Here's a quick project update: I've been busy with studies, work and other stupid things and I haven't had much time to spend with really important stuff - my weather station project. Now that I finally got more time and spent couple of nights building I got radio silence! Oh ****!
Currently I have no idea what's wrong. The remote unit is still on a bread board and it's been working fine in every test I've made so that's probably still ok. I have soldered main unit electronics to a perfboard. That's the most probable source of this problem. SPI comms between ATMega and RFM12B seem to be ok because I can read status from RFM12B. But when I start the receiver and wait for incoming data I get nothing. RFM12B should pull an IRQ line low when it has received one byte of data but it never happens. Same IRQ line seems to work when sending data (TX buffer empty signal) and on power on reset. I've been checking connections with continuity tester and analyzing SPI comms with oscilloscope but so far I've found no errors. I just hope I haven't fried RFM12B's receiver. It's been really cold in here (around -30°C) and means a lot of static electricity buildup and I haven't really been too careful with it (yeah, I have an antistatic wrist strap somewhere).
Good news is that I already have Twitter feed up. Now I just have to fix this thing so I can tweet something. Luckily I have a couple of spare RFM12Bs lying around somewhere (I think). I think I'll try replacing the radio with another one and see if it helps.
Here's couple of pictures of the current state of the main unit:
LCD in the box
Solder side of perfboard
Component side of perfboard
Last but not least: Merry Christmas!
Update Dec 25, 2010:
Got it! Stupid Santa stole capacitors from regulator output. I figured this out when I looked more carefully at signals between RFM12B and ATMega and noticed that they look quite noisy. When checking power rails I noticed that my caps were missing. After adding some caps everything is working again!