First , can you confirm the color of the wires of the psu connector going in to the motherboard ?

yes, the color of the wires are the same as in your picture.
as i stated previously, i took out the transformer
i'm considering bypassing that part of the psu and feed the voltages directly from a standard atx psu

What do you get when you measure in continuity/diode mode that point to psu connector ground (black wire) and 12 v (yellow wire) ?

to ground direct short = 0 ohms
to 12v rail = 22.4 ohms but the value goes down slowly over time

Can you desolder the coil , but only that end and lift it up a little so it wold brake contact with the board ? And then check on which side you have the short to ground ( the other contact of the coil to ground or at the psu connector) ?

ground to coil = 0 ohms (no connection)
ground to point in the board = short

now there are no shorts in between the 12v rail and ground

Can you confirm the lane in the pic ?
If so , do you still have the ram and the cpu on the board ?

yes, it's shorted to ground
i have the original processor and ram (celeron 533mhz and 128mb pc100)
also i have other celeron/pentium 3 processors for testing and many pc100 and pc133 modules

If they are installed on the board (ram and cpu) can you remove them one by one and check the status of the short ?

mother with everything = short
mother with cpu = short
mother with ram = open
mother "empty" = open

so, it seems that the short is in the cpu socket or some related component...

Ok , now with nothing installed on the board measure between '1' and '2' for continuity while the pin of the coil is still not connected .

short but the two are 12v

So you have a short between '1' and one of the 12v pins ? If that is correct , that should be the vrm stage for the cpu and it's compromised ('feeding' 12v in to the cpu). Lift both pins of the selected components and test/check for short.

okay, i'm gonna do those tests later
if 12v went directly to the cpu, it may have killed it...

Not always, theoretically there is short circuit protection , over current. You have to test the cpu on a working board.

well, i've lifted the legs in both components, now there's no short to be found
maybe now the cpu would not get any power?

Keep checking that stage , it's not normal to have 12v just pass trough it. (Your initial 'short' to ground on the 12v rail was the cpu , if you solder everything back(including the coil/inductor) and put the cpu in the socket , the'short' on the 12v rail will be back).
Also did you check for short between all the 3 contacts of each component ?

Any luck finding your faulty component from that stage ?

sorry, i was a little bit busy
the big one is a MOSPEC S20s40C, schottky barrier rectifier, it has no continuity in between legs
the small one is a Phillips Semiconductor PHD69N03LT, N channel MOSFET, has continuity in between legs
in both components i've tested in every way possible that I know (reversed polarity,using the board ground)