Did low voltage cause the pc to fail?

We would need to know the nature of the PC failure. The only component of the PC that could possibly fail due to a mains power problem is the power supply, because its output is regulated -- it either outputs the right voltages at full current, or shuts itself off completely.

It might over-exert itself stepping up a lower-than-normal input voltage...

Also, power cutting on and off repeatedly (a lot of "flickering") might damage the power supply or even, I suppose, a disk drive or something. Did that happen?

Either undervoltage or arcing can damage equipment.

Unervoltage - also known as a brownout - does not damage electronics. Unfortunately too many take what they learned about motors and expolate that to knowledge about electronics. One must first learn basic concept before making assumptions such as low voltage will damage a computer.

The Intel specs are quite bluntly clear about this. AC mains voltage can drop so low that incandescent bulbs are at less than 40% intensity. Even at that low voltage, a computer with a full load of peripherals must power up and run normally. Where is the damage? It only exists in a world where people don't learn simple technologies nor read manufacturer numerical specs.

What is a shutdown? Power is removed. Electrolytic capacitors slowly discharge. The computer suffers a brownout - diminished voltage. So the computer is damaged? Of course not. Computer must work just fine until voltage becomes too low - and then it must switch to a no more output mode. IOW a shutdown is a brownout that eventually becomes a blackout. But if brownouts - low voltage - cause hardware damage, then a shutdown will also damage hardware.

Can computers be damaged by a brownout? Well if it is, then the computer fails to mean numerous industry standards and defacto standards that have exists longer than most every reader here. The answer to the OPs question is found in those specs. If undervoltage damages the electronics, then the human knows exactly what the reason for failure was - the human.

there

damage

arcing

Intel doesn't make PC power supplies. Sometimes we don't know *who* makes them. I wouldn't put it past them to have something that would fail (or at least blow a fuse) when subjected to undervoltage.

After all, undervoltage requires the switching power supply to draw *more* current (as it gets less voltage).

In an ideal world this would be so.

In the real World, it's quite possible that brownouts can cause switchmode power supplies to fail.

--- sam | Sci.Electronics.Repair FAQ Mirror:

Repair | Main Table of Contents:

+Lasers | Sam's Laser FAQ: | Mirror Sites:

Note: These links are hopefully temporary until we can sort out the excessive traffic on Repairfaq.org.

Important: Anything sent to the email address in the message header is ignored. To contact me, please use the feedback form on the S.E.R FAQ Web sites.

In the real world, only those with a bean counter mentality have power supplies damaged by brownouts. When a power supply is damaged by a brownout, then the reason for failure is directly traceable to a human who failed to learn basic technology.

Again - what is a power down or shutdown? A brownout that eventually becomes a blackout. If brownouts damage electronics, then so does turning the power off. This alone is a damning fact. Brownouts do not damage properly constructed electronics. Shutdown also does not damage those electronics.

Even in those cited links, I find no reference to brownouts causing electronics damage. The CBEMA specifically states that all equipment must not be damaged by low volts. They and other industry standards - including the computer industry standard originally created by Intel - are quite blunt about this. Brownouts don't cause electronics damage. Stated bluntly even in specs.

Intel defines how computer power supplies must operate. IBM, Dell, AMD, and a long list of other responsible manufacturers also demand same standards be met. But this is old technology - older than most lurkers have even existed. However many bean counters buy 'dumped into North America' power supplies to cut costs. Power supplies that violate basic Intel requirements. Then those silly bean counter mentalities try to blame others; then claim brownouts damage electronics. Where is the logic in that? Its called corruption.

Numerous industry standards for ATX power supplies were the same defacto standards of 30+ years ago. Such defective supplies are not found in brand name computers. Dumping only works at great profit to the Asian manufacturer when a computer assembler does not even have basic electrical knowledge - buys power supplies on price rather than upon technical specifications.

If anyone says that a brownout can damage properly designed electronics, then we have a benchmark for a bean counter mentality masking as technically knowledgeable.

Those who think otherwise are then invited to explain how the individual components inside the power supply fail due to brownout. That's right. I design at the component level - not just rack and stack black boxes like a computer assembler who need not have any electrical knowledge. Please feel free to describe how that electrolytic capacitor or power transistor is damaged by a brownout. One must be that knowledgeable to foolishly claim brownouts damage electronics.

