1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

PSU Guide

Discussion in 'General Hardware' started by panchoman, Sep 13, 2007.

  1. de.das.dude

    de.das.dude Pro Indian Modder

    Joined:
    Jun 13, 2010
    Messages:
    7,474 (5.02/day)
    Thanks Received:
    1,936
    weird thing, my power went off for a second and came back(they might have switched grids)
    i was just browsing the internet, and nothing happened! all the lights and stuff went off and on, even the TV, but my CPU managed to live off just the charge in the caps?
  2. d1nky

    d1nky

    Joined:
    Jan 21, 2013
    Messages:
    3,557 (6.65/day)
    Thanks Received:
    1,204
    lol thats the extreme mobo you have!
    de.das.dude says thanks.
  3. McSteel

    McSteel

    Joined:
    Nov 19, 2012
    Messages:
    562 (0.94/day)
    Thanks Received:
    269
    Now I feel called out :p

    The older unit Frick linked to is not entirely relevant in this case. That design (PC7016) may be similar to the one used in the iPower 85H 750 (PCA011), but the latter has been tweaked so as to hold up even at 50°C intake (even though that's not AcBel's official standing).

    700W is nearing the platform's design limits, with 12V hitting ~95mV of R/N and drooping to ~11.7V. However, it is sufficient for 2x "HD7930", even in GPGPU 100% usage (like cryptocurrency mining or something useful). Longevity might be an issue, since the PSU uses a grabbag approach to filter caps - you may find Teapo, Aishi, Ltec, Elite, Lelon, all depending on the batch and your luck. That paired with a Yate Loon sleeve bearing fan could mean that you shouldn't leave your rig unattended, and that you should probably pay extra attention to dust buildup in the PSU. The minor rails are completely overlooked in this PSU, so don't load them with more than 15A if you don't really have to. 5VSB is perfectly fine even with over 3A, though.

    Other than the caveats above, your PSU is adequate for the task.

    The 85H is a "retail adaptation" of what AcBel makes for Dell desktops. Though Dell uses LiteOn and Hipro as well, and sometimes Delta in their home/light gamer machines.
    Frick and silkstone say thanks.
  4. silkstone

    silkstone

    Joined:
    Nov 1, 2008
    Messages:
    2,821 (1.36/day)
    Thanks Received:
    474

    That's a great summary, thanks for spending the time on it. I understand much clearer now. There were a couple of other reviews of the unit, one Chinese, and one local.
    But, I didn't feel too confident about their reliability.

    By the minor rails, you mean the +5v and +3.3v?

    I'm running 1x SSD, 2X HDD and 1x Laptop HDD along with the overclocked i5 and motherboard.
    Would that be putting much load on the minor rails? I'm not fully sure what uses the 3.3v and 5v rails apart from the cpu.

    My main concern is the safety. If it dies, I don't mind grabbing another as they are about 1/2 the price of equally rated seasonic/corsiar etc. units here. (They are only sold at a few places for a big markup) But, I don't want to have to worry about it taking out the other components. It's kind of why I went back to them as my old (crappy one) failed gracefully and just stopped being able to accept a high load on the 12v rail. From what I read, the safety seems fine, but I'm a little worried about the 5v rail now.
  5. McSteel

    McSteel

    Joined:
    Nov 19, 2012
    Messages:
    562 (0.94/day)
    Thanks Received:
    269
    Yes, by "minor rails" I mean 3V3 and 5V ones. 3V3 is used primarily by RAM, some GPUs even use it for their own VRAM, and by some PCI/PCI-E expansion cards, mainly for minor logic ICs. 5V is used by HDD/ODD logic, USB devices, major logic ICs (audio, network, PCI-E switchers, TV, modem, etc), SSD logic and memory, and signaling diodes. CPU uses 12V exclusively as of s478/s462 era, as do PCI-E VGAs. Some AGP cards use 5V, but they're not exactly "big spenders".

    With what you have, you won't see more than about 40W total 5V + 3V3 load, worst case. So no worries there. Large RAID arrays and a forest of USB devices can get the minor rails usage pretty high in some cases, but that's way under 1% of the population.

