3.3v PSU Issue.. what now?

Fizzer

Fizzer

Fizzer

Associate
Joined
23 May 2006
Posts
28
Hello

I have recently purchased a Gaiinwood Bliss 7800GS+ 512Mb AGP.

Initially I thought the card was faulty but later found that the 3.3v rail on my motherboard had burn't out.

Burnt out 3.3v Motherboard

Now the supply I am using is an Akasa PaxPower 400w and specs are:

+3.3V - 30.0A
+5V - 28.0A
+12V1 - 14.0A
+12V2 - 15.0A (dual 12v rails)
-12V - 0.5A
+5Vsb - 2.0A
3.3V and 5V combined load - 130W

In the system is two sticks of Geil PC4000 Ultra memory can't find any power specs for these.
The memory has always had to be run @ 2.8v for stability (this is common with Geil).

Now suspecting my motherboard I replaced it with a Gigabyte GA8661 (SiS661 chipset) from another system.

I ran Prime and SiSoft burn-in on the old graphics card (6600GT) for a couple of hours to make sure of system stability.

Then I put the new 7800gt back-in and Benchmarked a bit and played games.
Now after running a couple of bench more bench marks this happened.

Burnt out 3.3v on ATX connector

Now the only common factor is the power supply and the NEW graphics card.

Is the 7800gt pulling too much current off the 3.3v rail. I know the specs are tight for AGP.

Or was the supply faulty, I have noticed that the 3.3v rail in my old ABIT motherboard monitor was a bit low. I can't monitor the Gigabyte one as thier isn't a monitor for it.

I am buying a TAGAN 580w PSU BUT the specs are:

+3.3V - 26A
+5V - 30A
+12V1 - 20A
+12V2 - 20A
-5V - 0.8A
-12v - 1.0A
-5VSB - 2.5A
Max. combined 3.3v + 5v - 180W

Yes I realise that the Tagans combined 3.3v, 5v rail is 50w more. BUT is the 2 sticks of memory and the 7800GT pulling that much current. :confused:
ALSO the Akasa supply has a higher current rating for the 3.3v rail :confused: :confused:
I don't want to buy a new supply and end up blowing that due to faulty Graphics card.

Any advice would be really appreciated.

Thanks
 
I think this problem is partly due to the connector itself being unable to cope with the current draw, but i have seen this on another motherboard and it was caused by the design of the motherboard using the 3.3v rail for the cpu power rather than the 12v rail.

Theres lots of info and a solution to this here. Looks like the same problem to me.

I take no responsibility for any damage cause through doing that. ;)
Contact the motherboard manufacturer and ask if this board or early revisions of this board used the 3.3v rail for cpu power.
 
Thanks for the links. I will look into this.

Problem is this has happened on two different Motherboards now the IS7-E and the Gigabyte 8S661.
Both have run solid and stable for somtime now with no issues until I put the 7800GS+ card in. The Gigabyte in a SFF case with a limited 300W supply.

The only thing common was me moving Graphics cards to the 7800GS and staying with the supply.

Can some one confirm if this card will put more draw on the 3.3v rail???

If not then why the heck has this suddenly happened to two diffrent boards. I have to get another supply and looking at getting an Abit AV8 Motherboard and Tagan 580w supply.

I don't want them to go the same way because of a card issue.

Disclaimer: I am old enough and stupid enough to know that all actions decisions and ultimate f... up's are all my own and I will not hold anyone but my birth parents responsible for my inadequicies :D
 
Last edited:
Hello

Just realised the links to that issue don't apply. Firstly it's an old issue with the older motherboards and... More inportantly mine is the 3.3v rail NOT the 5.5v

atx_burn_dia.jpg


Diffrent pins you see mine are the 1-2 and 11 that have blown, 5.5v issue on CPU effects the 20-21 pins.

Just thought I had better confirm that.

So anyone have any ideas??

I've been searching and I am now stumped. Any thoughts on this 3.3v burn out please ?
 
Last edited:
Have you contacted abit? (looks like an abit).
Try asking them first.
Also try reading up on agp and see if it uses the 3.3v rail. Try wikipedia.
 
The Geforce should be pulling most/all its power off the molex's, not the AGP slot. Are you sure the damage to the ATX connector wasnt done when the IC7's power connection was damaged?

Or perhaps when the connector was overheated the first time around, it oxidised the contacts, causing bad contact with the new motherboard, which would likely cause additional heat generation.

Apparently the AGP slot is only rated for 25W load, hence the requirement for external (molex) power source for the card.

Im still wondering, if the Geforce is pushing the 12V rail too hard, and the PSU's voltage regulators are tossing everything it has at 12V and letting 3.3V drop. Not too sure here, but the motherboard's voltage regulators (the ones giving 2.8V to your ram), might be increasing amps to compensate for the voltage loss, and overloading the connection.

Thats the best guess I can think of, anyone got a better one?
 
Last edited:
I dont think any P4 motherboards ever use the 3.3V, or 5V rails for the CPU.

P4's run off the 12V rail, hence why the ATX+12 was often called P4 compatible. It was AMD XP's that ran off the 5V rail.
 
Thanks for the replies people.

All I read pointed that the P4 didn't use much in the way of 3.3 + 5 combined But the AMD's did.

Probably a combination of a dodgy PSU and new card.

Well I am sourcing an AMD board to go with the Athlon64 I got and when the Tagan supply arrives that will be enough. :)

Thanks for Help
 
alternativley you could getaway using an oldskool psu like mine. my psu has something like 46a on the 5v rail and thicker than average wires in the atx connector block. question is, can you still get a psu like mine.

closest i have seen is the qtec 650w. yes i know its qtec but it was manufactured for old skool setups with lots of current on 5v and 3.3v rails.
 
Qtec are measured for Peak load though, not continous, or at least they used to be. Wouldnt touch a qtec if you paid me ;)
 
You must log in or register to reply here.
Back
Top Bottom