[Sulfur]

I'd been having troubles with burning out power supplies, causing persistent crashing. I traced the problem to inconsistent power in my house (apparently caused by power lines being grouned through trees they were touching. I live in a very woody area). I bought a new power supply (550W) and a UPS (345W). Now, I didn't notice the UPS' wattage til just the other day, and I bought both of them a month ago. My crashing problem has been returning.

So, I'm wondering if the UPS not supporting enough wattage could have damaged my power supply directly, or if perhaps the lack of enough wattage made some of the power not 'backed up' and could have indirectly caused failure by allowing the faulty power to do damage through the UPS.

Edit: or perhaps, is it simpler. Whenever the power supply tries to get more than 345W from the UPS, it cannot - so the computer restarts (or occasionally the program just closes)?

2nd edit: nope. Problem persists with UPS out of the system.

Any insight would be great.

Edited by Sulfur, 29 May 2006 - 08:30 PM.

[Advertisements]

There's a possibility you may have damage to PC from all your outages. I'm sure you cleaned out PC after replacing power supplies and checked condition of fans, etc..

Possibily download Everest Home Edition or SpeedFan and see if you can monitor some of the specs to if you have radical changes in power supply voltages, fan speeds, hard drive problems.

Possibly see if the power company can monitor your line voltage coming into your house using an event recorder. The UPS normally is only rated for what it can handle under brownouts or no voltage conitions for a certain length of time using backup batteries in a dc to ac configuration. Normally when connected to a working ac line it will handle your PC ok.

Ron

[Major Payne]

There's a possibility you may have damage to PC from all your outages. I'm sure you cleaned out PC after replacing power supplies and checked condition of fans, etc..

Possibily download Everest Home Edition or SpeedFan and see if you can monitor some of the specs to if you have radical changes in power supply voltages, fan speeds, hard drive problems.

Possibly see if the power company can monitor your line voltage coming into your house using an event recorder. The UPS normally is only rated for what it can handle under brownouts or no voltage conitions for a certain length of time using backup batteries in a dc to ac configuration. Normally when connected to a working ac line it will handle your PC ok.

Ron

[Sulfur]

After watching speedfan a while I noticed a few irregulaties in the voltages. I'll just list the stats:

VCore - running from 1.62 to 1.66
+12V - From 12.6 to 12.8
3.3V - running from 3.07 to 3.12
VCC - from 5 to 5.1
5VSB - running from 4.95 to 5

All of my fans are working. All the temperatures are quite fine. CPU gets up to 45C under full load, and hard disks rarely break 30C.

Memtest86 shows my memory as working fine. Mainboard, video, and sound card drivers are all updated. Windows is updated. DirectX is 9.0C.

Major Payne - thanks alot for you're reply. I am still curious, however, regarding the effects of a UPS that doesn't support as much wattage as a computer pulls. I'm fresh out of extra power supplies to test things with, and would rather know if this could be damaging to them before buying another and blowing it too.

Two ideas I just had:
1- perhaps the Video card is getting hot. I don't know how to check this things temperature. Any tips?
2 - One of my RAM slots is burned out. A stick of RAM got put in upside down, which resulted in some charring and horrible smells. The slot is broken, and not in use now. Maybe the damage to this slot affected other parts of the board?

Thanks in advance.

Edited by Sulfur, 31 May 2006 - 07:59 PM.

[Major Payne]

Voltage fluxuations really don't seem that bad, but were these monitored under your heavy use of PC? If you have access to one of the Digital Multimeters with a temp probe, you might be able to check card's temp. RAM problem could have had a backlash to other components. A 550 w PS is more than enough to handle just about anything. What is the mean time between failure of PS? Don't use SpeedFan, but maybe it has some way to check what you need to check. I would go through every item and see if there is anything you can monitor or looks suspicious.

Any one else got any ideas just step in here.

Ron

[Sulfur]

Major Payne wrote:
"What is the mean time between failure of PS?"


Do you mean how long til the errors popped up after installing a new power supply, it's about a day. However, I've moved to a different house recently. So, I shouldn't be having any PS failures due to bad power anymore. However, with a UPS I shouldn't be anyways.....

Edited by Sulfur, 02 June 2006 - 01:04 PM.