Sam Goldwasser wrote:

That's all good and nice reasoning which the designer already considered when he designed a power supply that is not damaged by brownouts. There is even a circuit inside the supply that cuts off power IF the supply cannot maintain required output voltages. Same circuit also sends a signal to motherboard. Again, no damage to hardware.

Intel does not make power supplies? Underwriters Laboratory

- UL - does not make anything. Therefore UL standards don't exist either?

If a computer power supply is damaged by the brownout, then the brownout is not a reason for failure. That failure is directly traceable to the human who typically buys on price rather than first learn basic electricity concepts.

If the power supply does not come with written specs - things they actually claim to do - then one should assume the worst. These same 'discount' power supplies are sold to computer assemblers who would blame the brownout rather than blame themselves. Brownouts do not damage properly constructed power supplies. Unfortunately those supplies cost more money. Bean counter mentalities fear spending money.

I bought my power supply from some guy wear> Intel doesn't make PC power supplies. Sometimes we don't know

Here in the US we have 240/120 supplied to homes. If the grounded conductor fails, the voltage can divide up as 100 and 140 or as 80 and 160 and so on depending on the loads at the time. Maybe that's what happened and it was overvoltage that fried the computers.

And, of course, every PC ever produced anywhere in the world IS by definition completely, utterly, and totally compliant with Intel's specs, right? :-) :-) :-)

Ah, to live in such a world....

Bob M.

Or so you hope.

One of the things I used to do (in a past life, but for my current employer) was to supervise the environmental, etc. qualification testing of new products. Do you think we never EVER found problems with power supply design that showed up in the AC line compatibility testing?

It would be a truly wonderful world in which everything was designed as it should be, in which everything was actually BUILT per those designs, and in which the components that went into that construction never ever had any sort of problem of their own. Just as soon as you find that world, you let us know.

Both Intel-generated specifications exist, and UL standards exist. And so, according to you, everything ever designed and built automatically complies with those standards? Hallelujah! I can call down to the test lab right now, and tell those guys to take it easy! There's nothing more for them to do!!!!

First, it's hardly reasonable to expect anyone who buys a power supply to be sufficiently educated in power supply design so as to recognize a good design or a bad one at first glance. (Hey, if they're THAT good, they should be designing and building their OWN, and then I know several places that will likely want to hire them!) Second - you again show an awful lot of faith in the mere presence of "written specs." Many aren't worth the paper they're printed on.

....as long as they were properly designed AND constructed, built from perfect components, and have suffered absolutely no ill effects due to age, electrical or mechanical stress, etc., since being built. By this same sort of reasoning, I should be able to claim that NOTHING which is "properly designed and built" should ever fail. Once again, the reliability guys will be SO glad to hear that...

Bob M.

May I ask just how many years experience YOU have in power supply, design, and construction in the commercial world (i.e., actually shipping this sort of product in volume)?

Bob M.

Your ignorance is really showing up. If the specs are not met, that supply goes back on guarantee and it eventually will drive this supplyer out of business. In Europe we have now a 2years guarantee by law and believe it or not, when a part fails because of undervoltage the seller has to take it back and repair/replace it.

It is a criminal offense if UL-specs are not met, and if any personal injury happens you will pay big bucks. There is a responsability involved and if you fake the CE or UL or whatever rules apply in your country, you will be prosecuted.

Mr Meyer, you must be one of those bean counters. What a stupid and arrogant commentary. The consumer has a lot of laws on his side and especially in the US you better not deliver any sub-spec mercendise declared as being ok. If you buy at the surplus store or Ebay, you might end up with what you are descibing, but this stuff is not guaranteed meeting the specs.

--
ciao Ban
Bordighera, Italy

or

So if it fails within that 2 year period, great; do you believe that all suplies which fail due to an undervolt condition are within their warranty period? Or that simply because a failure DOES occur within the warranty period, that this MUST be indicative of a sufficiently widespread problem such that the supplier WILL be "driven out of business?"

Again, if your model is correct, power supplies that don't meet spec should never be found in the real world. I claim that such things do, in fact, exist. Care to bet which of us can support their position through evidence?

