Alright then. I have access to a hot air station. Where would I obtain solder paste and an appropriate syringe?

I have looked at many design recently--like all the Arduino clones on Kickstarter and pretty much any prototype circuit--they either use breadboards and Arduino shields or complete SMD components. I understand that SMD components are used because they are smaller and cheaper DIP components. I also completed a board recently where the cost of the PCB was more than all the components combined. That's a little ridiculous. I want to move to SMD parts. However, I am at a standstill.

It is difficult to hand solder even a relatively large SOIC-24 even with a nice Hakko-FX888 with a curved tip, although my solder was pretty thick. I tried to flux the pads and drag solder, but I seem to create many bridges and removing even just one is miserable. For this reason, I want to move to non-iron based methods such as the reflow skillet with solder paste, however I do not know where to buy it. I heard that it must be kept cold or it will lose its chemical properties, and I hear it is applied with a syringe. Is there a tutorial and accompanying product that I may buy to hopefully solder TQFPs 32s, LQFP 100s, and maybe QFNs?

If I understand correctly, to use the JTAGICE II or up, you have to use it through Atmel Studio and there are no cheap clones available. What application would we use with our own JTAG debugger to debug with the Arduino IDE? And you need all source code in order to debug, correct?

I am very excited to finally see the Due in the wild. I am eagerly looking for its soulmate, the JTAG debugger. What are our options for debugging 96kb of code? Can we use low-cost debuggers like the LPC-LINKs or are Atmel's programmers like the JTAGICE II our only choice?

Interesting..so in series with the cap. That would limit the rate of charge and discharge to about 22mA max. I wonder why that fixes the problem.

I would recommend changing the baud rate on the bluetooth device, otherwise other boards of whatever entry you changed will not work without more modification to boards.txt. I am not sure about less distance at 57600--you have to determine this empirically.

UPDATE:
I added a 220Ohm resistor from -cap->GND.... AND IT WORKS! no more surge problems!

220 between 5V and ground? Or somewhere else?

[in hardware]

You will have to use another, higher voltage power supply with your own regulation to reach the stage before the 5V. I used 9V. (this might solve the issue by itself without the giant cap)

The phenomenon we are experiencing resets the Arduino without letting it run any setup code it seems. You need to enable the WDT in hardware so that it is always on by setting the WDTON fuse setting on the ATmega328. You will need an ISP programmer like the AVR ISP mkII for this (preferred) or you can use another Arduino, but the ISP programmer is much cleaner and quicker. After you find your programmer, you can either use the Atmel Studio 6 "Device Programming" panel or just command-line avrdude (quicker but easier to brick your chip). Then the WDT will always be on for I believe 64ms, which you can change in the first line of your setup and throughout your program.

I tried connecting a 2200uF capacitor 5V>GND, but this did not solve the problem.

I replicated the problem by plugging in a fridge, a microwave, and a water cooler into a power strip and then connecting the arduino usb to the other plug in the same outlet. When I flip the power strip on/off, the arduino eventually hangs and loses connection.

This is a really annoying problem. :/

I think that is going a little far. You only need to simulate with the appliances that will be used there. (and the microwave does not use any energy when plugged in)

I put my cap across the power supply output, not the 5V. Energy stored in a capacitor varies quadratically with voltage. My circuit ran for three days and never reset. Even if it does, I turned on the WDT for two seconds as a redundant safety measure.

If you have already verified two-way communication without a USB cable while not on the roof, then it should work on the roof unless a wire comes loose or the asymmetry between Bluetooth transmit power causes one of the two signals to degrade.

I do not know why your computer is locking up while trying to program over bluetooth. If I had to guess, it is getting stuck at the connection establishment, something a USB cable does not have to worry about. I would just end the Arduino process or disable the bluetooth adapter in device manager instead of rebooting. Again, I do not recommend programming over bluetooth. The JY-MCU already provides an easy to use serial connection but was not meant to program.

Also when I said "up arrow (on either RX or TX)", I said "either" because I could not remember which one it was. The pins don't do the same thing.

You seem to know what you are doing so I trust you can troubleshoot.

1) Try using a very simple sketch that just outputs a message every second with just the Arduino first. Program it with a USB cable, then remove it from the computer and attach the JY-MCU so that the up arrow (on either RX or TX) goes to TX (pin 1) on the Arduino. If you connected it correctly then you should be able to see your data.
2) When you pair and open the serial port (you have two-either one should work), the LED should shop blinking and be solid on. This means the serial port is open and you should be seeing data. You cannot send AT commands in this mode. If you want to send AT commands to change baud rate and name, follow steps in post #3. You cannot upload sketches in this mode unless you press the reset button at the right time. Uploading sketches requires a reset wire which the JY-MCU does not provide. I made an ATtiny circuit to watch the LED and reset the microcontroller so that sketches can be uploaded, but uploading code is five times slower and sometimes it fails. It is flawless if it actually finishes an upload though.

Give this one a try.  It looks like the error trapping is more complete...
http://arduino.cc/playground/Main/DHT11Lib

Thanks for this suggestion. The new library works exactly like it should and has better error reporting like you said. I was able to create my nice ATtiny thermometer with seven-segment display.

Thank you Tom Carpenter. This was an easy fix for my >4k sketch.

Hello there to everyone who has experience,

I am trying to use Arduino-tiny for the first time. I did this so that I could use the Adafruit DHT11 sensor library. I have successfully implemented the example sketch on an Arduino Nano (mega328) However, when I connect the one-wire sensor to the ATtiny, it reads as zero. I know I have the pin right because when I disconnect the wire I get "Read failFailed to read from DHT". I am running at 8MHz internal oscillator.

Can someone look in the DHT library code read() function and see whether anything would run differently on an ATtiny?

Serial output:
Humidity: 0.00 %   Temperature: 0.00 *C
Humidity: 0.00 %   Temperature: 0.00 *C
Humidity: 0.00 %   Temperature: 0.00 *C
Humidity: 0.00 %   Temperature: 0.00 *C
Humidity: 0.00 %   Temperature: 0.00 *C

You have to use data compression and trade some CPU cycles for your memory. Avoid long strings of text and large arrays, or decrease resolution. For example, in the SAE J1939, many different values need to be represented using only 8 or 16 bits, so there is a different multiplier for each type of value. If temperature goes from -256C to 256C, then you can store it as an 8-bit value and multiply it by 2 instead of using a 16-bit value. The tradeoff is that your resolution is now 2 degrees instead of 1.