Interesting Post Mortem findings after a PS dies
Well I had an Antec Trupower II 550w SLI power supply die on me last week when something blew in side of it and replaced with an Ultra 600w SLI PS. About a year ago, I started having a problem where none of my BIOS settings were saving if the PS was disconnected or there was a black out. I replaced the battery and it didn't help. Thought it would be something I had to live with. Well put in the new PS, and 2 days latter we lost power for an hour or so. When I turned on the PC, all of the settings were saved and I didn't have to reset everything in the BIOS. Two problems solved with one fix....Just curious, any thoughts what might have been happening?
That's a good question. I don't have the answer, but I have an idea. :idea:
The ATX form factor specifies that the ATX PSU must always supply power to the motherboard, even when the "power" is off. This is important for stuff like Wake-On-LAN (WoL), Suspend-To-RAM (STR - S3 power state), etc. The motherboard probably uses this life-line to maintain the BIOS settings as well.
It's possible that the motherboard thought it could use the life-line power while the PC was off (instead of the battery) to keep the BIOS settings in-memory, but for some reason it wasn't enough juice (or stable enough) to actually hold the BIOS settings in-memory.
Just a theory. On that note, with Flash memory being as cheap as it is, I'm surprised we still need batteries at all... :!:
(edit: curiosity killed the cat)
Well now that its blown, is it possible to see whats gone pop? You could end up with a spare PSU o sell or use. save a bit of cash.
If you can see inside it, you might be able to see debris or the offending device without having to open it and put yourself at risk from shock.
Maybe it can be repaired? if its out of warranty.
5/10 you might find its a blown internal fuse. High Ampage overload fuses usually go with quite a pop, just watch for glass, as some of them are glass cased.
Just don't go too far and start bodging things, I don't want you to end up getting shocked or start an electrical fire.
consult a competent, qualified engineer if you have doubts. Don't go fiddling, unless of course you are an engineer.
I wouldn't recommend opening a PSU. The capacitors in PSU's still retain electricity, enough to seriously hurt you. What probably happened is one of the capacitors leaked or fried, that supply small amounts of power to the motherboard when it's turned off, which gives power to the BIOS chip so it doesn't lose settings.
Quote:I thought the battery is for the CMOS chip?
I wonder if maybe, during POST, the PSU caused the motherboard to think something wasn't right, and resulted in the motherboard BIOS rolling back the settings to the default settings? Just another theory.
CMOS = Complementary metal–oxide semiconductor
BIOS = Basic Input/Output System
CMOS is a type of chip (as opposed to the TTL). It describes how the transistors are wired together in the IC to form logic, if I remember correctly.
The BIOS is stored in a CMOS chip. The terms are generally used interchangeably when referring to the BIOS.
You're right about not opening up the PSU. It's generally a bad idea, unless you know what you're doing. Even if there was a user-servicable cap that's obviously bad, chances are that you've strained the other components so much that your PSU still won't be 'repaired' by replacing the part (Usually the regulators will pick up the slack if a cap goes bad. These tend to get really hot, which we all know makes chips go south fast).
It would make sense that the mobo would use the life-line power as opposed to the battery, if it's available. I don't know enough about it to say for sure, though.
i hear that, hence why its best to look through any grills in the PSU.
Got zapped at work a few days ago repairing a pilot Arc board for a welder. Provides a high voltage spark by switching current on and off through a coil/transformer, bit like a cars ignition system.
600V DC charged caps, ow, hurt like a bee hatch. 8O
still can't hold my hand steady.
Quote:The safe way to handle a PSU is to disconnect it and let it drain out of the case for 1 day. That is for a good PSU, not sure about a faulty one.
Caps can hold charges for a lot longer than a day; especially big ones.
Best to stay out of it completely.
A day? That's really not feasible nor is it effective. Capacitors can hold a charge for much longer than a day. Switch the PSU off and then hit the power button on the computer. Fans spin up for like 500ms, caps get drained.
As for the question in the OP: I have no clue why a PSU change would affect CMOS clear. Are you sure you didn't change the CMOS clear jumper setting to the correct setting when you changed out the PSU? I've seen motherboards that will not clear the CMOS unless there is no power to the motherboard and motherboards that will let you turn the system on while the CMOS jumper is set to clear. If you had both characteristics in the same board and had it set to clear that would explain it... that's my only idea though.
I've actually found more bad inductors in switching PSUs than caps. Especially if they get hot, the big uncovered inductors tend to snap off from the heat constantly expanding and contracting the wire.
It's a bad idea to try and repair a swithing PSU. If you touch the wrong pins, you could discharge a cap into another component and reverse-volt fry something.
Unless you know what you're doing, really, just RMA it or ebay it to someone who does. It's unlikely you'll have the parts required to repair the PSU, anyway. With PSU's, when one component goes, there's a good chance it takes 3 or 4 more with it, especially the regulators.
yea I would not even consider it after the "big bang" - but if you catch it at the right time maybe, have seen over two dozen postmortem smartpowers and only two that were salvageable (were intermittently crashing, no smoke yet) so recapped them first chance I got and they are still running today.
Mobos are much more pain, you need the temp controlled soldering iron. I recapped one about a year ago, had to get one of those and was a major pain - all for a lowly compaq athlon xp1600 board. Makes you wonder if it's worth the hassle, really just wanted to see if it can be done easily and took a lot longer than I expected. Well compaq wanted over $150 for a new board and I spend about $15 on new caps not to mention the blood and sweat :wink: doubt I will do that again anytime soon since I sold my solder iron
Yeah definately. I hate desoldering. I could never justify the $300 price tag for a temp-controlled soldering iron when I could run to a radio-shack and grab a cheap one for $15. I'd just go through a LOT of flux. I had a nice iron in the lab at work, though.
I've never tried repairing a computer PSU; mostly low-voltage switching power supplies for electronic cash registers. I had the schematics, so it was pretty easy to look at the bad part on the schematic and see the 2 or 3 components it likely affected when it went.
Manufacturing was so inexpensive, that most of the time it was cheaper to just replace the power board than to try and repair it (I think they were like US$15 or US$17). When a cap went, it was usually done and done. If an inductor or regulator went, they were usually pretty easy to identify and replace.
I never really liked analog electronics, though. Too much math and not enough payoff I'll take a 40MHz 8051 over a switching PSU any day of the week
I'm not sure that computer PSU's are SMPSU's (switch mode).
SMPSU's make quite alot of noise, both electronic and audiable, I always assumed they were linear?
Anyway, thats is the best way to discharge a caps safely I've found, I use a 11W R47 resistor, quick discharge and energy disipation compared to using a 1/4W equivilent.
I find that both caps and inductors are both suspect, but more inductors for the same reason of snapping legs due to thermal changes.
But caps can go leaky from time to time, I find that tantalin caps are a sod for doing that. I've tested PCB's with tantalin caps used as decoupling caps, and 9 times out of 10 the power rails are short.
But the standard electrolitic ones are fairly robust and hardly fail, unless they see AC, in which case its a mess!
I think i might try and see if i can take your idea and buy a busted PSU of ebay and see if i can bring it back to life. I would post what I've found/done on here if thats ok?
if your old psu was faulty and initially on boot doesnt supply enough steady power to start the system, boards from asus and gigabyte etc detect the bad start and reset the cmos settings to have a fault free start (also works when you do a bad overclock - the bios detects it and resets your settings) this may be the reason, and if you ask me, the setting sucks when you set all your settings and it "automatically" looses your settings etc.