Voltage Creep on cpu?

[FireBean]

Hey everyone, Long time lurker, first time poster. I've been on Guru3d and [H] for a few years though...

I have a strange issue that I'm trying to isolate. My dumb-ass killed my AWESOME power supply. It was a Silverstone OP750. What happen is that I had the computer running, and the cover off the PSU. The cover was off so that I get some some dust bunnies hiding in there. I know that a quick blow job *giggles* would of been sufficient, but I'm have OCD when it comes to cleaning. But one of my cats though it would look like fun to play with it, so I had to put the cover back on. I touched one of the heat-sinks with the cover while it was still running. *sigh* It does not power on anymore.

So, i bought a used CM 650W from a friend. It works, but i noticed that the voltage going to the CPU is no longer rock solid. It creep upward as the CPU become stressed. It's almost linear to the amount of stress that I put on it too. At 1.35, it will hover around (1.4 - 1.41v). Now, if I overclock, and push the voltage to 1.525, it will jump past 1.6V. The highest I have ever recorded was 1.63v.

So is this a POS PSU issue, or did I happen to fry something small when I shorted the previous PSU? I don't have a cooler on the mosfets, and I cannot see anything that is burned.

Here is the new powersupply. It's no Silverstone, but it gets the job done. Cooler Master RS 700 PCAA E3

 

[bryanl]

It's not the new PSU's fault unless its +12V output is very low because the CPU core voltage is determined by a voltage regulator built into the motherboard, and I wouldn't be surprised if it maintains proper core voltage even when fed just 10V.

You have enough OCD to clean the inside of the PSU but not enough OCD to first unplug the AC power cord? There's over 300V inside, enough to kill you. The damage was likely limited to the high voltage side, like the fuse or high voltage transistors.

 

[C1]

The actual PSU voltages have various tolerances (+-5% to +-10%) : see http://pcsupport.about.com/od/insidethepc/a/power-supply-voltage-tolerance.htm

It may well be that the Silverstone regulated outstandingly compared to the CM which might be varying its MB inputs substantially. Even though, like Brayanl states, the CPU regulator should compensate. You may have just had a PSU/MB combo that worked together well, but since fried your one in ten thousand Winchester rifle (based on movie "Winchester 73" - Lin ends up winning the prized one-of-one thousand perfect Winchester rifles).
===============
http://www.ebay.com/itm/ws/eBayISAPI.dll?ViewItem&item=190577947237+&item=190577947237

 

[BonzaiDuck]

I don't think the OP mentioned which CPU, which socket and motherboard this relates to. Boards lacking solid-state VR components -- over time -- showed occasional need for readjustment -- possibly for "capacitor aging." Many of us have seen it.

I suspect the phenomenon is a "thing of the past," but we shall see. On the other hand, if the OP is using a mobo and CPU released within the last two years, my thoughts may be largely irrelevant. . .