Electrical Problem Unknown Cause

Discussion in 'General Hardware' started by PogiMan, Oct 10, 2011.

  1. PogiMan

    PogiMan Geek Trainee

    Likes Received:
    0
    Trophy Points:
    1
    I have a PC that was just recently built. Since then, I had two experiences of "electrical crashes" (for lack of a better way to describe them).

    The first time, when I turned on the power, I saw that there was current passing through it because the fans were rotating and the LED lights turned on, but no power getting into the motherboard -- the HDD lights didn't light up, and the monitor didn't receive any video signal. The second time this happened was three months later, and the experience was similar.

    The solution I learned to solve this scenario: (1) Unplug the machine; (2) Unplug the cords from the PSU to the motherboard and CPU; (3) Remove all the RAM sticks; (4) Wait for a few minutes; (5) Slowly put back each RAM stick and test.

    I just recently tried the solution without removing the RAM sticks, but this didn't work, and so when I tried removing the sticks and putting these back, the system became okay again. I suspect there is some static built up somewhere, and maybe the RAM has a bigger role in this problem, and the act of unplugging the power and removing the RAM discharges it.

    Hardware specifications:
    CPU: Intel Core i7-2600
    Motherboard: Intel DH67BL [B3 revision]
    Video: No discrete GPU
    RAM: Kingston 4 GB DDR3 SDRAM x 4 (16 GB total memory) [verified compatible with motherboard]
    HDD: Seagate Barracuda XT 2 TB x 2 (4 TB total storage)
    PSU: Cougar GX1050 (1050 W)
    Chassis: Antec DF-85
    USB peripherals: Razer Lycosa keyboard, Genius mouse
    Monitor: Samsung LCD.

    Other incidental information: I used one of the styrofoam packaging as a "base", so the machine sits on top of it. I'm starting to suspect this might be a cause, but no solid proof yet. I've looked up more information concerning static discharges and the triboelectric effect; long-story short and I now have a 1 cm thick wooden board that separates the base of the chassis and the styrofoam it sits on. The wooden board came from a part of my computer table; let's just say the chassis was too big to fit into the wooden table's space for the desktop, which is why I just sat it on the styrofoam packaging in the first place. I did this so that hopefully the wooden board will mitigate any static that might be because of the styrofoam packaging and the contact with the chassis. Again, I don't have any conclusive proof that this could be a reason, so this is just a "just-in-case" measure. (Now that I think about this some more, I might eliminate the risk[?] altogether by removing the base and putting my machine on top of something else -- thoughts on this as well?)

    I have not been able to successfully replicate this problem. Frankly, I hesistate to even attempt to do so for fear of damaging my system beyond repair.

    Questions:
    1. Does anyone have an idea of what's going on here?
    2. Is there something wrong, and what is it?

    Any insight will be much appreciated.
     
  2. Ghostman 1

    Ghostman 1 Mega Geek

    Likes Received:
    85
    Trophy Points:
    48
    Try a different power supply to narrow down your problem... Never heard of the one you have... Power supply could have a short in it..
     
  3. Wildcard

    Wildcard Big Geek

    Likes Received:
    26
    Trophy Points:
    28
    Hi,

    I would also definitely get rid of the styrofoam as it could build up static electricity and cause harm to your computer components. Static electricity can kill a cpu. A cardboard or wooden box would be much better than styrofoam to set the pc on.
     
  4. PogiMan

    PogiMan Geek Trainee

    Likes Received:
    0
    Trophy Points:
    1
    I hear you guys... I definitely got rid of the styrofoam now... And after some digging around, I'm now planning to get a replacement PSU.

    To share some of my findings:

    From this source (with some paraphrasing, so not an "exact" quote): http://www.overclock.net/power-supplies/719397-ripple-its-effects-overclocking.html
    From this source, http://www.anandtech.com/show/3992/1000w1200w-highend-psuroundup/9, I made the following conclusions (hopefully I interpreted what I read correctly -- somebody keep me honest if that's not the case):

    It is known that the Cougar GX 1050 has a high voltage output for the +3.3V rail upon startup. Based on AnandTech reviews, it can start as high as 3.4V at only 10% of the load. Ripple quality still seems to be within ATX standards. It is known to have high ripple at the 110% loads for both the +3.3V and +5V rails; the +12V rail is okay. In any case, my impression is that it is known that voltage output for the rails can fluctuate.

    It is also known that my motherboard, Intel BH67BL, has some form of automatic voltage regulation on the memory so that it is within the 1.2V - 1.8V range (http://www.intel.de/content/dam/doc/product-brief/desktop-board-dh67bl-media-brief.pdf).

    Theoretically the motherboard should be able to handle voltage fluctuations, but the effect of ripple can throw this off. Thus, it's further possible (though not proven, just my speculation) that ripple for this PSU is also not good even when not at 110% loads.

    Possible causes of electrical crashes: Possible root cause is the high output voltage fluctuation in the +3.3V rail, where ripples could possibly still happen (again, no proof of this, just a suspected cause!), resulting in the need for discharging, and in particular, the need for RAM reseating.
    1. During the first time that the electric crash occurred, it is possible that the +3.3V rail had a high output, and it was beyond anything the motherboard was able to handle.
    2. In the second time that this happened while the machine was in use, it is possible that the voltage fluctuated, thus sending a high output, also caused the system to hang.

    Short-term solutions:
    1. If the crash happens before the machine is actually in operation, then there's nothing that can be done but to follow the reseating solution I already mentioned. One thing for me to try if there is a next time: Keep the RAM sticks in place and test if this also works.
    2. What I can try to mitigate is the electric crash once the machine is in operation. This can be done by increasing the load on the system. I have a couple of VPCs that I can load up that will do the job fine.
    3. Something else to consider: Reduce the number of RAM sticks used. With relatively less memory running to work with, the system will have more load to deal with in its memory.

    Long-term solution: Replace the PSU with a recommended unit. I'm seriously considering the Antec High Current Pro 1200W now, based on feedback from AnandTech, unless someone else has other suggestions.
     
  5. Wildcard

    Wildcard Big Geek

    Likes Received:
    26
    Trophy Points:
    28
    Hi,

    That looks like a really nice power supply and I read some good reviews about it. It should work well for your system if you buy it, but you may want to look at a wattage calculator before you buy. Based on the information you provided about you computer, the power supply seems to be overkill for what you need. I tried the Asus wattage calculator and also the newegg one and both said you only need around a 400-500 watts power supply. Now, I am not saying that you shouldnt buy this one you are looking at, because if you decide you want to upgrade and start doing SLI or Crossfire down the road it would do well with that setup. However, if you are not planning on that, you will be paying more than what you really need for your current setup (but it will work for your computer without problems).
     
  6. PogiMan

    PogiMan Geek Trainee

    Likes Received:
    0
    Trophy Points:
    1
    Thanks for the feedback. Yeah, I agree with you that it is overkill for my current setup. I'm looking at this as more of a long-term investment. Eventually I'll want to upgrade my rig, and rather than buy a new PSU when that times comes, I plan to re-use this. I don't know how much wattage the new components I will put in will need 4-5 years down the road, so a little overkill now could help "future-proof" for my needs later on.
     

Share This Page