Frequent crashes, looking for advice.

[ghostjeep]

Hello all. My PC took a huge surge a few months ago and even though it was hooked to a surge protector it got zapped. I took it into a local shop where they said the motherboard was dead and possibly the RAM. The guy said that he ran a mem test on the RAM and it first showed problems but then after a few more tests they showed ok. So I replaced the mobo with the exact same model. I also purchased more RAM and took out the old. I was running 7950's in SLI but also decided to upgrade the card to a GeFroce 8800gt (evga). Afterall of this I am getting frequent crashes when playing video games (WoW, and Tabula Rasa, Sins of a Solar Empire). Sometimes it will just lockup the screen, sometimes I get hanging sound w/ lockups, I get the BSD occasionally with a bad spool error, or sometimes it just resets itself. The only time I have these problems is when playing games. Any ideas on what to check for next, or how to check it?
Here is some system info

Operating System: Windows Vista™ Business (6.0, Build 6000) (6000.vista_gdr.071023-1545)
Language: English (Regional Setting: English)
System Manufacturer: System manufacturer
System Model: System Product Name
BIOS: Phoenix - AwardBIOS v6.00PG
Processor: Intel(R) Core(TM)2 CPU 6600 @ 2.40GHz (2 CPUs), ~2.4GHz
Memory: 3070MB RAM
Page File: 876MB used, 5444MB available
Windows Dir: C:\Windows
DirectX Version: DirectX 10
DX Setup Parameters: Not found
DxDiag Version: 6.00.6000.16386 32bit Unicode

And the psu is:
Apevia / Iceberg / 680-Watt / ATX / Dual 80mm Fan / 20/24 Pin / SLI-Ready / SATA-Ready / Blue / Power Supply
The RAM is:
A-Data 1g DDR2 800

Thanks in advance for any help or info.

 

[roadrunner197069]

Well if you still have the same PSU it probably got damaged also. The spike that damaged your mobo came through the PSU first.

 

[Nik_I]

Well if you still have the same PSU it probably got damaged also. The spike that damaged your mobo came through the PSU first.

that sounds like the most likely cause of the crashes.

 

[RetiredChief]

Agree with previous two post. The most likely path would be through the PSU. You didn't indicate the cause of the surge. Load increases normally cause a decrease in voltage, while a lighting stike (Not necessary for direct hit) will cause large spike and can also come in on coax or ethernet lines.

That said, best way to check PSU is also most expensive - swap it out.

You coud try this:
Down load CPUID, ATI Tools (If 3d view will work on your 8800 card), and either TAT, prime 95, or orthos.

Run CPUID, Note Voltages and temps (Nominal/min/max)
Run Tat, or orthos, or prime 95 (To load CPU) - Do voltages / temps stay with in Specs
Voltage Specs are normmally +/- 3%. Fluctuation should be 2% or less But should not drop you below the 3% point.

Stop CPU loading, and run ATI tool, Show 3d view (or another program that will load your graphics card. Monitor GPU temps and +12 V
Same criteria on voltage.

 

[ghostjeep]

Thanks for the replies, Ill check it out asap. As far as the cause it was the circuit was overloaded and sent a surge thru the lines and blew out the actual outlet that the surge protector was on, this was ( as I was told I'm not an electrician ) from too many amps being drawn and old/faulty/improperly set up circuits in the house. It didn't blow a fuse or trip the breaker there was actual damage to the outlet and circuit the room was on.

 

[RetiredChief]

Not sure weather PSU/surge protector was the cause or the outlet - a case of the chicken and the egg. The Power supply could have caused the overload because it failted or as a result of a defective outlet; or the outlet was deteriated to the point that it was it in itself caused the PSU to fail.

I've replace several outlets in my house as (1) loose fiting, (2) Zero volts out (wire had come loose). One of the failure modes for outlets is that the contacts make a loose connection which causes (1) posible arcing (not good), or (2) increases the rate of oxidation which inturn eventually causes the electical contact to overheat. PS I would like to ZAP the electrical guy that wired my house (Can't shoot him)

Note: I perfer a UPS, one that corrects for broun outs. I really perfer the ones with a regualted output, but they are much more expensive (My last one cost $450 and lasted about 5 to 7 Years) Reason being, when the vacuum clearer is turned on, you can see the light on that circuit flicker.

 

[ghostjeep]

Hi, I ran all that I the test I could and here is a couple of the numbers I'm getting.
It appears the +12 varies from 11.73 to 12.83.
The gpu varies from as low as 64c to as high as 74c.

Does anyone know if these numbers look too far out of whack?

 

[jsc]

The 12 volt output should not vary and the high limit is technically out of tolerance. If separate checks (safe and simple place to check is between the yellow and black wires on a hard drive power connector) with a digital multimeter confirms that, it's time to swap out the power supply to check it.

GPU temps tend to run higher than CPU temps.

 

[RetiredChief]

The fluxuation in the +12V (if valid) are excessive. If you are refering to the test I posted. Also, As jsc pointed out 12.83 is out of spec but could be a "software glich" Also as jsc pointed out, if you have a DVM you can verify the voltage. But looks like that PSU is a good canidate to swap out.

What you want to note is that the nominal values are stable at idle. When runing test, ie ATI tool 3 D view, You will see a DECREASE in +12 V, But it should also be stable at this lower value (Monitoring current value)

EX My +12 is rock solid @ 12.10 V, When I run ATI Tools 3D view it drops to 11.86 BUT is solid at this lower value.

The amount of decrease Depends on :
(1) Where the Measurement point is - It may be a IR drop on the Mother board
(2) quality of PSU and
(3) The degree (percentage) of change in Current and also percentage of Current drawn / max Current capacity.

jsc - I think you will find that the +12V MAY l decrease with large change in current draw ( ie going from Idle to 3D gaming.) This is one of the test they do on PSU's in terms of % regulation, going from min to max loading. If the loaded +12 v is only going fro 30 % idle -> say 60% loaded And Based on My system YOU ARE CORRECT. Measuring the +12 as you said only drops my +12 by 0.03 Volts BUT I have a 850 Watt PSU about 20 % load at idle and max of 50% loaded
I think the delta (based on (1) above is the reason for the change.

DVM measurements:
+12 V Measured at 24 Pin connector, +12 V @ 8 Pin connector (MB), +12 V measured at 6 Pin GPU , and +12 V at Molex (HDD) connector.
Results idle voltages measured 12.15 (Molex) -> 12.18 (8-pin), when loaded +12 V whent to 12.14V (only a 0.04 drop @ MB.

Edited, based on DVM measurements