how much can you up a 12v rail on a motherboard before bad bad things happen?

[BobbyBubblehead]

I ask because I am over volting cpu`s and ram... board etc.

audio network experiments... and generating a good square wave

but anyhow... anyone took a motherboard 12v to the edge before?
where is the edge?


note. I have a atom D525 ready to die if need be

low power high voltage computing.... maybe the future... maybe the last you hear about it

there was some droop.
so I cranked 12.5v in and got 12.024 (varable 12v bench supply)
vcore/3.3 and 5v all stable and constant.

suppose I should use the multi to check also but.... ya know its late here and no magic smoke so far nah the pico will clip and cut out if I went too far... done that boo boo before

Edit: I even over volted the router....

 

[EarthDog]

I wouldnt put it out of spec, that is for sure. There is a reason for the spec.

How are you adjusting the 12v rail in the first place? Do you have a PSU that does that or something? Im a bit confused here at what you are getting at...

 

[BobbyBubblehead]

using a bench power supply to feed DC into a pico unit (that splits the voltage)

now theirs some "custom" noise suppresion on the DC power in... part of the experiment.
I think its generating the droop and making the pico`s job harder to perform.

so what we have here is 12v being altered and regulated by the pico... then hitting the two phase regulation on the motherboard and dropping it some more.

in essence the regulation stages are taking the 12v under desirable levels.

just got me thinking is 12v on the nose best or can we get give it a tiny bit more and not kill it long term.

 

[BobbyBubblehead]

should note this does just correct the board voltage.
all the drives are on another PSU fixed at 12v.

even the fans are on a seperate isolated power supply.

crazy experiments a go go

 

[BobbyBubblehead]

well I just had a search round.
12.8 for dare devils.

seems +/- 5% is the normal limit

so sorry to have bothered you really... just thought someone may know off the top of their head... took more time to explain than find out

 

[Bobnova]

Pretty average 12V rail OVP shutdown point in a PSU is 13.4v.
Good protections ICs cut it off at more like 13.
It's still within ATX spec at 12.6000v, 12.6001v is over spec.

 

[BobbyBubblehead]

cheers bob

hope your well me old guru clockin buddy... merry christmas `n all that

 

[larrymoencurly]

The overvoltage protection of the PSU doesn't kick in until around 14V, sometimes 15V, but I don't see how increasing the 12V rail will help with the CPU or memory because they're fed from voltage regulators built into the mobo.

 

[EarthDog]

The overvoltage protection of the PSU doesn't kick in until around 14V, sometimes 15V, but I don't see how increasing the 12V rail will help with the CPU or memory because they're fed from voltage regulators built into the mobo.

See below for cutoffs...

But yes, it doesnt help at all... BobN may correct me on this but I would imagine that it will stress out the mobo power bits that feed the CPU/parts.

Pretty average 12V rail OVP shutdown point in a PSU is 13.4v.
Good protections ICs cut it off at more like 13.
It's still within ATX spec at 12.6000v, 12.6001v is over spec.

 

[Bobnova]

This is one of the better ICs:


This is a midrange Weltrend unit:


PS223, a bit looser:


Some units like the NEX1500 use a microcontroller and custom code to do OVP/UVP, its OVP on 12v hit at 12.83v, excellent.