graphics cards

[Phaedrus2129]

You really don't get it. The cheap units are cheap for a reason, not because the cost of manufacture has come down. Power supply manufacturers lie, they misrepresent, they cheat, and they cut costs in order to get the highest wattage number for the lowest price with the biggest profit margin possible and damn the competition.

You want to see the type of crap cheap brands pull?

http://www.jonnyguru.com/forums/showthread.php?t=6714&highlight=680W





Power supplies are not like CPUs, period. The same ideas of value and price don't apply. If you don't want to take the advice of someone who knows their ****, fine. It'll backfire. The X7 probably won't since it's a decent unit, my main qualms being on legitimacy and warranty, but if you go out and buy some $50 Apex "650W" and it blows up on you, don't come back to me crying because you'll get a big fat "atojiso" back in your face.

[chimera15]

I know. I watched like every 3dgammerman vid on youtube, and he really went into depth about power supplies and what they actually can do or not, and that they're often misleading.   He always stressed that the rail wattage is the important part, not the wattage on the box.  All said I haven't looked at the rail wattages on that one, but the fact that it's quad sli ready should say something for it.   The problem of course is the seller.

[Phaedrus2129]

That's also a misconception, marking "3dgammerman" a novice. The rail ratings are useful for spotting mid-range units that have been overrated, and for spotting ATX12V 1.x and ATX 1.3 units. But it isn't useful for ATX12V 2.x units, or when manufacturers just lie.

http://hardwareaware.com/review/linkworld-430w/

[chimera15]

Well to be fair I don't think he's done a power supply review for a year or two. lol

[chimera15]

Seriously though.  The likelihood of a power supply failing isn't a concern to me.  If it lasts a year, it's all good.    While I know there is a small chance of it failing and destroying the mobo processor, there's a far greater likelihood my computer just won't turn on one day, or like happened to me one time, my computer will experience brownout type symptoms and intermittent bsods cause my hd's were powering down cause it was failing, and if so I'll test the powersupply, find it's no good and replace it.  It's not like I'm building this machine for someone, if I was I would probably use a higher quality brand.

[Phaedrus2129]

Again, I'm leaving the RF-900 out of this, this is just general principle now.


Very cheap power supplies will blow up, or worse die silently. They also will damage your components, even if they don't blow up. Damage to your capacitors and VRM circuits from out-of-spec ripple is cumulative and irreparable. Bad enough ripple (a la the 3000mV on a certain Linkworld unit) will kill hardware in hours, even things as "simple" as fans.

And a power supply death is worst when it goes quietly. When there's a bang, that means it's a primary-side failure, and that means you have a big ol' transformer between your computer and the failure point. But when the PSU dies quietly or with only a small spark or bit of smoke, that means a secondary side failure, meaning a spike of "dirty" power hitting your PC in the face, killing almost everything on the affected rail (voltage rail, not OCP rail).

And bad PSUs do fail violently.

Skip to 2:00

[chimera15]

It looks to me like they were over volting it to get those results weren't they?  They were drawing 14 and 15 volts on the 12v rails, in addition to an unknown amperages on what rails?   They could have been pushing it far beyond 450 as it looked to me, the thing was completely maxed out.

I read all that.  I would take issue with their testing, as they're not using an actual pc requiring different levels of amperage at different times. I don't think they took into consideration peak wattages vs operating amperage either.   They also tested another power supply, that failed to explode because it had protection circuits in it.   I'm not sure how common or not these are, but I would guess that they're more common than not, since that unit is banned in certain countries.

[Phaedrus2129]

Gabriel Torres is a competent reviewer, and an ATE is superior to a PC for generating load on a power supply. The label is supposed to mark the power supply's continuous operational output. If it can't do that number continuous for at least an hour, then the label is a lie (also consider many units are rated at room temperature of 25C, rather than realistic operating conditions of 30-40C. Gabe and most other reviewers test at 40-45-50C, depending). Also, they have no control over what voltages the PSU outputs, if the voltage was out of spec that was due to the power supply, not the tester.