    As for safety, AcBel typically makes a point of it. And even though the model you're using has Weltrend's protection/control ICs (as opposed to superior Silicon Touch and Sitronix parts), it will not fry your components if and when it goes out.
    silkstone says thanks.
  6. silkstone

    silkstone

    Joined:
    Nov 1, 2008
    Messages:
    2,821 (1.36/day)
    Thanks Received:
    474
    Thanks for that. Makes it much clearer. I've been thinking of getting an extra 2TB WD Green, but then I will probably pull out the (failing) 500GB drive too.

    I guess that I should open the thing up, give it a good clean, and replace the fan as soon as it is out of warranty. I used to be a smoker and I live in a dusty city, my computer regularly gets clogs up with dust.
  7. de.das.dude

    de.das.dude Pro Indian Modder

    Joined:
    Jun 13, 2010
    Messages:
    7,474 (5.02/day)
    Thanks Received:
    1,936
    how come? bigger mobo does mean bigger consuption right?

    im guessing the voltage didnt totally drop to 0, it just went to 90-100v and then back up.
  8. silkstone

    silkstone

    Joined:
    Nov 1, 2008
    Messages:
    2,821 (1.36/day)
    Thanks Received:
    474
    Yea, Good PSU's can handle extremely brief power outages, just as long as you aren't doing anything extreme on your PC.

    You really should invest in a UPS with dodgy power as it'll start to kill the PSU. I've had lots of things in my house die from dirty power.
    Luckily, the apartment I've been living in for the last 5 years or so has a decent electrical system, so I don;t have to worry any more, but we still get random power outages.
  9. de.das.dude

    de.das.dude Pro Indian Modder

    Joined:
    Jun 13, 2010
    Messages:
    7,474 (5.02/day)
    Thanks Received:
    1,936
    no dodgy power...
    just a grid change or two.
  10. westom

    Joined:
    Dec 5, 2009
    Messages:
    180 (0.11/day)
    Thanks Received:
    33
    Describe a part at risk when a PC is doing something 'extreme'. And define 'extreme'.

    Posted previously was how so many *knew* heat was destructive. Speculation said so. Many did not learn that heat is a diagnostic tool to identify an already defective part. Many would blame heat rather than the previously defective part.

    This 'extreme' concept is also 'subjective'. Speculation exists when one did not learn by actually measuring current - learning the numbers. What number defines 'extreme'?

    What part is 'destroyed' by dodgy power? And what is dodgy power? Another subjective claim; devoid of perspective - the numbers. Anything that UPS might do to protect a power supply is already inside a PSU. Otherwise normal power cycling would also cause hardware failures.

    This UPS outputs pulses that peak at 70% higher than AC RMS voltages. Those pulses are why a UPS is potentially harmful to motors. Same pulses are ideal power for all electronics. The anomaly put into perspective with numbers. How does that UPS protect hardware when it output so 'dirty' power?

    Why did a power interruption not affect that computer? A requirement defined by specifications that existed even with the original IBM PC. Specifications require a computer to be more robust than other electronics.

    Long before recommending a solution, first define the anomaly with numbers. Many boxes are recommended to only do what is already solved inside electronics. Dodgy and extreme require perspective - defined by numbers. Exactly what part is at risk from dodgy or extreme?
  11. de.das.dude

    de.das.dude Pro Indian Modder

    Joined:
    Jun 13, 2010
    Messages:
    7,474 (5.02/day)
    Thanks Received:
    1,936

    :pimp: that is some deep shit yo :pimp:
  12. westom

    Joined:
    Dec 5, 2009
    Messages:
    180 (0.11/day)
    Thanks Received:
    33
    How to quickly identify a scam. That recommendation did not say why. And provided no numbers. That's basic layman knowledge. At least with layman who quickly identify and avoid myths and cons.
  13. silkstone

    silkstone

    Joined:
    Nov 1, 2008
    Messages:
    2,821 (1.36/day)
    Thanks Received:
    474
    I don't know what you are on, but can I have some?

    By extreme, I mean high load. For example video encoding, gaming, etc.

    UPS also generally eliminate noise, spikes, surges and lightning strikes that can hurt sensitive electronics.

    Furthermore, I've had 3 hard drives fail due to power outages causing the PC to lose power.

    If you want to pay me to spend time formulating a thesis and doing objective testing, you'll have to pay me. Until then, you are gonna get 'subjective' talk.

