture

last bump; still hoping that someone in /sci/ actually knows something about this stuff

We use ATMEGAs at work. We use the avr-gcc toolchain, which has good documentation. Also the datasheets and any other docs from Atmel. I'm not too sure about any books or tutorials other than what you'd find on Google.

>>996921
thanks for the input!
i use eclipse+gcc+avrdude, which works pretty nicely, but all my programs pretty much look like hacks and i was searching for guides to improve my programming style/make my programs more efficient

>>996947
to be more specific: everything i do looks like a hack and pretty ugly and although it works most of the time i want to change that fact

I know stuff about what you are trying to do but I've never read any books on it, I just learned it all by making shit for the last 15 years. PCB design is mostly just a big logic puzzle unless you are dealing with very high frequencies, just make sure the high current traces are wide enough and as short as possible etc.

pic related, a custom arduino board I've made a couple dozen of with an onboard RTC and EEPROM that is 33% smaller than a standard duemilanove board.

>>996959
cool stuff!

If you're looking to get into microcontrollers check out http://arduino.cc
I'm not sure what your skill level is

>>997001
well i designed the board i posted for myself, soldered it and programmed all libraries.
as i already said; the problem is not really the lack of knowelegde, but that it is mostly inefficient and ugly.
also my buddy already has an arduino mega128, pretty interesting concept to code stuff for a 8-bit controller with an objective-oriented language instead of C.

>>996947
That sounds more like a general programming question. The best thing you can do is have a deep understanding of C and read a lot of code. Reading and understanding code probably did more for me improving my style than anything else. Once you get something working, presumably after banging on it with a hammer for 3 hours, figure out why it works and why it didn't work before it started working. IOW, just because something works, doesn't mean it's done.

Also, just because it isn't done doesn't mean it's not "good enough". One of our programmers recently quit, and his code was all hacks that just barely worked, and the new guy has been doing essentially nothing but rewriting old guy's code so it doesn't make our eyes bleed. We never made a huge deal about it because this code was never going to a customer and the fact that it worked was more important than the fact that it was ugly. New guy has been making a HAL layer with preprocessor macros (the code runs on several boards), and reducing unnecessary dependencies.

>>997091
thanks, but what do you mean with "deep understanding of C"?

I mean you know the quirks and idioms.