welrod

Distinguished
Jul 16, 2011
21
0
18,510
I'm installing a new PSU, replacing an old 580w w a Silverstone 1000w st-g. The pc i bought was prebuilt and has some case fans and lights to worry about. In addition to a few confusing things (like a light and battery pack looking thing at the bottom of the case that have never seemed to work) I'm a bit confused by the way the fans were set up.

My motherboard is an MSI p45 Platinum and it has a 4-pin out called JPWR which I believe can provide power to video cards, according to the manual PDF for the mobo. But when they built my PC, it looks like they used this to start a daisy chain of fan power. It splits to a Y, one to the back exhaust fan and one to the female end of the side window case fan, which then males out to a 4-pin connector I can't see behind the drive shield wall, and then looks like it comes out of the other side of the drive shield wall and connects to a molex for the front chassis fan and possibly what looks to be another one for the front chassis light.

So I have a couple questions.. #1, to use JPWR to power all of the fans, as it looks like was the original setup (none of the SYSFAN ports on the motherboard were being used and still aren't) ... how is that different from the more normal either fan->psu direct, or fan->sysfan setups? should I try to emulate this with the new PSU setup? And, I know it's random as topics go, but is it safe for me to take off the OTHER side of the case (non-window side) without doing anything untoward to the mobo, which is of course still in place in the pc? That way I could see the wiring jumble that's behind the shield wall, BUT I don't want to take the chance of damaging anything as this is my first time doing this sort of thing and I've already had to bang my knuckles a bit into the edge of my video card, remove and replace a ram chip just to unhook the old psu.. any answers or ideas? thanks a lot btw people like me would have no hope without places like this ;)
 
Solution
JPWR1 12V power connector is suppose to be connected to a 4-pin Molex power connector from the power supply unit. Its purpose is to supplement the power traces to the PCI-E x16 slot so that the graphics card is able to draw more power from the PCI-E x16 slot than the motherboard power traces as designed allowed at that time. This is actually a kludge to handle the revision to the PCI-E power spec revision that happened around the time that motherboard was released so that they didn't have to redesign a new motherboard.

The daisy chaining of cooling fans plugged into JPWR1 was done by someone who didn't have a clue what they were doing.

For those power supplies that don't have a power switch and when the power supply is still plugged...

welrod

Distinguished
Jul 16, 2011
21
0
18,510
and another thing.. if i wire somethjing like a front chassis light, or a front chassis fan, directly to my PSU, and my PSU itself does not have an on/off switch (silverstone st-g) , how then will any signal ever be sent to the fan or chassis light to turn off, when my PC is turned off by way of the OFF button???? Is the PSU aware that when the 24/20 pin MOBO wire stops asking for power, it's time to turn off? Or do they stay on all the time this way?

Could this be why the original supply of power to these was through JPWR, maybe to have MOBO control over these devices' on/off status, maybe to work around not being able to reach all devices to SYSFAN ports on the MOBO?
 

welrod

Distinguished
Jul 16, 2011
21
0
18,510
I am deeply in need of help... JPWR looks like it is meant to RECEIVE power, where this whole time I thought it was being used to PROVIDE power to my fans, the way it was originally set up.. could it be that the old PSU had a daisy chain providing power to the front fan and chassis fan, the side window fan, the ccl inverter, and *THEN* JPWR (for rudimentary GFX power?) and THEN lastly the back exhaust fan? THey say you can't hook 4-pin peripheral cords (molex) up wrong, so I'm pretty sure I'm right that JPWR receives power.. right?

And if JPWR provides power to the GFX, but my GFX card has a requirement for two 6-pin PCIe cords right from the PSU, can I plug them into the PSU as seems to be normal, and then still send power to JPWR?
 
JPWR1 12V power connector is suppose to be connected to a 4-pin Molex power connector from the power supply unit. Its purpose is to supplement the power traces to the PCI-E x16 slot so that the graphics card is able to draw more power from the PCI-E x16 slot than the motherboard power traces as designed allowed at that time. This is actually a kludge to handle the revision to the PCI-E power spec revision that happened around the time that motherboard was released so that they didn't have to redesign a new motherboard.

The daisy chaining of cooling fans plugged into JPWR1 was done by someone who didn't have a clue what they were doing.

For those power supplies that don't have a power switch and when the power supply is still plugged into an AC outlet and the system is turned off the power supply unit is not completely off. The +5VSB rail of the power supply unit must always remain on or the motherboard and other rails of the power supply will never turn on when you press the chassis power button.

 
Solution

welrod

Distinguished
Jul 16, 2011
21
0
18,510
So, okay! Well, does this mean that it's best to leave JPWR *not* powered, as my video card is receiving feed from two 8-pin->6-pin PCIe connectors right from the PSU, now? Or should it still receive molex feed, and should that be direct from PSU?

What I've done is this: I wired up the fans to nodes on a 6-pin->molex cord from the PSU directly. Tried to give everything its own node, but I was one node short so the front fan got to use the other end of the Y from the front chassis light wire. The last molex node on this chain (direct from PSU) was given to the side window fan, which then fed its other end of the Y out into the back fan, which sent *its* other end of the Y to JPWR, where the chain terminates.

This fired up for 2 seconds then on and off, and I tried unplugging the molex for the ccfl inverter, and it ran fine.

Well, not *fine*: it's very hot now, idling at 70c on the cpu. This is hotter than it was before with the 580w heat-machine PSU... so I'm still a bit disturbed and troubleshooting that.

I've been told I should take off the (stock) HSF, clean, re-paste, and re-seat, and that'll fix it.. I've never done anything that hardcore before.. wondering if my wiring might be creating any extra heat through its complicated structure that I should rewire and try first before going at the CPU (as a relative noob) ?

Also I haven't managed the new fat PSU cables so I could be creating the heat problem that way, will try tying them to chassis walls and such first, to see if that helps. But 70c seems like a lot, I'm guessing everything needs a little doing.
 
You should still have a 4-pin Molex connector from the power supply connected to the 4-pin Molex connector next to the PCI-E slot on the motherboard.

Cooling fans are low power draw devices so daisy-chaining a few of them together is safe.

If you leave the side panel off of your chassis does the temperature drop? If the temperature does drop with the side panel off then your chassis has insufficient airflow and that will need to be corrected.

Removing the CPU HSF and cleaning and reapplying the TIM (Thermal Interface Material) would also be recommended. Also thoroughly clean any dust out of the HSF assembly.