    Tell me what colour the sky is in an objective, with experimental data and references, tell me how long it takes you. Then you'll understand the absurdity of your post.
    Last edited: May 19, 2013
  14. Frick

    Frick Fishfaced Nincompoop

    Joined:
    Feb 27, 2006
    Messages:
    10,380 (3.40/day)
    Thanks Received:
    2,083
    Uh, gaming or encoding isn't extreme. I'd say extreme is when you push the PSU beyond it's rating for prolonged periods of time.
  15. silkstone

    silkstone

    Joined:
    Nov 1, 2008
    Messages:
    2,821 (1.36/day)
    Thanks Received:
    474
    Hi - load then. If you're drawing a high amount of power and the power goes out for a millisecond, then your computer is more likely to reset than when you are drawing a lower amount of power. Maybe extreme was the wrong word, but it's late here and I couldn't think of a different way to say it.

    [Edit] - Try it, I have, in the past. Run OOCT and flick the power switch on and off real quick, you PC will probably loose power. Do it while sitting at the desktop and your pc will not bat an eyelid.
    Last edited: May 19, 2013
  16. westom

    Joined:
    Dec 5, 2009
    Messages:
    180 (0.11/day)
    Thanks Received:
    33
    Many power supply parts suffer less stress when load increases. Unfortunately that hard reality is contrary to speculation that assumes more power means more strain.

    Learn from numbers. Changes you have assumed are major are actually small. Most of the power consumed by a computer occurs both when doing 'extreme' and when doing nothing.

    So which part suffers significantly increases stain during gaming? And why?

    UPS does not eliminate noise, spikes, surges, lightning strikes, etc. However if you know otherwise, then post specification numbers that say so. Your subjective claims have no merit without hard facts - ie numbers. Some of the worst noise and spikes seen by a computer are generated by a UPS. Also why a UPS may harm an electric motor or power strip protector.

    You have only assumed disk drives destroyed by a power loss. How does power loss differ from power off? It doesn't. To a disk drive, all power offs look exactly same.

    Now had you said which part in each drive was damaged, then maybe the claim had merit. Speculated was power loss caused damage. A conclusion based only in observation is classic junk science. You used same speculation to assume a UPS does 'cleaning' that the manufacturer even does not claim. That speculation is mostly found where advertising replaces hard science.

    Since you know 'extreme' means more stress, then describe which part is being stressed. And why a UPS averts that stress. Since you know power loss damaged disk drives, then explain why same power off (called shutdown) did not do that damage. Since you know a UPS does all that power 'cleaning', then let's see those numbers. Otherwise it's just another myth or con.
  17. silkstone

    silkstone

    Joined:
    Nov 1, 2008
    Messages:
    2,821 (1.36/day)
    Thanks Received:
    474
    I am talking about loos of power from the mains, not stress.

    A PSU is neither a motor or a power strip. Show me how a UPS harms a power PSU. And show me how they do not regulate noise.


    It damaged the drive as the drive was reading/writing/spinning at the time. With a normal power off, the heads park themselves and the platters stop spinning. With an abrupt power cut they do not. One of the damaged hard disks was on my laptop, as I used to run it without a battery. When the power died, the hard disk got a load of bad sectors that built up with each power cut. I always run it with the battery in now, and never had another problem with the HDD and bad sectors before switching to an SSD. Anecdotal again. But if you want to refute it, you do the tests and risk your own hardware.

    I was not going to open my HDD and forensically examine it, the RMA was more important.

    I never said extreme means more stress. your second point is repetition which I have addressed above.

    I understand that the Active PFC is meant to filter out all the nastiness in the power, but when equipment dies, and fuses blow due to power spikes I don't see the problem in adding a second layer of protection.

    When switching to battery power, a UPS can be harmful to a PSU, agreed. Some PSUs won't work with certain UPS's. It's why you should buy a decent one.

    http://superuser.com/questions/113113/why-are-brownouts-so-harmful
    http://www.jonnyguru.com/forums/showthread.php?t=3964

    I'm really not going to spend any more time looking into it, unless you can show me that a ups is more harmful than "dodgy" input power.