[Major Payne]

Actually it was the mil spec MTBF before something fails. In this case, I meant the Power Supplies. Sorry for not being more specific on it. Post back if the move helps solve the PS failures. Normally, on a UPS that is properly installed, you should have better protection against voltage fluctuations.

Ron

[Johanna]

Because of the detail involved in diagnosing the UPS/PS/Burnt Memory, I am going to move this to Hardware.



The UPS and the PS are compatible, btw.



Johanna

[Sulfur]

Thanks for the move.

I was a little unclear about two things earlier.

The failure's aren't consistent in their timeframe. The computer only encounters the problem when I enter CPU and RAM intensive games. Once I do this, the failure usually happens between 1 and 60 minutes later. On a few occasions it's taken 3 to 4 hours for failure to occur (or for me to quit before failure occurs).

I was living in the location with the bad power for one week after installing the new PS and UPS. After 1-2 days of this week going by, I encountered the failures. After moving, I continued to encounter problems.

Eliminating the possibility of UPS and PS incompatability, here's what I've got.

Memory/Motherboard error due to burnt out slot
Power supply error due to DOA
Power supply overheating
Video card overheating (very unlikely)

Is there any possibility of the UPS being DOA resulting in fluctuating power leaking through? I assumed not, as it seems any UPS failure would cause it simply not to power things.

Anyways, I'm in the process of getting another PS to try out. Also searching for something to test video card temp with.

[WinCrazy]

What make & model power supply do you have now ?

I'm glad you are getting another power supply. The one you have is not operating as well as it should. The major outputs should be within +/- 5% of their nominal voltages. The 3.3 volt rail is sagging. This could easily cause random crashes. Your 12 volt output is high but I don't think that would cause a crash.

Good voltage ranges are as follows:

+3.3: 3.135V to 3.465v
+5: 4.75V to 5.25V
+12: 11.4V to 12.6V

[Sulfur]

I put in a power supply that had been broken back when I had no UPS and had house with bad power.

The symptoms back then were the same as now. This power supply acted just like the other, in that it crashed after 40 minutes of a game running. However, voltage stuff was different.

Vcore 1.62-1.66
+12V 12.6-12.8
3.3V 3.2-3.25
VCC 5.02-5.07
5Vsb 4.98-5.02



My question: if a power supply was damaged by voltage fluctuation coming from the source (outlets), would the damage be reflected in bad voltage values? Or, would it intermittently fail like I experienced without voltage problem prior to the failure (and presumably some kind of problem at the time of failure)?

I'm going to test my 'broken' power supplies on someone elses computer to see if they are broken, or perhaps it's something else.

[WinCrazy]

Damaged power supllies are most likely to produce voltages that are wildly out of range including no output at all on one or more rails.

Supplies with sagging outputs are usually due to either being overloaded or due to bad designs.

Since you have tried multiple supplies which all resulted in system crashes I think it's safe to assume the crashing problem is not due to the supplies. Continual crashing even after you used a UPS backs this guess up.

We should look to other possibilities that are more likely. Since the CPU temp under load is good the most likely suspect is RAM failure. The best way to check this is with Memtest86, bootable off a CD. It is available as an ISO image of a CD and is burned with a program such as Nero, Roxio or the free CDBurnerXP. Make sure the BIOS is configured to try to boot from the optical drive before it tries to boot from the hard drive.

Let Memtest86 run for 2 complete test cycles. There should be no errors. If there are errors and you have multiple sticks of RAM run the test with only one RAM module installed to isolate the bad one(s).

[Sulfur]

Yeah, I ran memtest a bunch. In one of my earlier posts I mentioned this briefly. After 2 cycles no errors, and a few other times I ran it once or twice with no errors. It's thouroughly checked.

Well, I just noticed something I've overlooked about speedfan. It can keep a log. So, at the time of failure I didn't notice any odd voltage changes. CPU temp got up to 49 right before failure, but I've had it idling at that before without crash (when I had a lesser CPU fan).

Edited by Sulfur, 07 June 2006 - 02:38 PM.

[WinCrazy]

Well, all the usual suspects leading to crashes have been discounted. What's left is that something in the XP installation is causing the system to crash and it's not an often used peice of code. Reinstalling the OS would tell you if this is the cause.

Another slim possibility is that the motherboard may have a failing component on it.