cheapo rig shut down problems

[Karn]

Rig keeps either shutting down or freezing ever 4-8 hours. Psu and cards have been checked in my main computer and seem to run fine. I know the best thing to do would be upgrading but it's just not in my budget right now. Any ideas what the problem could be or what I can check would be helpful.



Rig spec's (from an old Gateway GT5082)
cpu       AMD 64 x2 3800+
mobo    FOXCONN C51GU01
ram      2gig hynix pc3200u-30330
psu      cooler master GX 750w
gfx       sapphire 7990, evga 750ti
windows 7

[tom99]

  most likely your gears overheat somewhere then cpu shutdown for safely.

[Karn]

I have it in a home made wind tunnel  (big box with box fans at either end, not just for cooling it also helps keep my cat out) gpu never get's above 65c

[onetwentyfive]

Rig keeps either shutting down or freezing ever 4-8 hours. Psu and cards have been checked in my main computer and seem to run fine. I know the best thing to do would be upgrading but it's just not in my budget right now. Any ideas what the problem could be or what I can check would be helpful.



Rig spec's (from an old Gateway GT5082)
cpu       AMD 64 x2 3800+
mobo    FOXCONN C51GU01
ram      2gig hynix pc3200u-30330
psu      cooler master GX 750w
gfx       sapphire 7990, evga 750ti
windows 7

your psu could be failing after a couple hours, that happened to me twice. eventually i got a bigger psu. if it's just *pow* shutting completely off out of the blue then it's probably the psu. in fact, it could work alright in a regular computer because it will only fail if there is massive power being drawn. just a thought.

[Rookie41]

Had a quick Google seach for some of your components.

Your PSU says 750W, but according to this review http://www.hardwaresecrets.com/article/Cooler-Master-GX-750-W-Power-Supply-Review/917/10 ,
power usage above 600W can cause high noise levels on the +3.3V and +5VSB "rail" which can result in overload/damaged components.

A quick rough power calc for your system at 100% load (from various sources online):
Radeon HD 7990 - 375 W
GeForce GTX 750 Ti - 60 W (according to spec, but review sites mention 100+W under full load)
AMD 64 x2 3800+ - 89W
mobo+ram+hdd - 40+5+7

All up for about 576W (more if your GTX 750 Ti actually uses more than 60W), awfully close to the 600W danger zone tested by Hardwaresecrets)

So maybe run this setup without the GTX 750 Ti and see if it still crashes (using about 516W then).

Your PSU seems to have issues with so called "crossload" : Cross load means high 5V and 3.3V with low 12V or vice versa.
More info about crossloading can be found here: http://www.playtool.com/pages/psucrossload/crossloading.html
TLDR: "If the two loads get too unbalanced then the voltages can get far enough out of their desired range that the power supply has to turn off to avoid causing serious problems. The subject of relative loads on different voltages is called cross loading."

My 2 cents,
give it a go or ignore it if you wish.

[MaxDZ8]

^ Very interesting word of advice.

As a side note, I had lockups related to both northbridge temperatures and RAM (!).
I'm still investigating the issue, but I can tell the lockups were greatly reduced after I pointed a fan to the NB.

This is especially the case for cheap mainboards as they cut on everything not being reviewed, quality coolers are the first thing to go. Your NB cooler seems fairly capable however.

Also check out capacitors and voltage regulators on mainboard.

[Karn]

Had a quick Google seach for some of your components.

Your PSU says 750W, but according to this review http://www.hardwaresecrets.com/article/Cooler-Master-GX-750-W-Power-Supply-Review/917/10 ,
power usage above 600W can cause high noise levels on the +3.3V and +5VSB "rail" which can result in overload/damaged components.

A quick rough power calc for your system at 100% load (from various sources online):
Radeon HD 7990 - 375 W
GeForce GTX 750 Ti - 60 W (according to spec, but review sites mention 100+W under full load)
AMD 64 x2 3800+ - 89W
mobo+ram+hdd - 40+5+7

All up for about 576W (more if your GTX 750 Ti actually uses more than 60W), awfully close to the 600W danger zone tested by Hardwaresecrets)

So maybe run this setup without the GTX 750 Ti and see if it still crashes (using about 516W then).

Your PSU seems to have issues with so called "crossload" : Cross load means high 5V and 3.3V with low 12V or vice versa.
More info about crossloading can be found here: http://www.playtool.com/pages/psucrossload/crossloading.html
TLDR: "If the two loads get too unbalanced then the voltages can get far enough out of their desired range that the power supply has to turn off to avoid causing serious problems. The subject of relative loads on different voltages is called cross loading."

My 2 cents,
give it a go or ignore it if you wish.

Should have added I already pulled the 750ti and it was still happening. Man really wish I had read that psu review before I bought it. lol it figures the parts I got from a PC I found sitting in a trash can work fine. One of the parts I paid for is likely the cause of the problem, and my damage the rest. Gotta love the irony. Thank you for the help

[Rookie41]

Also check out capacitors and voltage regulators on mainboard.

Good tip as the OP's motherboard is a fair few years old some capacitors could be bulging (mostly on top) or even leaking at this stage.
Quick and easy to check, just shutdown the rig, unplug the power and open the case and inspect the motherboard.
Some capacitors could be blocked by the heatsink and might be hard to see.

Should have added I already pulled the 750ti and it was still happening.

Check the motherboard for leaking/bulging capacitors. If that's all good maybe try and replace the PSU with another one (other pc, a friends ...)
It always helps if you research hardware before you buy it.
Hardwaresecrets could have just been unlucky with that PSU and yours may not have the same problem. It's hard to tell.

Replace the PSU with another one (don't go buy a new one until you're 1000% sure your PSU is the problem) and see how you go.
Could even be an OS or driver issue .. who knows.

I have it in a home made wind tunnel (big box with box fans at either end, not just for cooling it also helps keep my cat out) gpu never get's above 65c

Check CPU temp too btw, just to be sure. Altho it sounds like your wind tunnel is providing some good cooling.

[Karn]

Been working just fine since my last post. All I did was take out the ram rearrange it and put it back in(don't think this had any thing to do with it). The other was turn off the intake box fan leaving only the exhaust fan. I think that maybe the cpu cooler may have caused a dead air spot behind it. Thats the best I can come up with lol but at least its working for now. Thank you to all that posted I learned a lot 

[Rookie41]

Glad to see you got it working!

Hope it stays that way!
Happy mining!

[Karn]

Rig went down again so I pulled it apart and found most of one type of capacitors (e61b) are bulging (9 out of 10) and 2 of those are leaking a very small bit. I'm guessing that means the death of my cheapo rig , and the start of hustling up a new cheapo rig.

[kingscrown]

1 card may be fault and doing it - i had this issue

[Karn]

1 card may be fault and doing it - i had this issue

Of course I can't be 100% sure it's not one of the cards. I did check each of them out in my main computer and they mine and pay games just fine.

Edit
Any one know of a place I could get a really cheap mobo that would work with my cpu/ram and maybe support 3 gpu's (2 750ti, 1 280 non x)?