[williamjoseph]

when it comes to power supplies, i know one thing, if its light, its cheep (aka cheep parts, cheep construction, inferior quality by leaps and bounds). If it feels heavy it is expensive (aka superior quality)  I once bought a power supply from 3btech.net. it was a 3btech branded psu promicing 700 watts for $45. i blew it out on my first run on crysis running this. I proptly replaced it with the PSU listed in linky. no problems since. did i mention that the difference in weight was astounding to myself?

[ch_123]

Clearly Chimera's opinions on PSUs are as sophisticated as his opinions on trackballs.

That, and he has never had one fail on him, taking out two graphics cards in the process.

[chimera15]

Clearly Chimera's opinions on PSUs are as sophisticated as his opinions on trackballs.

That, and he has never had one fail on him, taking out two graphics cards in the process.

You have? I think that's pretty rare. A google search of "power supply failure destroyed graphics cards" returned one result on the subject from 2008.  And that was a pentium 2. lol It's clear you're a snob about psu's as you are about trackballs.  Although I don't know why being left handed would effect your choice of them. lol

I've had lots of power supply failures over the year.  None on c2d+ generations.  None have destroyed components on my boards.  In addition I worked as a repair tech for a short time and my main job was repairing systems that would not boot.  I repaired hundreds of systems.  It was a rare machine that had a power supply problem.  The few that did didn't have any fried components.  It was normally the mobo that got fried from lightning strike over lan or modem lines.  Power supplies are pretty rugged these days, you really have to stress them out, even the cheap ones, to blow them, and even then they rarely take any components with them in my experience.

The fact that there a number of bad ones out there is without doubt.  The fact that there are ones that are banned in certain countries because they don't have standard protections on them, and that they would explode doesn't surprise me, but I think those are pretty rare as well.  There are also batteries that explode and cause fires in laptops, and Toyota's that don't have breaks, but what are the chances of you actually seeing and getting stuff like that? It's like 1 in 1000 odds at best.

Unless you're designing a new model line for a company or something, I don't see the point of knowing about every component of a power supply, and testing it for quality.  If it does the job, and has reasonable protection, it should be good enough.   It's just a power supply, people have been making them for 80 years.  Before computers they were made for radios and junk.  

Hell, the power supply on my apple ][ died on me, in 1979 and didn't take any components with it.  Back then, that was immature technology, and power supply failures were very common.  These days it just doesn't happen that often relative to the number of people using computers and the amount of psu's out there.

[chimera15]

I bought that rocketfish 900w for $53.  The seller responded that it was tested and worked great.  We'll see what I get. lol

[ch_123]

Yeah, if it's not on Google, it must not exist... Oh dear God, people on the internet...

So, how do you explain two graphics cards on two completely seperate computers with only the same PSU in common dying in the same way? (and same PSU itself just refusing to go sometime afterwards)

It is well established that dodgy PSUs kill parts, and cause all sorts of generic wankery. This isn't some myth to make people buy expensive CPUs. I have seen all sorts of computers that would blue screen, reboot randomly, and behave in all sorts of weird ways that was ultimately traced back to the PSU. I was in contact with a guy who's DVD drive would only read CDs and not DVDs. Tried everything and it couldn't work. He tried a different PSU in the computer for the lulz and suddenly it worked again. Puts back in the old PSU, breaks in the same way. That's one snobby DVD drive, amirite?

Amazingly, facts exist separately to your own experience with things. Crazy, I know...

[chimera15]

If you read the thread, I explained that I had seen those same brownout symptoms, so I'm aware of them. That type of failure is more common, but as you stated, putting in a new psu fixed the problem, and the parts weren't destroyed.

That type of failing is a heck of a lot more common than the psu failing and permanently destroying parts.

Google may not be a credible source for reports, but it certainly can be looked at for the frequency of reported events such as this.

There should have been more than one complaint of persons when searching for that term if it was at all common.

[ch_123]

I'm pretty sure I've seen a bad PSU take out some RAM too... Can't remember the specifics though.