    Oh, and where I am the power can be very fickle. You sometimes notice the lights getting brighter or dimmer for periods of time.
    Last edited: May 19, 2013
    AsRock says thanks.
  18. westom

    Joined:
    Dec 5, 2009
    Messages:
    180 (0.11/day)
    Thanks Received:
    33
    Learn how drives worked even when heads were moved with motor oil. Every drive (back then and today) first learns about power off when voltage starts dropping. Only then does it stop writing and park heads. All power offs (shutdown, yank the power plug, nuclear power plant shutdown) appear same to all disk drives. If AC power loss causes disk drive damage, then so does shutdown.

    Power loss from mains causes no damage. International design standards over 40 years ago even demanded it. All low voltages were defined in a chart that contained this phrase in all capital letters: No Damage Area.

    Low voltage can be harmful to motorized appliances. Low voltage and AC power loss do not damage electronics. If you know otherwise, then cite a threatened part. And manufacturer datasheet that defines that threat.

    Power loss from AC mains does not damage electronics or a disk drive. As was true over 40 years ago when I was designing or analyzing failures on this stuff (even on one drive that moved head with motor oil).

    Please grasp why a UPS can damage a motor or power strip. Because 'clean' power from a UPS is often a myth. Some of the 'dirtiest' power in a building comes from a UPS when in battery backup mode. That 'dirty' was even defined with numbers.

    Or more facts. View same from a utility:
    http://www.duke-energy.com/indiana-business/products/power-quality/tech-tip-03.asp
    Figures 1 and 2 demonstrate 'dirty' power from a UPS. 'Dirt' made completely irrelevant because electronics are so robust.

    Active PFC addresses one problem created by electronics. Active PFC does not filter out 'nasties'. Which 'nasty' are you referring to? Numerous anomalies exist. Each one has a name. PFC is about electronics doing less harm to AC mains while increasing efficiency.

    'Nasties' is another "subjective" term often promoted when basic electrical concepts are unknown.

    Why do some PSUs not work with some UPSes? Figures 1 and 2 from that utility demonstrate the problem. 'Dirty' power does not harm the UPS. But can confuse active PFC. So the power supply simply powers off - without any damage.

    Figure 1 and 2 provide numbers that separate myths and hearsay from what actually exists. UPS does not 'clean' power as hearsay so often claims. UPS does not protect hardware as is so often promoted by advertising (subjective claims). Power loss and power off are same and not destructive to every disk drive.

    What did the Wikipedia quote forget to mention? Anomalies are also made irrelevant by what exists in PSUs. To say more means each anomaly defined by a number. Your every concern is based only in subjective reasoning - what also creates junk science.

    UPS does near zero hardware protection. Sufficient for advertising to claim 100% protection - subjectively. Too many are informed by advertising; not by spec numbers and hard facts. Again, if you know otherwise, then numbers are in every claim.
  19. westom

    Joined:
    Dec 5, 2009
    Messages:
    180 (0.11/day)
    Thanks Received:
    33
    Normal power for all electronics is even when incandescent bulbs dim to 50% intensity. As even defined in international design standards. Computers are required by ATX standards to be even more robust.

    Meanwhile, dimming lights are sometimes an indicator of a human safety threat. Fix the reason for dimming rather than cure it with a UPS.
  20. silkstone

    silkstone

    Joined:
    Nov 1, 2008
    Messages:
    2,821 (1.36/day)
    Thanks Received:
    474
    I don't know. You are one of the only people I know that says power outages can't cause bad sectors and that UPS are bad for a PC.

    What UPS are those graphs from. It would be wrong to say that all are the same.

    https://www.google.com/search?q=sud...ome.0.57j62l2.7511j0&sourceid=chrome&ie=UTF-8

    The other thing that my UPS does is allow me to ground my computer, so I don't get shocked every time I touch a metal part, and it grounds my LCD, printer and case individually.

    I've not seen objective reasoning either way, apart from running a PSU on a 240v square wave system is a bad idea. So until I do, i'll stick with my anecdotal evidence.

    http://www.electrosafe.co.nz/power2.htm

    Computer glitches, lockups and hardware damage can be the result of poor power quality. Different types of problems will have different effects on the operation, or even life expectancy of equipment. Spikes, Dropouts, Power Failure or Blackout, Power sag or surge, High frequency noise, Normal and/or Common mode noise are all quality problems that affect the quality of the AC power being supplied. Knowing what these problems are, and what some of the consequences are of having these problems, can help in the process of identifying what can be done to help protect equipment. Some are caused by Lightning, Electrostatic Discharges, Ground Surges and Faulty wiring. With the explosion of computer network technology, computer systems are also susceptible to network and phone line surges, as well as the various power supply surges .

