Room Management System – 74HC595 shift register noise discovery

I have been banging my head on the wall for nearly a week now trying to sort out some instability problems. A couple of days ago, late in the evening I found a small mistake in my soldering. After making about 100 soldering point in a row I most have gotten a little tired and didn't solder the data pin properly to the board. Fixing it solved the problem for this evening. The next evening, - the problem was back. Reading through a lot of forum posts and rechecking the data sheet of the 74HC595 I found that this particular shift register in known for noise problems. Nothing to worry, we just need to add a few filter capacitors and the problem is solved. What I did is the following. The original 0.1 uF capacitor connected to the latch pin of the first shift register of our chain, going to ground has to go. This capacitor which is even used in the Arduino tutorial “Serial to parallel shift out” is preventing the latch pin to go HIGH fast enough before you bring it back down again.

The next step is to add a 0.1uF capacitor between VCC and ground on each shift register as close to it as any possible. Step number 3 is to add a 10 uF capacitor in the power line as close to the power source as any possible.

The last two images are only to show clearly what I have done. The following image shows it in the experimental setup with all the components as far as it is developed. 

That also brings me back to the analogue smoothing algorithm from Tom Igoe. If you didn't take it out yet and you want to go with it, please feel free to leave it as it is because there is nothing wrong with it. It was just co-instance that the stability problems where starting to occur after its implementation.