It's funny to see my idea was already invented.

I just tried a first setup with my new steelballs and the plastic hose and the magnets.

hose, cause if I put a magnet outside some of them aligns in zick-zack
curves, like  .`.?.?.?.?.?.?. 
sothis is happening cause the hose diameter is too big versus the 10 mm ball
diameter.
Also the hose has pretty big friction onto the balls.

I use a magnetic field that is less strong. Don't be greedy, just allow gravity to do the work. 

This is also interesting. Not very clear, it uses the smot's vertical area.

(http://www.overunity.com/index.php?action=dlattach;topic=291.0;attach=3905;image)

All,

Try the following arrangement of balls, magnets, and springs to make a reciprocating gauss rifle.

<----

////_____00[]0_____00[]0_____00[]0_____00[]0___0_\\\\


When the ball reaches the far end the spring will send it back in the opposite direction:

---->

////_0____0[]00_____0[]00_____0[]00_____0[]00_____\\\\


The ball will have accumulated a lot of kinetic energy. The spring at the end will return most of that energy and send the ball in the opposite direction with more force than when it initially started, force that will continue to build.

How do we harness this energy? The steel balls will have remenant magnetism for a brief period of time after leaving each magnet. What happens if we place inductive coils beside or underneath the track?

I tried this before, in various ways.  It doesn't work because you need the initial steel ball to collide with the magnet. 

Referring to msurucu's drawing, as the initial steel ball (left) approaches the magnet, the field increases its speed giving it more energy.  On the right side of the magnet, the second ball just transfers the momentum but the third ball is far enough away to only feel a small effect from the field.  When the initial ball collides with the magnet, it transfers the increased momentum through the second ball and into the third one, which is weakly attracted and goes flying away - what energy lost in the momentum transfer is small compared to what the magnet gave the initial ball.

When you insert a forth ball like in msurucu's drawing (which would be the second ball on the left), the initial ball never gets the speed boost that it does when allowed to collide with the magnet.  So when it strikes the added ball on the left, what little energy that it did gain is attenuated in the momentum transfer and the system fails. 

I also tried this idea with electrostatically charged balls.  Unfortunately, macro-scale electrostatic forces are so much smaller compared to the magnetic force.  To get the equivalent you need so much voltage that it will cause breakdown on the balls and you'll loose all the charge. 

I abandoned this idea

Hmmm... May someone please give me an explanation of why it should NOT work?

I need to understand what prevents a gauss gun from becoming a perpetual motion machine.

Inspiring video:

http://www.youtube.com/watch?v=ZP6_vXus0F4

I see now... The steel ball that will be "launched" needs to be at a certain distance from the central magnet, in order to the magnetic attraction force that is getting it stuck to the central magnet be weak enough to not prevent it from being "launched".

I understand now... It needs to be unbalanced. You can't put the same number of balls at the two sides. You need always to have more balls at the side where is the ball that will be "launched" than in the side of the "trigger" ball.

So it will never work in a closed loop. Let's forget the gauss gun...

Simplified the design. 

Hi,
As the magnetic ball falls, the idea is that it's slowly moved out of range of the ferrous metal.
I ordered some small balls at Dealextreme to do some simple experiments with a seesaw.
This is still just an idea, the final design will probably look completely different. Maybe i'll need to design on that has multiple stages or something.

For the one with the bar magnet and steel ball i need a long bar magnet that sticks out well above and  below the contraption so the ball won't stick.