So i tried re-routing and lenghtening the wires but nothing helped. Even put the mobo on a stand so the under board wires were hanging away from the board to avoid problem, nothing helped.
I read about the 68nF cap fix on GND-PLL_BYPASS points on some forums, but i didn't have one so i used a 100nF electrolytic cap i happened to have handy and i actually got it booting quickly in ~5 seconds every time.
But, the problem was it froze randomly in the dashboard or startup logo.
So i bought a couple of 68nF metallic polyester ones to try, but with that it did the same thing. So i thought maybe its a wrong value cap so i went on to put different ones in series to lower the capacitance, to see if it helped.
So i went trial and error with the caps until i had 47nF-68nF-68nF-100nF(electrolytic), in series and then it was perfect! No freezes (in 30 minutes) and booted quick, tested many times.. This makes about 16,5nF as i calculated, but interesting here is that when i tried one 15nF one, it wouldn't glitch at all. And when i tried with 22nF one it worked but was freezing again.. I also had some of the normal polyester caps in series to make about 16,4nF and it would work almost, but froze in about 7 minutes.
So i'm thinking, could it be the one electrolytic cap helped stop the freezing, or is it that accurate about the capacitance? Or maybe there is something else at play here...
I also tried the 68nF inline CPU_RST but the had no effect on this one.
Edited by Vilpo, 22 March 2012 - 11:25 PM.