OB d-i-y: ATX PSUs

Sitting on the "bench" across the room is a new upper-midrange motherboard (Gigabyte GA-8IK1100), just filled with a 2.8G P4, and an immodest 2GB of ECC PC2700 RAM. (It's going to replace the Athlon mobo processing these very keystrokes, which is slightly prone to misbehaviour under both Windows and *BSD. And I run enough long compiles and similar that I won't have non-ECC RAM, which restricts my choice of mobos quite heavily, sigh.)

It's running its first powerup right now - Memtest86, of course. The weirdness, tho', was connecting a fancy "silent" PSU - no rubbish here, a Tagan TG480-U01. Did the usual minimum install - memory, CPU, heatsink+fan, vid card, and spare KVM-switch inputs for mouse and kbd. Plugged in the PSU, incl the 12V aux supply. Turn on PSU switch, momentarily short the "power switch" mobo pins. "Almost" nothing - a little LED lights on the mobo by the DIMMs, to indicate standby power is being supplied (so meaning DON'T fiddle with the DIMMs or the PCI cards!), but no full-on PSU turnon (and I did have an old HD and a CDROM to provide a 12V load too). Shorting the "power" pin for 4s turns the little LED off again, a momentary touch lights it, but again, as we say in the Valleys, dim bootio, dim PSU-fullonio.

Disconnect PSU, short its ATX-pwr pin 14 (green wire) to adjacent black (ground) pin. Nothing. Dim sausigio. I was sure 'normal' ATX PSUs were supposed to switch-on while pin 14 was grounded? And indeed, nipping down the road to PCWorld (I know, I know...) and picking up a less fancy-pants ATX PSU (still 450W-nominal, yada yada) yields one which not only comes on full with pin 14 grounded, but powers up the mobo perfectly well.

I call that more or less conclusive evidence of misbehaviour by the Tagan PSU, but just before I make an arse of myself sending it back to the fine people at overclockers.co.uk - does anyone with relevant PC-assembly experience want to clue me in on any simple oopsie?

Oh, and in the realms of "obvious when you think about it": the processor fan, running under the mobo's own speed control, is doing nice slow revs, and the hs heatpipes are barely warm to the touch. But as there's no case fan moving any air, the *memory* modules are letting off 'noticeable' amounts of heat! (Kinda reasonable the moment you think about it, since Memtest86 is full-on accessing all of RAM just as fast as the memory controller will let it!). An infra-red thermometer claims to see temps of 55C or so on the DIMMs (and on a few spots of the similarly unfanned vidcard). Still, if they survive a 'burnin' like this, they should be happier in the case with reasonable airflow!

Stefek

Reply to
Stefek Zaba
Loading thread data ...

Yup, indeedy

Me thinks you have your answer....

Lets assume that your turned on the nice rocker switch on the back of the PSU for a moment... ;-)

And it has not got an input voltage selector slider thingy that is slid (or not slid) in the wrong (or right) place, nor somewhere betwixt the two of them, then, your PSU is in need of buttering and eating...

(i.e. is toast)

55 hardly counts as warm in reality... I would rather have em a bit cooler, but they will survive that indefinitely (well until your shiny new processor seems hopelessly inadequate for running solitaire, and hence you upgrade it to something that won't fit in the motherboard and wants DDR7 RAM or some such). Oops my MB expansion potential cynicism is showing ;-)
Reply to
John Rumm

HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.