Either way, life is too short to play a lottery with low end PSUs, especially when the price of a good PSU is less than the parts that you could end up replacing.

[gr1m]

I'm pretty sure I've seen a bad PSU take out some RAM too... Can't remember the specifics though.

Either way, life is too short to play a lottery with low end PSUs, especially when the price of a good PSU is less than the parts that you could end up replacing.

That was my story. 1 stick of RAM and a motherboard.

[ch_123]

If you read the thread, I explained that I had seen those same brownout symptoms, so I'm aware of them. That type of failure is more common, but as you stated, putting in a new psu fixed the problem, and the parts weren't destroyed.

That type of failing is a heck of a lot more common than the psu failing and permanently destroying parts.

Google may not be a credible source for reports, but it certainly can be looked at for the frequency of reported events such as this.

There should have been more than one complaint of persons when searching for that term if it was at all common.

Even still, I don't want to have to figure out whether there is some software problem, RAM problem, heating problem, PSU problem etc etc. They tend to cause weird problems that can manifest themselves as failures of other parts. In the example that I used about the DVD drive, if that happened to me, I would have gone and bought a new drive unnecessarily. Just a gigantic waste of time and money even if they don't fry parts (which they do).

You seem to imply that this is a regular occurence? Would it not make more economic sense to buy a good PSU that will last as long as the other components last rather than having to replace **** PSUs every once in a while? Those cheap things add up.

[gr1m]

http://www.nvnews.net/vbulletin/showpost.php?p=1531458&postcount=19

http://forums.n4g.com/tm.aspx?high=&m=828595&mpage=1

[chimera15]

Even still, I don't want to have to figure out whether there is some software problem, RAM problem, heating problem, PSU problem etc etc. They tend to cause weird problems that can manifest themselves as failures of other parts. In the example that I used about the DVD drive, if that happened to me, I would have gone and bought a new drive unnecessarily. Just a gigantic waste of time and money even if they don't fry parts (which they do).

You seem to imply that this is a regular occurence? Would it not make more economic sense to buy a good PSU that will last as long as the other components last rather than having to replace **** PSUs every once in a while? Those cheap things add up.

No, because it's a $ over time consideration.   You buy a psu for $50, that you take a chance running continuously lasts 1-2 years, vrs one for $200 that may still fail from various reasons, like lightning strike, with a system's overall lifespan being at most 2-3 years on average.  Even those selling for $200, there's a chance that they're actually repackaged crap.

 It seems like the brownout thing does happen a lot, especially in previous generations.  I haven't had it happen to me with newer c2d's yet, but perhaps it will as the systems I built age.  As a repair tech, not just in that one job, but over 20 years someone building and maintaining at least 5 units at a time in my own house, and more for friends and people that know I know hardware, I've seen a lot of these units with strange brownout symptoms like that.   Not just on ones I've built either. Usually they're on units that had been running continuously for years, and usually it does effect the drives in some strange way, more than the mobo/cpu/graphics card.  Yeah I get your point about replacing the faulty indication.  I've done this as well, but even so, it doesn't always add up. Besides I have a lot of parts for testing faulty systems now. lol  Sides new cd/dvd roms/standard hd drives are super cheap as well now.

I also like to keep a spare good psu on hand to test this, for instance the one that I replaced my mother's failed unit with was that spare one.  So if nothing else this one that I just bought will fit that role.

[ch_123]

Well, for me, my time is worth more than the dollar. If I'm saving up to build a system, I'll gladly make sure to accommodate stuff that isn't going to cause me problems. If I'm spending money, I like to spend it on stuff that works.

[chimera15]

Maybe it's because of my early experience with computers, especially my apple ][ psu failing.  The apple 2 power supply cost like $400 or something ridiculous.  $ doesn't always equal quality.  There are a lot of different factors at play,  and with computers it's a potshot if you get something that works or not, and they fail for seemingly random reason, even reputed good parts.  So there's no reason to pay out the nose for something basic like a power supply if you can meet the necessary requirements of your build for less.