    More: http://www.nwasco.com/faqs.cfm?CategoryID=7
    More, with a table showing how damage to HDDs can occur: http://www.aelgroup.co.uk/pdf/wp_powqal.pdf
    More: http://www.controlledpwr.com/help-regulator-technologies.php
    More: http://www.criticalpowerresource.com/9problems.pdf
  21. westom

    Joined:
    Dec 5, 2009
    Messages:
    180 (0.11/day)
    Thanks Received:
    33
    How many designed power supplies? How many worked on disk drives even a decade before the IBM PC existed? How many fixed computers by literally tracing each fault to an IC. Only replacing that defective semiconductor. And then defined, in design reviews, why that part failed. We even opened semiconductors to use a microscope; to determine why a failure happened. How many never did any of this and still make recommendations? How many read datasheets before concluding? Or did most recite popular hearsay?

    Well, there was once an 8" floppy drive from Interdata that was designed defectively. Therefore power on or a transient would corrupt some sectors. An example of why Interdata did not last very long. Sector corruptions was unacceptable even that long ago.

    Obviously, irrelevant is which UPS created that Figure 1 an 2 waveform. Apparently you still miss the point. A third time: that 'dirty' UPS power and even 'dirtier' UPS power is made irrelevant by circuits already inside all electronics. Routine is for a UPS to create 'dirtiest' power. Since electronics are already so robust. To even protect from 'dirtiest' UPS power.

    Many computer 'experts' have no electrical knowledge. So overseas manufacturers can dump into the market supplies that are missing essential functions. All they need do is withhold spec sheets. Then it is legal. They do so because, as you have demonstrated, so many make claims while not even citing one spec number. A computer can crash on 'dirty' power because a computer assembler selected defective hardware. Because that assembler did not even know why every power supply makes even 'dirtiest' UPS power irrelevant. Because that assembler ignored spec numbers.

    Why are scams so easily promoted? Immediate suspect a scam when claims are only made subjectively.

    Well, more citations that only made subjective claims. Meaning no useful facts. A symptom of junk science reasoning. Also called advertising or propaganda.

    You claimed subjectively, long ago, what a UPS might do. Were challenged to provide spec numbers for each claim. Spec numbers. How many posts and still no technical numbers? An example of why so many recommend a UPS to do what even the manufacturer does not claim it does.

    Well one citation did provide some numbers. Its numbers, repeatedly, contract what you have posted.

    Another number was 88.5% of AC power problems were transient related. Let's add relevant facts. 1) A potentially destructive transient typically occurs once every seven years. 2) A transient that no adjacent UPS even claims to protect from.

    No facts (numbers) demonstrate an adjacent UPS protects hardware. Not even defined is one 'at risk' part. A fact that must be known BEFORE making a recommendation. Meanwhile, industry standards say internal protection inside electronics even protect from 'dirtiest' UPS power. And says so with numbers.

    The concern is a rare transient that can overwhelm that protection. That solution is different, located elsewhere, is proven by over 100 years of experience, is recommended for all homes, and is only understood by an informed minority.
  22. silkstone

    silkstone

    Joined:
    Nov 1, 2008
    Messages:
    2,821 (1.36/day)
    Thanks Received:
    474
    A UPS power is only dirty when running on batteries. Even then it is cleaner than the power in this country.

    I think the transient event was once every 7 years in the UK. - they also say this;

    "In 1974 most ICT equipment used linear power supplies which are largely unaffected by small
    transients and completely immune to most forms of noise. Today most ICT equipment uses switch mode power supplies whose susceptibility to transients and common mode noise is far greater."

    Also this from another site: "Spikes and/or surges are sudden increases in voltage that last less than 1/60th of a second. Large spikes can burn or melt electronic components causing instant failure. Repeated “hits” from small spikes gradually erode components and can shorten their useful life. Most spikes originate inside your home or business and are caused by other electric equipment such as when the compressor in your refrigerator kicks on, or when copy machines, laser printers and large motors are turned on and off. A smaller percentage of spikes originate outside your building, caused by lightening strikes, short circuits in electric lines and large equipment used in businesses. While these spikes are less frequent, they tend to be more serious. Lightning strikes to a telephone line, cable TV system or satellite dish also carry voltage spikes. Devices that are connected to more than one wire, such as cordless telephones, answering machines, computer modems and TV’s are especially vulnerable."

