Linux - HardwareThis forum is for Hardware issues.
Having trouble installing a piece of hardware? Want to know if that peripheral is compatible with Linux?
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
Do you agree with me that the computer's power supply is the component for which the prospect of an unexpected breakdown should worry me the most? As far as I know, it's the component whose failure can be the most destructive: some years ago my dad's computer suddenly died, and when he took it for repair, he was told the PSU had blown and taken the motherboard with it. But besides this, it doesn't seem as easy to monitor the condition of the PSU. There are these "sensors" programs that monitor and report the condition of the most important electronic components, but the only way I know to examine the condition of the PSU is to plug a tester into its cables. I bought and used one, but it occurred to me that the primary usefulness of these CPU testers (possibly even the only serious purpose) is to figure out after the computer breaks down whether the power supply is the problem. They indicate only whether the power supply is delivering power, not whether the power supply is close to failure.
The only failed component I have ever had, that took out some other component, was a power supply. And it took out the motherboard. Probably the CPU too, but when I scrapped the dead system, I only retained the hard drives, DVD burner, and a few misc screws and fans. So I don't really know if the CPU and memory had been destroyed or not. Sockets on mobos change so fast that there's not a lot of call for trying to salvage an older CPU.
Ever since, other than installing good quality power supplies rather than the bargain basement stuff, I don't do any special worrying about the PSU. If it goes, it goes I guess. Just buy a good one and hope it doesn't. Same thing for hard drives. And of course, backup your system well so when something DOES go, and it will, you're not in too bad of a situation.
In theory you could preventatively replace the PSU when you notice the voltages become unstable (vary quickly with time), or drop to near critical value. However, this won't prevent damage from sudden failure.
I always invest more extra bucks into PSU & mainboard before thinking about any other components. Bad PSU's don't protect well against lightning strikes, over/under voltages, electricity burts or things like that. Nothing more foolish than to buy costly hardware and then go for a cheap PSU. Most important component. Good manufacturers build very capable PSUs that can handle any electric problem very well, also offer 120mm coolers (silent) and good power throughput.
When that is done, i usually also go for a slightly better mainboard at least, saving me later fiddling if the board somehow won't satisfy me. When these two components are fine, i can add whatever i desire.
Sorry but quality/good PSU usually do not fair well with direct lightening. Sure better filtration for normal interference (EMF/RFI) but the quality you are paying for is tolerance and stability. UPS does provide protection for dips/brownout/power outage cycles and does limit EMF & RFI.
If you value the equipment then be sure to use power protection via UPS or even Surge suppression. I prefer UPS for my systems throughout my lab or personal equipment(s). Simple surge protection is generally a one time event, MOVs' do not heal but fail.
Distribution: Debian Wheezy/Jessie/Sid, Linux Mint DE
Power supply failure which damages the other components is extremely rare. Not impossible, but rare. In the past 20 years that I do computer maintenance, I might have seen 10-20 power supply failures. But all of them failed with absence of voltage.
I did not maintain thousands of computers, but the odds already indicate it seldomly happens. Before that period I think I have seen it happening once, but that was < 1994.
It makes sense that a power supply fails by stop working, not by producing too high a voltage. Power supplies are exclusively the switching type. High stressed components are likely to fail. Highly stresses components are the components involved in switching: capacitors, FETs and diodes. If those fail, the switching stops. No output anymore. There are a number of very lightly stressed components, to measure and control the output voltage. If those fail an overvoltage might result.
This is different from linear power supplies where the regulator mostly is a series regulator. If that one fails, you'd have an overvoltage in the majority of cases. So the assumption of power supplies damaging other circuits might be an urban legend from the past. As a matter of fact it happened to me once that a friend of mine created a short between +12 and +5 of his linear computer power supply. While his 5V crowbar was not in place yet. Thyristors were expensive. That blew off most of the IC caps of his Apple II clone.
So if the power supply is of reasonable quality there is not much to worry about. Hardware is so cheap that the one time time would happen is an acceptable risk. The risk doesn't deserve the attention. Backup are infinitely more important.
And don't forget that on the mainboard itself a number of voltage regulators are built as well. They reduce 3.3V to an accurate 1.342 V (??) for the processor core. Any overvoltage would damage the processor in microseconds.
A close lightning hit causes serious overvoltage in power supplies leading to immediate failure of the power supply and attached components. Or soft failures. Or failure a few months thereafter.
Okay, "lightning strike" might have been an exaggeration on my part, but in my case i had an experience with a lighning strike in my house that destroyed my phone/DSL hardware, my old PC (power supply + mainboard gone byebye) and made the lights in my room flicker big time while at the same time my workstation with an excellent PSU didn't even twitch, it just kept grinning while compiling a piece of software. Saved me some bucks because some other hardware in that box wasn't cheap.
Edit: and yes, that brave PSU that saved me so well from further damage gave up about 6 months later, i.e., it seems to have been affected at least somehow. Nevertheless, that was the only thing i eventually had to replace.
Power supply failure which damages the other components is extremely rare. Not impossible, but rare. In the past 20 years that I do computer maintenance, I might have seen 10-20 power supply failures. But all of them failed with absence of voltage.[...]jlinkels
I agree with jlinkels, I think that most PSU failures are component failures where the voltage decreases. I would say only a power surge could realistically produce an increase in voltage (possibly due to lightning).
Power supply failure which damages the other components is extremely rare.
I am not as long in the hardware business as you, but this is exactly my observation. While in the consumer space I found cheap PSUs to be the most common case why a system doesn't work correctly (or not at all) I have only seen one or two cases where a PSU damaged the rest of the system in several hundred repaired PCs. While I recommend to anyone not to save with the PSU and get a high quality one a failing PSU is for me not scariest failure that one of my machines can have, since they still are relatively cheap in comparison to the other parts, like CPU, mainboard or a gaming videocard. This is the reason why I always spend a few Euros more to get parts with decent quality, this saves money in the long run.