[ch_123]

That's not really a valid comparison, because buying replacement parts from vendors is extortion.

And I don't buy the most expensive PSU on the market. I read around and I find one that is going to be reliable. If it happens to be expensive, so be it.

[gr1m]

So there's no reason to pay out the nose for something basic like a power supply if you can meet the necessary requirements of your build for less.

That's where you're wrong. There is reason, but you just don't see it.

[chimera15]

That's where you're wrong. There is reason, but you just don't see it.

I see it, I just don't agree with it.

[chimera15]

That's not really a valid comparison, because buying replacement parts from vendors is extortion.

And I don't buy the most expensive PSU on the market. I read around and I find one that is going to be reliable. If it happens to be expensive, so be it.

My point is that you can't tell from reviews what is going to be reliable or not.  Companies can pay people to review products, reviews can be wrong, and it just isn't a major factor on an individual basis, as you may get the one car out of 1000 that doesn't have breaks even so.


You're paying out the nose for what people think won't fail 2-3 years down the road, is no different than buying a psu from apple that cost $400, and is a 1000% markup on the parts.  Besides there's a lot of over-engeering that goes into most of those overpriced psu's.  Why pay $200 for a psu that is built to last 10 years, when the average lifespan of a computer is 3.  My mother's system's 8 years old cause she doesn't need much power for what she does, email and whatnot, she's had 1 psu failure in the last few months on an 8 year old system that's been running continuously..and I didn't pay anywhere near a major amount of the original one that failed, and it died cause it was full of dust, not from manufacturer fault.  Her system was never anywhere near a gaming rig of course.




If I were going to build 50000 systems, I would want to review every part that went into a psu to make sure it was quality, but on an individual basis stuff like that just isn't that reliable or important.

To me it's a lot like you're arguing that I should buy a $200 dvdrw, when there are ones out now for $30, that are perfectly fine and do what I want.  The return on the cost from $30 to $200 just isn't that much.

[gr1m]

But $30 DVD drives cannot harm your other parts. I should just stop wasting my time because keep in mind that we're having this argument for your benefit.

[chimera15]

But $30 DVD drives cannot harm your other parts. I should just stop wasting my time because keep in mind that we're having this argument for your benefit.

You saw that Mythbusters where they got cd-roms to explode didn't you?

[gr1m]

No, but apparently, neither did you because you said in your earlier post that cheap DVD drives are fine.

Besides, there's nothing wrong with $30 DVD drives because people use them a lot and not many problems are reported. I use them exclusively myself (the $30 SATA 22x DVDRW drives, be it Asus, Samsung or whatever) and they do not cause undue problems. Why are you being infuriating?

HEY GUYS CHEAP DVDS ARE FINE
YES
NO THEY ARENT CHECK OUT MYTHBUSTERS OLOLOLOLOL

Do you do this to everybody that tries to help you?

[chimera15]

It was a joke. It's not my fault that you read it seriously.  If you had seen the episode you would have laughed.

[/URL]

What's really interesting is reading the comments of this video and seeing how many people are claiming it's happened to them.

[hyperlinked]

To me it's a lot like you're arguing that I should buy a $200 dvdrw, when there are ones out now for $30, that are perfectly fine and do what I want.  The return on the cost from $30 to $200 just isn't that much.

Type less. Read more.

I don't think anyone's arguing that you should always spend more money. I think they're just questioning your rationale that cheap is most likely to be as good or better than expensive. Sometimes you do get what you pay for and that goes for both ends of the price spectrum.

ch_123 and gr1m are explaining why they like to build their systems a certain way. Maybe they don't throw their systems out the door every 2 to 3 years and most people certainly don't ditch their computers every 2 to 3 years. They're making their purchasing decisions based on a lot of the same criteria you use except that if they have reason to believe that a higher priced item is likely to last longer or be safer, it's worth a premium. They didn't say that means they'll buy the $40000000 Apple branded PSU. It simply means that if they believe the cheaper PSU has a 25% chance of ruining or even just reducing the life of something else in their system and end up costing them a few hours to shop and install a new one, they're willing to spend the extra $20 or $40 or whatever they believe to be a reasonable premium for the extra assurance.