    "Uninterruptible Power Supplies or UPS provide a battery backup for computers. UPS devices insure power quality during sags and interruptions protecting both data and hardware. Most UPSs provide 10-16 minutes backup time, allowing you time to close files and exit programs. Most UPSs provide surge suppression as well. A UPS must be UL listed.

    Retail outlets usually have a good selection of surge suppressors, and UPSs can be purchased at computer equipment stores. If the information on your computer is valuable, we strongly recommend it be UPS protected.

    REMEMBER –
    It is better to be safe than sorry and taking the time to purchase a properly rated surge suppressor when you purchase electronic equipment is money well spent."
    (From a public utilities company)

    I also found this: http://www.rbs2.com/outage.pdf
    Bottom of pg. 33 power company pays out for damaged computer equipment due to power cut.

    http://www.rbs2.com/utility.pdf
    The guy who wrote that is an electrical engineer and there are a few cases regarding power cuts damaging equipment, some won and some lost.
    He acknowledges that irregular input electricity and cuts can harm modern equipment and UPS can protect against damage.

    http://www.adrc.com/ckr/hdd_myths.html
    "Myth 2: Can Power cuts causes bad sectors ?
    Typically, modern hard disk incorporates certain techniques to park the heads wherever there is a power cut. However, severe power surge or unstable power supply could cause electronic or RW head failure which in terms create platter damage or bad sectors. Generally power cuts may just cause logical data corruption."

    What numbers do you want? It is difficult to understand.

    Losing data and bad sectors being created by sudden power loss is a reality. I'm not going to try to break my drives trying to prove it to you though. There is plenty of evidence out there already.
    Last edited: May 20, 2013
  23. silkstone

    silkstone

    Joined:
    Nov 1, 2008
    Messages:
    2,821 (1.36/day)
    Thanks Received:
    474
    Here is a typical electricity pole.

    [​IMG]

    A better organized one.

    [​IMG]

    A montage of various electricity poles.

    http://vimeo.com/30252956

    There is no load balancing. Nothing is grounded. The neutral wire isn't always properly earthed and there is a too high neutral to earth voltage.

    An LA times article about electrical safety here: http://articles.latimes.com/2010/jun/05/world/la-fg-vietnam-safety-20100605

    If you don't believe power surges and power cuts can damage hard drives, do an experiment yourself. Turn off the computer at the mains while your disk is reading and writing. Repeat 20 times and then scan for bad sectors. I'm not sure how you would simulate a power surge to test that. But i did find this: http://www.dtidata.com/resourcecenter/2007/08/07/hard-drive-recovery-power-failure-surge-brown-out/
    More evidence that power surges can damage HDDs.



    You wanted numbers - Here are some numbers from a study done in S.E. Asia: http://www.copper.org.sg/sites/default/files/publications/files/UNEP2012-PQ_0.pdf
    Or you can buy this paper: http://ieeexplore.ieee.org/xpl/logi....ieee.org/stamp/stamp.jsp?tp=&arnumber=343378
    another study
    http://cdn.intechopen.com/pdfs/14957/InTech-Consequences_of_poor_power_quality_an_overview.pdf

    And here is an advert - http://chasepower.com.au/documents/Resolve-Power-Quality-Problems-with-a-UPS.pdf

    And a technical explanation: http://ecmweb.com/sagsswellsinterruptions/when-does-poor-power-quality-cause-electronics-failures
    Last edited: May 20, 2013
  24. de.das.dude

    de.das.dude Pro Indian Modder

    Joined:
    Jun 13, 2010
    Messages:
    7,474 (5.02/day)
    Thanks Received:
    1,936
    i live in a metro, so all our power lines/ phonelines are underground.
    only the internet and cable lines are overhead.
  25. westom

    Joined:
    Dec 5, 2009
    Messages:
    180 (0.11/day)
    Thanks Received:
    33
    Underground or overhead make no difference to electrical anomalies discussed here.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page