Homemade arduino fails

I bought just the main parts to save money from this kit: Breadboard Arduino Compatible Parts Kit Retail - RTL-10422 - SparkFun Electronics I was able to assemble and test it by using an UNO board to upload sketches. The version on the bread board worked without any problems so I decided to transfer it to a proto board. Since I am using a DIP socket, I can pull my atmega out to program it so I decided to omit the FTDI parts. (The breadboard version had these omited also) So the setup would be a 5v regulator with the filtering caps, crystal with capacitors, an LED with resistor on pin13, and a reset button with resistor. For some reason I can not figure out why this set up will not work. I checked the regulator, 5v is going to VCC, AVCC, and GND, GND are grounded on the atmega. The capacitors and crystal are wired correct, the LED is wired correct (I tested it with external power while not connected to the arduino) and my reset button is wired correct, 5v when not pressed goes to reset pin, 0v when pressed and it is held high with a 330ohm resistor.

In my frustration, I pulled it out, put it back into the UNO board and it worked fine, pin 13 blinks like it should. The only thing I can not test is the crystal which is the only thing I think can be bad. When it is on, I get 0 volts on Clock 2 pin and about .38v on Clock pin 1. When I have the the atmega in the uno board, it is about 0.46 and 0.86 on those pins and it works fine. Is there any way I could test this? I do not have any spare crystals or 22pf caps.

schematic and picture would help diagnosis

held high with a 330ohm resistor.

This is too small, 10K is the smallest resistor recommended in the data sheet.
It also needs a 0.1uF decoupling capacitor on the chips power rail as close to the chip as you can get it.
But like you said it could be the crystal or the crystal's caps.

Check your wiring, and check that the caps connected to the crystal pins really are 22pf if they are not the exact same ones you had working on the breadboard. If those caps are too large in value, the oscillator won't work. OTOH it may work without the caps connected at all.

I don't have a camera right now, but I will describe everything best as I can. I used the schematic and assembly instructions from here and omited the FTDI components since I am using a dip socket. (I can pop it out and plug it into an UNO to program it)

I uploaded a sketch to make an led on pin 13 blink and it worked 100% on the bread board. When I moved it to a protoboard, it quit working.

I removed all of the components and assembled the circuit "in the air" with no sub-straight. It worked fine.
Then I put everything back on the proto board and it failed again. I double, triple checked my wiring and tested the components individually and they work.
I removed the ATmega and left the reset button, led, crystal and power supply (and all applicable caps and resistors) on the proto board. The circuit works fine with the ATmega NOT plugged into the dip socket. So naturally I took a closer look and a volt meter to see if something was not making a connection and everything looks fine. It seems to have something to do with the dip socket, I just need to have it since I do not have an FTDI cable.

My guess is that you've got a short on your protoboard. I suggest you examine it under a magnifying glass looking for solder bridges between pads, also check (with the processor removed) using a multimeter that there is infinite resistance between each clock pin and everything else.

Did you check your reset button? Is is holding the reset pin low by any chance?

I checked the dip socket again. The two pins that would be the serial pins were shorted. There was no visible connection though and it was in the megaohm range. Couldn't get it to go away and destroyed the socket trying to get it off. Everything worked fine until that dip socket came into play. I am rewiring another up with out the socket to see if that will help

I got it to work. The dip socket was some how causing the trouble by shorting the bottom right two pins out. I couldn't see a physical short though but bypassing it solved all of the problems. phew!