You can use any piece of metal with the shape of a cube (or dice), the only requirement is that it needs to be made of metal because is going to be soldered... and also his proportions needs to be very well adjusted
Right now... the most accurate way to do it i figure is by using a copper sheet with 03 thickness and cut it in cubes of 03x03x03. This way there are 2 sides of the cube with the exact meassure (mechanized by the machine that made the copper sheet in a industrial process)
I have a theory of what can happen with the use of "pillars" taller than the original, I don't know if it makes much sense but here I am.
CPUs further away from the board with the same BGA size, as mentioned, due to the stress of temperature variation, these BGA's may crack, that "extra" height added later may occur, when the solder ball cools down, it should not be able to contract enough and be "pulled" both up and down, leading to a crack at one end of the contact.
@vyktormvmpay25 are these pillars not removed along with the old weld when reballing? in the PS3 reballing videos I had never noticed this detail.
When the BGA ball gets solid is not an sphere anymore, is mostly like a "fat pillar" (with a belly at the middle all around)
If you ask me to choose in between 2 extreme options (outside of the aceptable ranges) for the shape of that "fat pillar" i would choose to reduce the gap the most posible
You know... if we increase the gap a lot the BGA balls are going to solidify like "thin pillars" and are going to be very fragile
All this depends of the original diameter of the BGA ball of course because his diameter indicates the amount of material
And also it depends of the gap... personally i would try to reduce the gap (vyktor said is 0.3mm so i would try to reduce it to 0.2mm), but also i would try to use balls with a smaller diameter (to reduce the amount of material)... and before the solder solidifyes i would "press" it on top to squeeze the balls the most posible, lol
But this is the kind of experiment i would do in scrap boards only, because i think is a good idea but im not completly sure
You know... when the solder gets fluid everything goes a bit out of control, like trying to control water, you dont know well how is going to move, is the kind of thing we can theorize as much we want, but i think even the professional engineers would need to do some tests in a laboratory with a microscope recording video in HQ to see how the solder ball changes his state in between solid->liquid->solid to understand well where are the limits