need bigger psu?

[gazix]

I've got a duron 1300 w retail fan (no overclocking)
abit kt7a raid mobo
dvd-rom, cdrw and dvd-rw drive
7200 rpm hard drive with drive bay cooler (2 mini fans)
5400rpm hard drive off the raid controller (need to use a power
splitter)

running a generic 350w power supply now, no plans on upgrading or adding more devices to it (maybe a dvd-rom/cdrw combo drive when my dvd-rom dies) or a case fan or two

[markjs]

As long as you are booting ok and the system is stable then I see no need for a new PSU.

[grover]

If a cheap PSU is overloaded, it may work for a time, but it's a ticking time bomb and they often cause damage when they go. If it's a slow failure, you might have a few days warning to start load-shedding and getting a new one on order, but they just as often die in a cascade failure, with working rails spiking as other rails fail, and sometimes take down your mobo or components with it. If you can afford it, it's worth peace of mind to upgrade, but even an upgrade isn't a sure thing, as even the best PSUs can die randomly. But they generally do last a lot longer!

In the end, it's basically a numbers/probability game... With a generic 350W PSU on your rig, you should most likely be OK. But it never hurts to upgrade!

[SexyMF]

If a cheap PSU is overloaded, it may work for a time, but it's a ticking time bomb and they often cause damage when they go. If it's a slow failure, you might have a few days warning to start load-shedding and getting a new one on order, but they just as often die in a cascade failure, with working rails spiking as other rails fail, and sometimes take down your mobo or components with it. If you can afford it, it's worth peace of mind to upgrade, but even an upgrade isn't a sure thing, as even the best PSUs can die randomly. But they generally do last a lot longer!

Not meaning to poke but I don't see how you could get spikes from a Switched mode power supply. There is a fixed switching frequency which chops up rectified mains. The outputs are achieved via transformer action, the input being the chopped up recitifed mains. Each winding is seperate with feedback winding keeping things in check. Only certain failure modes like primary winding shorts could cause higher secondary voltages.

People here are forever talking abou slow faults and immenant PSU failure but it is rarely backed up with electrical/electronic proof - or even a picture.

Generic PSUs do not deserve such a negative image.


I will add if the switching FET fails as a collector-emmiter short you may see a primary voltage spike, this will be seen in proportion on the seconday outputs (your voltage rails) - that may be very bad.

[grover]

Who needs a diagram? Not like the generic transformers come with a troubleshooting diagram with which to prove their shortcomings... In fact, I'd be very surprised if all PSUs were switched; you can accomplish the same thing with a single 5-terminal unrectified transformer, someone will be selling a budget PSU with a single 5-terminal unrectified transformer. At any rate, a transformer with a heavy load that suddenly drops becomes a heavily charged inductor- which would almost instantly discharge upon any availible circuit as a voltage spike; that spike going back into the switching MOSFETs can overload and transfer a spike just as well as a wire would have.

I'd expect high quality PSUs to have filtering and protections to prevent this from happening, but not all PSUs are created equal! There is a big difference between a $15 PSU and a $70 PSU and it's not just the quality of the fan.

What I *do* know is that regardless of the electronics of the failure mode, PSUs that die run a disturbingly large chance of taking part of the computer with it when they go. This is reported on this forum time and time again. Whether that's from a voltage spike (seems the most likely cause) or from undervoltage problems or something else entirely is almost moot; cheap PSUs die early deaths and an early PSU death is a serious risk to the rest of the PC. Unless it's that failing components kill cheap PSUs, that could be it, too, and we've got it backwards, heh