It doesn't mean they expect that the more expensive one will always be better or live up to the hype. That is always possible as well and you're correct to point out that a lot of times you're getting cheated out of your money. The difference between what you argue for and what they're arguing for is that if $40 is on the line, you'll pocket the $40 and take your chances. Their point is that if there's a reasonable possibility that the $40 will save them extra hassle or cost, they'll give up the $40 for the peace of mind.

You're simply placing different bets. You can be cheap and get burned. You can be lavish and get burned. The fact that you did or didn't get burned doesn't mean someone was right or wrong. Nobody's being an idiot. Not you. Not them.

Just for once, try this. Say "I can understand why you'd want to do that, but I think you're wasting your money. It's not the way I'd build my systems, but to each his own."

[ch_123]

My point is that you can't tell from reviews what is going to be reliable or not.  Companies can pay people to review products, reviews can be wrong, and it just isn't a major factor on an individual basis, as you may get the one car out of 1000 that doesn't have breaks even so.

It isn't just as black and white as 'How long does it last?'. Bad PSUs will have all sorts of issues - such as inability to deliver correct amperage under load, excessive heat, weird noises when pushed under load, and of course complete and utter failure to deliver the rated wattage - either where the system just turns off, or the PSU burns out, complete with smoke, loud bangs, and sometimes flames. All these can be tested for out of the box, and if your PSU can get over all that, the odds are that it's going to last quite a good deal longer than one that doesn't.

You mention issues with the objectivity of reviews, and how they are limited by the quality of the one unit they receive. This is true. This is why I read every review I can find, and also read anecdotal evidence from people on the internet who make use of them. And usually you can spot the sellout reviewers - usually the ones that give everything five stars.

Ultimately, you seem to be saying that buying good PSUs is pointless because they could fail within 2-3 years. Well, nothing in life is certain, so yeah, they could fail. But they are an awful lot less likely to explode into flames after a few moths compared with some 'whitebox' junk. Besides, 5 year warranty on a Corsair-brand Seasonic unit? Yes please!

I think you're making an argument where there is none.

You're paying out the nose for what people think won't fail 2-3 years down the road, is no different than buying a psu from apple that cost $400, and is a 1000% markup on the parts.

Ehh, no.

Seriously, unless you are prepared to qualify that statement, it's like "the grass is not green".

Besides there's a lot of over-engeering that goes into most of those overpriced psu's.  Why pay $200 for a psu that is built to last 10 years, when the average lifespan of a computer is 3.  My mother's system's 8 years old cause she doesn't need much power for what she does, email and whatnot, she's had 1 psu failure in the last few months on an 8 year old system that's been running continuously..and I didn't pay anywhere near a major amount of the original one that failed, and it died cause it was full of dust, not from manufacturer fault.  Her system was never anywhere near a gaming rig of course.

Asides from the fact that computers these days have a much longer useful life than 3 years, you seem to contradict yourself by talking about 8 year old computers.

The fact that your mother's PC isn't some mad gaming rig illustrates the point further. If a bad PSU is going to flare out on some low end computer, what chance is it going to stand with a high end CPU/Graphics card or loads of hard drives.?

If I were going to build 50000 systems, I would want to review every part that went into a psu to make sure it was quality, but on an individual basis stuff like that just isn't that reliable or important.

Shame that major OEMs don't think like that.

To me it's a lot like you're arguing that I should buy a $200 dvdrw, when there are ones out now for $30, that are perfectly fine and do what I want.  The return on the cost from $30 to $200 just isn't that much.

As pointed out above, the DVD drive isn't going to kill the expensive parts of your computer when it fails.

[chimera15]

I agree, this is really an argument about nothing.  It's just a matter of my experience building units, and what I think the failure rate and chances I'm willing to take of lower end psu's and people who have the money to spend more on building a better unit.  The difference between paying $750 total, and $900 is a breaker for me to build this system.  I don't have $900 to spend, I have $500 really, and I'm pushing the limits to get this unit built at $6-700.


