Do you have problems with GROUND?

[Stormrider]

When I test the buttons on the control panel, sometimes I can see that if I press a button and keep it pressed, it blinks. I thought it was a bad soldering or a wire which is not properlly installed. However, I've found out its a problem with ground. If I touch the metallic shaft of the stick with my hand or if I put both my hands resting on the wooden board at the same time, the buttons work as expected, so they work all the time when pressed. That's the reason I don't notice wrong operation during gameplay, since my hands are touching the panel. I have attached the end of the ground wire to a  metallic part of the base of the stick, and also touches the metal base which closes the stick underneath. Shouldn't that be enough? Why does the pcb still needs my body to carry the ground to the floor? Could it be because the antislippery rubber feet of the box prevent this from happening. Finally, what could I do for the new panel I'm about to make in order  to be sure that ground will be working all the time? The only thing I can figure out is taking the ground wire out of the box, but that's not a clean solution.

[surface tension]

You shouldn't have your 0v wire connected to the control panel and joystick base. It's not an earthing wire... it's a return for the 5v for the microswitches.

Try removing the connections to anything other than the switches.

[NoOne=NBA=]

Assuming that you used single chunk of wire to do all the grounds, the ground wire should start at GND connector on the encoder, connect with every switch, and then end back at the GND connector on the encoder.

This creates a huge loop, and makes it so that any single wire break will not affect the entire system.
You will need two breaks for anything to quit working.

I've actually seen games where an op ran dual ground loops.
These require a minimum of FOUR breaks before you will lose any controls.

[lordtodd75]

The only problem I have ever had with ground is that sometimes when I fall I land on it....it hurts!!!   

edit: added awesome smiley to my witty comment

[Ginsu Victim]

Assuming that you used single chunk of wire to do all the grounds, the ground wire should start at GND connector on the encoder, connect with every switch, and then end back at the GND connector on the encoder.

This confuses me. I'm using a Mini-pac with the ground harness. I connected the GND wire to the harness, then ran the ground to each microswitch. When I ran out of switches, I just tied up the rest of the GND connectors and tucked them out of the way.

What are you saying? I don't see anywhere to connect it back to the encoder...

[Stormrider]

Yes, the ground wire starts in the pcb in a ground point and then goes from switch to switch. When I reach the last switch, I didn't know I had to take that end and wire it back to the origin. I don't understand why. I'll try that now and see if it works.

[Ginsu Victim]

You don't HAVE to wire it both ways. Here's this:

[SavannahLion]

Yes, the ground wire starts in the pcb in a ground point and then goes from switch to switch. When I reach the last switch, I didn't know I had to take that end and wire it back to the origin. I don't understand why. I'll try that now and see if it works.

The idea behind it is to alleviate or reduce the likelyhood of a particular kind of failure from occurring. Looping the ground wire creates two potential paths for the current to travel through. If a single break occurs at any point in the wire, the current will take the alternate path to ground. I can't imagine a break in ground being instantaneous, a more likely scenario is that a break in ground would cause intermittent behavior. Intermittent behavior is often the most difficult to diagnose. When a failure does occurs that disable buttons, it makes it easier to deduce where the break occurred and make the necessary repairs.

Here are some crappy pics I drew to give you a better idea on just how the buttons would fail under specific circumstances.

I think the assumption here is that your ground circuit isn't complete for some reason. In other words, there's a break in the line. When you make contact with the metal components of your cab, you're bringing the line potential down to ground which gives the appearance of the buttons working again.

[SavannahLion]

It doesn't. Here's this:

That is a very basic wiring diagram intended to get you up and running with a minimum of fuss. It doesn't encompass other valid wiring designs.

[Ginsu Victim]

I understand the idea of running GND from both directions, I'm just showing that you don't HAVE to.
I'll be changing my setup to run GND both ways, though. I hadn't thought about it until now.

[NoOne=NBA=]

SavannahLion was right, I am assuming that you don't have a good connection on your ground wire.
If you can make ANY difference in the circuit, something is drastically wrong.
You have copper wire that is supposed to be making a connection from the switch to ground.
That should come up as a dead short, vs. your very high resistance.
If the circuit has higher resistance than your body, something is not hooked up right.