Problem w/ External Power Supply

I am having a problem with an external power supply that I am using with one breadboard project.

I am connecting my power supply as follows: +ve 10 volts to the hole indicated by the red arrow -ve to the ground strip on the breadboard

There is an lm7805 onboard.

When I do this, the LED marked by the green arrow lights up as it's supposed to. This indicates 5v presence on the breadboard. The problem is that my sketch doesn't run as supposed to, in this case it is simply a variation of the blink sketch.

If I connect a battery pack (~4.5v) directly to the breadboard power strips, then the LED marked by the green arrow illuminates, and my sketch also runs properly.

In order to rule out the possibility of a faulty power supply, I have tried the following setup on another project.

The blink sketch runs as it should.

Please guide me in troubleshooting the interface between the power supply and the first project. The lm7805 chip I have used on both breadboards is the same.

well, 1 thing i noticed is the lack of capacitors around the 7805, could be getting all kinds of noise which would mess with the cpu

Yes, I understand about the lack of filtering.

I have already ordered components for the following:

This is still puzzling though, as the circuit works fine on the second breadboard project, but is problematic on the first.

Some decoupling around the AVR wouldn't go amiss either. Digital noise from your circuit can wreak havoc with the processor. A couple of 100nF ceramics close to the 5V and Gnd connections either side of the chip should do it. Since the second circuit has a real Duemilanove, it will already have the decoupling in place.

I think you've hit the nail on the head pluggy.

The duemillanove does indeed have decoupling on-board, a detail which I missed, and this is probably what's allowing the power supply to work with the second breadboard project.

Thanks, to pluggy and osgeld for their fine assistance.

I have ordered components as per both their replies, and will post back when I have implemented their fixes.

The power supply capacitors fixed the problem.

I also used one 100nF ceramic capacitor between VCC and GND, not that it had a discernible effect as the power supply capacitors were the solution.