Magnets levitation stability

I'm willing to make a levitating stand based on standard magnets. Both the stand itself and the levitating object would have magnets attached, of course the same poles heading each other. I have two questions respecting this little project:

1. How should I arrange the magnets to make the floating object stable? I don't mean not moving at all, just keeping vertical position. The object would be rather tall (2.5:1 height:width). Is it even possible to maintain vertical stability without superconductivity? I suppose that choosing just one, flat magnet in the base and one in the object wouldn't suffice (would it?). How about three magnets in the base?
2. I was wondering about neodymium magnets for this purpose. The weigh of the object wouldn't exceed 1kg, but I think that for maintaining 5-10cm of distance, they would be necessary - am I right?

The image and this gif outline the purpose of this undertaking:

• Google Earnshaw ' s theorem – docscience Jan 30 '16 at 15:41
• You cannot reliably engineer a levitating magnetic field using permanent magnets. You either need to use electromagnets and feedback controls or superconductors. – docscience Jan 30 '16 at 15:45
• How to Build a Magnetic Levitating Top but for a bonsai, I don't know if it is enough. Anyway, it will be unstable without a complex regulation. – user46925 Jan 30 '16 at 15:47
• @docscience thanks for that. My quick research found out that there are some ways to overcome the theorem's limitations: 1. We can use a diamagnetic, like bismuth. However, I assume that the repulsion force is really low in that case (would it be enough for such bonsai?) 2. Using some supporting structure, like assembling a set of magnets so that the power cost (gravity) of flipping the structure would be superior to the force of the magnets' flipping ability thus the structure would stay levitating, like here: youtu.be/s0wjWYFlaBE?t=8m43s What do you all think about that ideas? – jalooc Jan 30 '16 at 16:56
• The most simple way to "beat" Earnshaw's theorem is by fast rotation. The theorem is only valid for quasi-static levitation, but the spinning levitating top (Levitron) works just fine. – CuriousOne Jan 30 '16 at 21:17