As I've already said, I don't want to build some rock solid ultra gaming computer, (well I do, I just don't have the money to do it and I know it) just an average gaming computer that will put me up with modern times, at a reasonable price.  I'll admit that I don't have a lot of money to spend, and a psu is a place I can skimp on because the likelihood is that that that $50 900watt off ebay will work perfectly fine for me for a few years, and that's all I really care about right now.  The likelihood of the psu failing is slim in my experience, especially if I don't go with an 2x 9800gt sli solution, which brings this thread back to what it's really about, and the main question.  Will one 460 beat dual 9800gt's, which I've yet to be able to answer.

If I was building this system for someone else, and they had the money to spend, obviously I would go with a more solid supply.  As it is, I'm willing to take the risk on that unit, and if it goes bad, I know enough about the issue that I'll likely be able to repair the system.

To answer the other thing about 8 years vrs 3 years.  Gaming computers are obsolete after 3 years, probably more like 1 at most if you're a serious hardcore gamer.   Most of my family runs my hand me down gaming computers. Email/browsing computers have a much longer lifespan.  You can still use p1's to browse and do email from like what, 15 years ago.  Obviously the class and stress put on the psu is completely different as well.

[gr1m]

Oh, if you're only gaming, there's no reason to go i7.

[chimera15]

Oh, if you're only gaming, there's no reason to go i7.

I know, that's another consideration, but I think it's the best bang for the buck right now. I could probably build a c4d gaming system for $300 right now, that would meet any requirement out there, but I am doing 3d/photoshop/editing stuff as well occasionally, but gaming is the main point I have for the system, cause my current c2d's are just slightly being overwhelmed on the highest settings for the games I play and plan to.

I can't imagine New Vegas is going to be extremely intensive on a system either, but I normally like to build a new system when benchmarks are 2-4 times what I currently own, and the benchmarks are showing that clearly.  I'd be aiming for an i7 that can get 300k crystalmarks which I think is very doable for a reasonable price.

[gr1m]

I would recommend a Phenom II 955/890FX platform. You'll save money, overclock the **** out of it, lose no performance in gaming when compared to i7.

[Phaedrus2129]

I read the last few pages, skipped this one, blah blah. Just one thing stuck out, about not being able to find good reviews? Here you are.
http://www.overclock.net/power-supplies/738097-psu-review-database.html
Shinji2k went to all the trouble of listing every "good" review from every reliable (ie., has testing equipment, knows what they're doing, not paid off) reviewer in the industry. Not every power supply has been reviewed, of course, but there are a couple of hundred units listed there with competent testing, dissections, and informed opinions.

[chimera15]

I read the last few pages, skipped this one, blah blah. Just one thing stuck out, about not being able to find good reviews? Here you are.
http://www.overclock.net/power-supplies/738097-psu-review-database.html
Shinji2k went to all the trouble of listing every "good" review from every reliable (ie., has testing equipment, knows what they're doing, not paid off) reviewer in the industry. Not every power supply has been reviewed, of course, but there are a couple of hundred units listed there with competent testing, dissections, and informed opinions.

I was talking about graphics cards I think, but nice site.  They don't have the rf900, but they have the 700...that one isnt modular though.

[chimera15]

I would recommend a Phenom II 955/890FX platform. You'll save money, overclock the **** out of it, lose no performance in gaming when compared to i7.

No, not into amd stuff, except maybe in laptops.  Not really into overclocking to extremes either.   I don't see risking the equipment, or at least the stability of the system for a slightly better framerate.  Don't have the money to do that either.  I could probably start pushing my current c2d's and meet the requirements I need for the games as well, but don't want to risk the hardware as I'll probably hand this stuff down.  It's why I don't need giant psu's.

[Phaedrus2129]

I was talking about graphics cards I think, but nice site.  They don't have the rf900, but they have the 700...that one isnt modular though.

The RF-900 is listed as the Huntkey X7 900W.