As to "my ignorance really showing" up, you're right - after

25+ years as an engineer working for one of the major computer companies, I clearly don't know squat about the realities of the electronics manufacturing business.

injury

Actually, it isn't, although that is a widespread misconception. The UL standards do not carry the force of law, nor is there any law making it a criminal offense to sell a product which is not UL certified or registered. UL compliance can and will enter into liability awards resulting from failed products, but that is a civil issue, not a criminal one. Further, simply obtaining UL certification basically just says that, IF built as designed from the specified components, the product can reasonably be expected to meet the specification in question. The nature of testing and statistics being what it is, this can never be a perfect assurance of 100% compliance, unless the product in question is subjected to extensive finished-product testing in 100% of the units shipped. Especially for consumer goods, this is basically never the case, with the exception of a few key specifications. And no testing can ever be a 100% guarantee of continued compliance for all time, for the reasons mentioned earlier having to do with aging, electrical and mechanical stress, and so forth.

Yes, but that's not what's being discussed here. Forging a UL mark is irrelevant to the topic under consideration, which is the reliability and failure modes of power supplies.

arrogant

the

Nope; I am not now, nor have I ever been, a "bean counter." It is simply a fact of life that NOT all products shipped will meet their published specifications 100%, and those which do will not continue to do so forever. Again, if you feel otherwise, please call up the head of our service and support organization, and inform them that everyone can go home now.

What published specifications really mean, from a legitimate supplier, is that IF the product delivered does not meet the specifications (which generally can be considered as being a part of the purchase agreement - i.e., you did not agree to buy THIS product, but instead you actually agreed to buy a product which met these specifications), then you have the right to return it and expect a replacement. But simply publishing the specifications by no means is perfect assurance that every product shipped/received WILL actually meet its specs when it reaches the customer. SOMETIMES, this is due to shoddy design and/or manufacturing - but even the best manufacturer in the world for whatever product you're considering still has the occasional failure-upon-delivery. Welcome to the real world.

Bob M.

Isn't the question whether the PC failed or not?

I wouldn't consider the power supply the PC itself. I guess you could argue that point but the power supply is a part that is made to fail, if necessary, to prevent further internal damage to the PC. I've replaced many power supplies on PCs that were struck by lightning (not directly, I'm sure), the power supply was fried but the PC itself was fine after a new supply was installed.

Hi Greg, I'm not going to take the time to read all of the answers you have gotten. Remember what Abe Lincoln said, Beleive only half of what you see and nothing of what you hear. My two cents is this: MOST switching power supplies and probably all used in computers are rated from 90 to 150 volts set to 120v and double that when set to 240. Our engineers, way back in about 1985 built one that was to tolerate 90 to

300 without a switch to set the input voltage. It didn't make it, but was great when the switch was added. Luck, Jim

X-No-Archive: Yes

Do you have the detailed description of failure? Poor quality power supplies have a disgustingly low hold up time and even the best power have a limited hold up time.

Hold up time is the amount of time the power supply can sustain proper outputs when input power is lost. Usually, when power goes out or browns out severely briefly but longer than the hold up time, the computer will reboot suddenly. If it happens at just the right time, it can corrupt the HDD. I've heard of power supplies that's so crappy that computer reboots when UPS tranfers to battery, because it it couldn't hold up during a ~5mS transfer time.

Undervoltage most certainly *can* damage electronics, some poorly designed SMPS's will attempt to compensate for the low line voltage and blow themselves up. Properly designed equipment won't do this, but there's a great deal of poorly designed junk on the market these days.

supply

be

If

I think anyone who actually works on real world products will quickly come to this same conclusion. Like it or not, there's a lot of pure garbage that ends up on store shelves. For every well designed piece of equipment I work on, I come across at least half a dozen things that make me want to find the engineer who designed it and smack them over the head with the thing.

Let's face it, the guy is a troll, spouting out all sorts of rubbish with nothing to back it up. The crossposting doesn't do anything to support his credibility either. It's a clear and proveable *fact* that some power supplies will fail when run at low line voltage, he may wish to stick his head in the sand and refuse to believe that such products are out there but they do exist and I would guess that a great many of them somehow manage to outlast their warranty period before failing, or fail due to other causes first.