[gr1m]

No, not into amd stuff, except maybe in laptops.

Wrong again. Intel is clearly superior to AMD when it comes to mobile products, so if you really want to be a fanboy, laptops are the best place to do it. For gaming desktops, the difference becomes trivial.

Besides, what does "not into AMD stuff" even mean? Are you into wasting money? I wouldn't think so based on your reluctance to buy a decent power supply. So why do you want to spend $100 more on CPU power you don't need, and refuse to spend $50 more on a power supply. Overclocking doesn't cost too much money. A $30 Scythe Mugen 2 is enough to push any Phenom II (maybe even the 6-cores but I haven't played with one yet) to 4GHz.

[Phaedrus2129]

Doesn't like admitting being wrong or misinformed, I believe.

[ch_123]

AMDs are unreliable, man.

*coughs*

[EverythingIBM]

You have? I think that's pretty rare. A google search of "power supply failure destroyed graphics cards" returned one result on the subject from 2008.  And that was a pentium 2. lol It's clear you're a snob about psu's as you are about trackballs.  Although I don't know why being left handed would effect your choice of them. lol

I've had lots of power supply failures over the year.  None on c2d+ generations.  None have destroyed components on my boards.  In addition I worked as a repair tech for a short time and my main job was repairing systems that would not boot.  I repaired hundreds of systems.  It was a rare machine that had a power supply problem.  The few that did didn't have any fried components.  It was normally the mobo that got fried from lightning strike over lan or modem lines.  Power supplies are pretty rugged these days, you really have to stress them out, even the cheap ones, to blow them, and even then they rarely take any components with them in my experience.

The fact that there a number of bad ones out there is without doubt.  The fact that there are ones that are banned in certain countries because they don't have standard protections on them, and that they would explode doesn't surprise me, but I think those are pretty rare as well.  There are also batteries that explode and cause fires in laptops, and Toyota's that don't have breaks, but what are the chances of you actually seeing and getting stuff like that? It's like 1 in 1000 odds at best.

Unless you're designing a new model line for a company or something, I don't see the point of knowing about every component of a power supply, and testing it for quality.  If it does the job, and has reasonable protection, it should be good enough.   It's just a power supply, people have been making them for 80 years.  Before computers they were made for radios and junk.  

Hell, the power supply on my apple ][ died on me, in 1979 and didn't take any components with it.  Back then, that was immature technology, and power supply failures were very common.  These days it just doesn't happen that often relative to the number of people using computers and the amount of psu's out there.

I was reading some of the PSU debates, and just want to share my little piece of cheese. I agree with chimera -- you can get a good deal on PSUs even if it's a lower price; there's always someone who's just selling it to get rid of it, rather than trying to rip people off; I never once got "cheated" on ebay, everything I got was exactly as described.
I've never had a power supply stop working on me before, and I've got some fairly old & abused computers (the dust in the PSU on the 300PL I got was phenomenal -- some of it caked on so hard that I couldn't get it off even with the strongest compressed air can). But it still works fine and probably will for years to come.

There is a limit to how "cheap" manufacturers can make PSUs, as, it has to fit within a certain safety guideline.

[chimera15]

Wrong again. Intel is clearly superior to AMD when it comes to mobile products, so if you really want to be a fanboy, laptops are the best place to do it. For gaming desktops, the difference becomes trivial.

Besides, what does "not into AMD stuff" even mean? Are you into wasting money? I wouldn't think so based on your reluctance to buy a decent power supply. So why do you want to spend $100 more on CPU power you don't need, and refuse to spend $50 more on a power supply. Overclocking doesn't cost too much money. A $30 Scythe Mugen 2 is enough to push any Phenom II (maybe even the 6-cores but I haven't played with one yet) to 4GHz.

It means I hate the intel graphics cards in laptops, that intel always seems to want to package with the majority of their processors. There's exactly 1 intel based tablet pc(which I own) that has ever been produced with a dedicated graphics card, and that was before the amd/ati partnerships. (although the new tm2's do have ulv's(yuck) and ati graphics cards. )


I'm not into AMD desktops cause I've owned early ones, couldn't stand them, seen and had tons of problems with them, and not willing to sacrifice capability for cost.  A cheap power supply doesn't sacrifice any of the capabilities of the system, which an amd motherboard/processor would, except perhaps extreme overclocking, which I'm not going to do anyway.

[EverythingIBM]

AMDs are unreliable, man.

*coughs*

Lefties...

=)

[Phaedrus2129]

There is a limit to how "cheap" manufacturers can make PSUs, as, it has to fit within a certain safety guideline.

No they don't. Only cert they have to make to be sold in the US is FCC, which says nothing on safety or quality, just makes sure it won't interfere with other devices. Leadman LP8860 units cost less than $5 to produce, and it shows.

Also, there's a difference between getting a deal on a nice unit on ebay, and buying a generic piece of ****.

It means I hate the intel graphics cards in laptops, that intel always seems to want to package with the majority of their processors. There's exactly 1 intel based tablet pc(which I own) that has ever been produced with a dedicated graphics card, and that was before the amd/ati partnerships. (although the new tm2's do have ulv's(yuck) and ati graphics cards. )


I'm not into AMD desktops cause I've owned early ones, couldn't stand them, seen and had tons of problems with them, and not willing to sacrifice capability for cost.  A cheap power supply doesn't sacrifice any of the capabilities of the system, which an amd motherboard/processor would, except perhaps extreme overclocking, which I'm not going to do anyway.

We just had a huge discussion on all the ways a cheap PSU screws you over. And now you say there's no difference? Failed reading comprehension, perhaps?

[chimera15]

No they don't. Only cert they have to make to be sold in the US is FCC, which says nothing on safety or quality, just makes sure it won't interfere with other devices. Leadman LP8860 units cost less than $5 to produce, and it shows.

Also, there's a difference between getting a deal on a nice unit on ebay, and buying a generic piece of ****.



We just had a huge discussion on all the ways a cheap PSU screws you over. And now you say there's no difference? Failed reading comprehension, perhaps?

Reliability doesn't equal capability.

[ch_123]

It means I hate the intel graphics cards in laptops, that intel always seems to want to package with the majority of their processors. There's exactly 1 intel based tablet pc(which I own) that has ever been produced with a dedicated graphics card, and that was before the amd/ati partnerships. (although the new tm2's do have ulv's(yuck) and ati graphics cards. )


I'm not into AMD desktops cause I've owned early ones, couldn't stand them, seen and had tons of problems with them, and not willing to sacrifice capability for cost.  A cheap power supply doesn't sacrifice any of the capabilities of the system, which an amd motherboard/processor would, except perhaps extreme overclocking, which I'm not going to do anyway.

As someone who been using AMD in their desktops for about 5 years, I respectfully suggest that your ****ty PSUs might have been to blame. Or perhaps you apply the same rationale to motherboards. Either way...

And yay to Intel for laptops, if only to get their Wifi cards...

[Phaedrus2129]

On the topic of PSUs taking out other hardware when they go, would it be possible to create...um, filters to put in between each connector and the target device (inlet and outlet plugs on said filters)?

I don't imagine such a thing would need to be too complicated. I'm not an EE guy but I think capacitors can have a smoothing effect and eat up overvoltages, right?

When connected in parallel, yes capacitors can filter ripple and noise, but they do nothing to stop overvoltage or a surge. You can also get that to a lesser extent with ferrite beads. Antec does that on some of their very high end models, ferrite rings and capacitors on the end of cables. It screws with cable management and looks ugly, so they've mostly stopped. Voltage spikes might be filtered with a metal oxide varistor, but those are bulky and can only do so much. Preventing overvoltage would require a voltage regulator... basically another power supply.

It's the designer/manufacturer's job to make sure there's sufficient filtering inside the power supply. If they're cheap enough to just chop out the caps and transient filter then they shouldn't be supported anyway.

[EverythingIBM]

So chimera, when's the PSU coming? I want to see how well it all goes.